#381 – An Interview with Derek Kozel

Derek Kozel: This is The Amp Hour Podcast. Released February 25th, 2018. Episode 381. An interview with Derek Kozell.

Chris Gammell: Welcome to the Amp Hour. I'm Chris Gammell of Contextual Electronics. And I'm Derek Kozell, electrical engineer and radio enthusiast. Welcome, Derek. How are you doing? I'm doing really well, Chris. Yourself? Good. Where are you coming to us from today? So I'm located in Cardiff, Wales. I moved over here recently.

Derek Kozel: Nice. And I'm sure it's a beautiful country over there. I've seen some pictures. And I heard that you have a new house that you are furiously adding antennas to. Is that a correct statement?

Chris Gammell: That is a correct statement. So here in the UK, pretty much everything is built out of stone and brick. And that means that we have nice sturdy walls. But entertainingly, you can't run cables through them. So what I have at the moment is just a hole in the ceiling of my office with a whole bunch of coax going up into the attic. Nice. Oh, so you still put the antennas internally. I have them going through the attic. And I haven't yet gotten up on the roof. And I'll be mounting some onto the chimney. This house is actually older than electrical wiring. Oh, yes. And so there's no metal in this house, which is phenomenal. Oh. But a little unnerving at times. Yeah. Okay.

Derek Kozel: So, well, you said a radio enthusiast. Tell us your story of getting into radio and where you work and play these days.

Chris Gammell: Sure. So, way back when, during university, I found the local amateur radio club at the university, W3VC at Carnegie Mellon, and joined them and passed the introductory licensing exam for amateur radio. And at the time, I was envisioning it purely as a hobby. You know, I still don't really enjoy math.

Derek Kozel: And so... No one blames you there, man.

Chris Gammell: Yeah, no one blames you there. I had had just enough of my signals and systems second year course to know that I was never, ever going to make this a career. That would be a horrible decision.

Derek Kozel: Yeah, right. You don't like doing Fourier transforms by hand, like the math?

Chris Gammell: You know, the Fourier transform, and then you get to the Laplace transform. The Z transforms. The Z transforms, the Hilbert, and you're just like, whoa. Why are we in totally... Couldn't the computer be doing this? And it should be. It absolutely should be. And I can get into a long divergence about the, you know, educational system of... Oh, yeah. How would you actually teach this type of material?

Derek Kozel: Yep.

Chris Gammell: But... Maybe we should. We can get there later, yeah. We can definitely get there later. And so, what I thought was, hey, I can push a button and I can talk. I'd be a great person for the amateur radio hobby. And so, yeah. I joined and spent the next couple years progressively coming to the reluctant understanding that this was really cool. And I really wanted to do a lot more of this.

Derek Kozel: Is that going to be... Is that like, oh, I am going to be a nerd?

Chris Gammell: I think that realization had come many years before. I know. Me too. But, yeah. So, I joined W3BC. I ended up, you know, inevitably becoming one of the officers because there were so few people willing to become one that pretty much everyone who raised their hand got some position.

Derek Kozel: Yeah.

Chris Gammell: And I... It was great. We had our own...

Derek Kozel: I actually don't know what... Like, so, an amateur radio club at a university. Like, what is... I mean, is that like doing like field days and stuff like that? Or what is the roles and the events that are happening?

Chris Gammell: So, entertainingly, field days during the summer. And so, no. Oh, yeah. Not at all. Right. A few people, of course, who live in the local area will come together and try and do one. But what we found was the main core of our amateur radio club was the fact that we had this wonderful room at the top of the engineering building right up in the roof. And we were able to install a whole bunch of repeaters in that room and then run coax up actually this chimney stack. Because Carnegie Mellon was founded long enough ago that the city electrical grid didn't run out to them at the time. So, they had their own coal power generator in the basement of the engineering building. And freshmen and sophomores had to take shifts piling coal into the generator. Way back when, of course. And so, now all that's left is this phenomenal building with this giant smokestack. And so, we stuck a whole bunch of repeaters up there. And basically, what you would do is you would walk out of your classroom, reach down to your hip because you were just this cool, turn on your radio and call out, you know, AG6PO, is there anybody out there? And we were a large enough club and had enough other people on campus licensed that pretty much inevitably, there was somebody else who had their radio on who was just walking around. And so, it was really actually a hugely social experience. Sort of. Sure. Hey, come on. Okay, sorry. What we would end up with is we had shack nights every Thursday night where, you know, whoever was available would go up into the radio shack. And we had a collection of nice but very old equipment for the most part and a couple different operating stations. And part of what we were doing while I was there was trying to clear out some of the truly museum pieces. Right. From the back. The boat acres. Right. Yeah. Yeah. And find new homes for them. Sometimes recycling, sometimes selling it on just to, you know, find a good home for it.

Derek Kozel: Yep.

Chris Gammell: And as part of that, one of the things that we were doing was trying to bring in more computers and bring in, you know, nicer test equipment and gain a more, I guess, engineering and scientific understanding of like what the condition of our wiring and stuff was in. Mm-hmm.

Derek Kozel: So, so, so, so, so this sounds like it was more like the clubs about using it. And then it was also, you guys are trying to transition into building. I mean, were you doing like, like QRP stuff, like low power and things like that? Or was it mostly just use and HTs and stuff like that? Yeah. Right.

Chris Gammell: Oh, I would never. But yes. Oh, sorry. Oh, sorry.

Derek Kozel: Chris just swore on there. Yeah.

Chris Gammell: No, they, the Baofangs. That's what I had. The Baofangs are astonishing radios for the price point. Yeah. And one of the interesting things is they're a little hit and miss. And by a little, I mean a lot, but they can actually be astonishingly clean in terms of their spectral emissions. Mm-hmm. We, we did a test at one point comparing it to a Motorola and several of the Baofangs we had were, were right up there at Motorola quality on the transmit.

Derek Kozel: Mm-hmm.

Chris Gammell: But we found on the receive side, they were pretty much garbage. Uh. Rather deaf. Mm-hmm. But yeah, I have a, I have a Yaesu, uh, VX6, which is a nice submersible, uh, handheld. And it, it covers a few different frequencies and it goes up to a maximum of five watts on the transmit side.

Derek Kozel: Oh, okay.

Chris Gammell: And you're like, hey, five watts, that sounds great. But that actually only gets you five, ten miles, depending on where you're at. I mean, if you're at the top of a mountain, yeah, you get a lot more.

Derek Kozel: Mm-hmm.

Chris Gammell: But it's really inconvenient to have to walk up to the top of a mountain every time you want to talk to somebody. Right, yeah. It's way more convenient to, like, drive up to the top of a mountain with a whole bunch of equipment, dump it on the side of the mountain, set up an antenna, set up a radio that's always on. And it's always listening on a certain frequency. And anything it receives there, it'll transmit out again on a different frequency.

Derek Kozel: Mm-hmm.

Chris Gammell: We call this a repeater. And they also tend to be much higher power. You know, 20 watts, 50 watts, 100 watts. And you stick that up as high as you can with great line of sight. And all of a sudden, now you're not talking five, ten miles. You're talking 20, 30, 50, 100 miles on a regular basis. So that was our big thing that we had going.

Derek Kozel: Repeaters always kind of felt like cell phone towers to me, too. Like, they're not networked necessarily. I mean, I know they can be networked, but it's more like, you know, you talk to the tower and then it basically amplifies your signal. So you don't have to have that handheld power like you're talking about.

Chris Gammell: Which, of course, is not what a cell phone tower is doing for the most part. But, yes. Right, right. Yeah. Cell phone towers generally are receiving you, decoding your voice data or your digital data. Right. And then putting it onto a network. Right, exactly. And that's the private cell phone provider's network. And then they'll switch it out and send it back. But very, very infrequently are you actually just sending to the radio at the cell tower and then going right back down to someone next to you. There's usually a lot of network involved.

Derek Kozel: Yeah, cool. So, okay, so how did this transition into other stuff for you then? So this got you into the, obviously, we're talking about, you know, more the analog side of things. But obviously, well, I mean, we don't even told people where you're from yet. So, yeah. How do we get to where you are currently?

Chris Gammell: Sure. So this was in Pittsburgh. And I was an engineering student. And so I was looking for summer internships. And ended up, through a school competition, one of the judges was from SpaceX. And they came in and they were, you know, judging. And they found our project to be the most boring, but I guess well implemented one or one of them anyways. And so I got an interview and a summer job out of that. And that's, that was my first time working with radio systems as a job. Wait, which project was this though? So my project was building a sensor, a wireless sensor network for the first stage of the Falcon 9. Oh, I thought you meant this is the school project. Is this the school project or something? Oh, no, no. That was the SpaceX project. But the school project, we were doing, again, a wireless sensor network just using little ISM band radios. The RFM-22s, I think. Cheap, easy, simple. But the problem that we saw was that you can get a bunch of these, but what you really want to do is get that data onto a computer. So we built a little board with a little AppMega on it that was bridging the ISM band radios onto a Wi-Fi radio. And so you could plug in your sensors, run it over the ISM band, and then it would bridge up to Wi-Fi. And all of a sudden you had data on your computer.

Derek Kozel: Wow. Nice. Okay. And then basically you went and did the same thing at SpaceX? Pretty much. Yeah, do that on a spaceship now.

Chris Gammell: But better. Right. Yeah. Make sure it actually works, huh? Yeah. So I learned a lot about reliability and actually planning what you're going to do before you do it at SpaceX. Right. So that was a really fun summer job. And it was also my first summer job that was really based on hardware. Previous to that, I had been doing software jobs. You know, I'd done some Android app programming for a summer job. I'd done QA for two summers, which was all like Perl and Java. Because in university, I was doing electrical engineering, but it was, you know, an electrical and computer engineering degree. And so it was a mix of both.

Derek Kozel: Right. Right. Well, there's always that hard thing, too, where it's like you have a student who's interested in hardware, who hasn't built that much hardware. And then, you know, you don't necessarily have a project that you can be like, hey, just go start from scratch. You know, like you got to give them testing jobs are always good for that stuff. You know, that's how I started. Repair is a great way to start, you know, if you can if you can find that work. So, yep. I think we all know that struggle.

Chris Gammell: Yeah. Yeah. Now, one of my one of my good friends who I hugely admire as an engineer, he had a job during middle school, which was doing Q&A on hardware on an assembly line. Oh, really? I think he learned a phenomenal amount from that. I'm envious of that experience.

Derek Kozel: Exactly. If you can get if you get hands on that soon, you like start to see these patterns that just show up over and over again. And the sooner you see that, the better off you are.

Chris Gammell: Yeah.

Derek Kozel: Cool. OK.

Chris Gammell: After after that summer internship, I, you know, confessed, confessed my sins and said, I want to be a hardware engineer and I want to be involved with with radios. Man. So you got the sickness. Oh, man. Do I ever. Right. And yeah. So now I now I now I do that as a job. I jumped through. So let's let's have the big reveal here. So what do you do now? Sure. So I work with software defined radios and I'm employed at Edis Research, which is a subsidiary part of National Instruments. So my full time day job is is sitting designing very fast, very good performance software defined radios. Nice.

Derek Kozel: And keen listeners of the show may recognize the name there. edis research uh back on show oh wait what was it 101 we had matt edis founder of edis research on this show uh with a a picture that we were apparently not supposed to use uh we've been yelled at since that i just grabbed a picture off of matt but uh uh yeah he was great too and um so edis research been that was six years ago so that's been going obviously much longer than that and matt's apparently on to new things but the company's still going strong and doing uh lots of sdr stuff so what uh like you said national instruments is tied to it what what are people doing with all this stuff these days i mean what are what are what are some of the the sdr applications that are most exciting most exciting um you should have given me

Chris Gammell: you should have sent that in an email about a week ago then i'd think about it oh there is so much cool stuff that gets done with software defined radio these days i i i started off at edis research doing actually part-time customer support because i had another job during the day wanted something to fill the evenings that was again radio related and we get questions from customers doing all sorts of really really cool applications and then i i've been happily able to go to the go to a conference uh for the last three years that's all based around software defined radio and we have seen you know radar systems that are scanning the skies for space debris uh we've seen people using software defined radios actually for uh ultrasound research oh wow because an electrical signal is an electrical signal uh in the end and the the adcs and the dax the analog to digital and digital to analog converters in these radios in the end don't really care what's generating that whether it's an antenna or a transducer so yeah we we see those uh there was a video that i'm pretty sure it was an edis radio but it it was a software defined radio for sure and this guy was building his own mri machine on his on his like desktop bench it was totally ludicrous and awesome uh yeah so i love seeing applications like that

Derek Kozel: and i i do really like that too of the um you know the the general feel of it right so like you're talking about there's a front end there's an adc there's processing the middle you're trying to do as fast as possible and then if you need to go back in the other direction you put it into a dac back to a front end and that's kind of it right and then you just kind of switch out those pieces is that does that sound right or is that am i missing a piece in there too no not at the high level i and

Chris Gammell: that's one of the wonderful things is if people like me do our job really well then you can pretty much treat a software defined radio as a at least a gray box where i like that you send you send the bit stream you know the the mathematical equations that you've you've written that are generating some signal on your computer you send it to the radio and it just magically appears at the antenna port and it's okay it's our job to change that signal as little as possible um but you end up with inevitably the analog realm happens yeah um or really interesting things like in most of these radios many of these radios i should say have fpgas on them and they're using that to do a little bit of signal processing um usually at very high sample rates and so most of that ends up being uh integer uh fixed width mathematics uh what what's the word i'm looking for um fixed point fixed point yeah fixed point math yep and so you end up with really strange little rounding errors yeah yep most of the time you're not going to notice but if you end up in a really really demanding application yeah you start

Derek Kozel: noticing those effects yep that's interesting uh so okay so that sounds like definitely fpgas and the processing piece and doing that fast is one hard part uh so where where are most of your where is most of your time spent i mean are you spending it on the fpga side are you spending on the software side are you spending on the antenna side like so where where where is the hard part of sdrs these days for especially for people getting started for people getting started it's the software absolutely okay

Chris Gammell: um because if you're just getting started you're probably not building your own hardware um antennas filters amplifiers yes all of those are not required but really really useful right right um and and i think they all end up underappreciated by beginners um yeah because again they're they're looking at this as a black box and they're saying an antenna is an antenna is an antenna right no no right i've got several pounds of paper next to my head right now that uh would tell you in about a thousand pages all the

Derek Kozel: reasons why that's not true i was gonna say i thought pounds of paper will not actually be a good antenna i

Chris Gammell: i thought you were gonna go that direction yeah no no but um uh antenna theory analysis and design um by blanis yeah one of my favorite textbooks of my entire university career hands down on i i was gonna say unreadable that's not true very difficult to read yeah right um well once you speak the

Derek Kozel: language though i'm sure that it does start to make sense right especially with like and that's i'm sure well we could very easily fall back in the education side of things right there there is a ton of math in the antenna side of things right especially if you want to start modeling it and changing things and

Chris Gammell: like getting optimal designs right absolutely and the hardest math i think i've seen anywhere has been intended design math um because you start getting into the realms of actual uh electromagnetism and it turns out that there aren't simple closed form solutions for most intended designs um and so you end up in this this really weird and wonderful gray area of we can kind of simulate it with you know variable granularity but you're not going to get the complete one defining answer right exactly

Derek Kozel: yeah that's fun but but but the problem being that they start with that math instead of what you're talking about with beginners is starting with a just doing it right i mean it sounds like sdr really just enables you to start transmitting and receiving and seeing seeing where all the pieces

Chris Gammell: fit together absolutely and i i've stayed with the amateur radio side of things ever since i got started it's still a huge part of my day-to-day life really i mean i can touch several several radios with my hand right now that are all off at the moment for the podcast but usually on and talking to local repeaters oh nice and i look at amateur radio and and there are some people doing phenomenally cool things with software defined radio uh in the amateur radio realm but most people are just getting started they've got a you know an rtl sdr or some other really small cheap receive only almost invariably software defined radio and they're looking at a waterfall they're looking at a representation of the spectrum over time yeah and they're saying oh look there's a signal here i'm gonna go reach for my you know regular receiver and i'm gonna go tune over to it and i'm gonna talk and that's they have pre-baked software that basically hands them that much right right it's

Derek Kozel: basically just replicating formerly expensive software expensive hardware solutions right for like looking at a waterfall in hardware um or even looking at just a spectrum in general right

Chris Gammell: yeah yeah they don't know where to go from there because where should they go derek i'm so glad you asked chris uh they should go to grin new radio um which is a software project based under the uh i i shouldn't i shouldn't dive into this but it's an open source uh software project that has been running for gosh probably at least 15 years now i think long enough to where they decided

Derek Kozel: to name it gnu anything yeah yeah i think if you these days if you started something you'd be like

Chris Gammell: you know we should really call this gnu whatever and um yeah and so it got that you know enviable name gnu radio and what it is is a toolkit it's a framework and it provides a lot of the building blocks of digital signal processing so people like me uh end up writing drivers for the various radios and that allows you to go from you know a signal on the air into your computer and all of a sudden now you just have a whole bunch of numbers that represent uh basically the analog voltage on the antenna what do you do with this you're not going to reach immediately for your you know local copy of

Derek Kozel: digital signal processing you're right textbook you shove it into excel and then what oh goodness

Chris Gammell: oh goodness people have done this oh no people have done this um that is not the recommended approach

Derek Kozel: yeah um so these are streams of data that's what people should realize too right it's like you're not it's it just keeps coming you know it's not a it's not a fixed thing

Chris Gammell: it really does and depending on how much bandwidth you're looking at uh how wide your signal is that can be a phenomenal amount of data um we i was having a um i was at the a conference two weekends ago fosdom which is the free and open source denmark conference uh where i ran into somebody really fun uh me and derek were hanging out there the whole time yeah that was fun it was great and one of the comments that got made there uh and i i can't actually vouch for the factuality of this but i i don't think it's that wrong is that the cern large hadron particle uh accelerator actually only produces something like 40 gigabits of data per second yeah and and we around the table kind of had to sit there and laugh a little bit because at the moment on my desk i have two radios each with two 10 gigabit ethernet connections attached to them going to my desktop which means i'm generating as much data on my desktop from these software defined radios

Derek Kozel: right as the entire large hadron collider i know you that blew my mind and and like and you said it was because because it's been such a long-term project though too right that the the the hardware that they have assuming this is you know a correct statement as well right we should say that it may have been updated since then but but even you know it sounds plausible enough right they're they're really focused on the front end there right because they have to be super sensitive to whatever these interactions are happening in the collider but then it's streaming back a lot of data but not as much

Chris Gammell: as you're saying here right i mean it's still a phenomenal amount of data there's no getting around that and and to kind of loop back around to the new radio thing the question is what do you do with that data um yeah and if you just want to look at you know a time domain plot like an oscilloscope trace of it that's pretty straightforward and simple if you want to look at you know apply a fourier transform and look at the frequency domain representation of it that's computationally difficult but a lot of time has been spent on making that a very fast operation and so you can pretty much do that to you know several hundred uh million samples per second especially if you start throwing a graphics card at it and stuff like that you can end up with really interesting plots and people have written wonderful open source software to do this and that's all fine and dandy but i'm sure the hadron collider is doing quite a bit more data processing and they must be throwing a huge computational resources at that and then they want to save it right right yep i'm not saving this data anywhere right you're just streaming it through and throwing it out throwing it out yeah yeah um and so yeah so i i stream this data into gnu radio which is this framework and i end up with just a radio source block and it's the source of samples so a whole bunch of numbers are coming streaming out of this and then i can basically drag and drop in the new radio companion graphical environment blocks that represent the data processing steps and so this could be something as simple as an addition operation i'm going to add two signals together or it could be something really complicated like receive the gsm cell phone signal be careful on that one i think there's other restrictions on that i i'm assuming of course that you're running your own cell phone tower of course using something like the open bts project um right which is a as a minor diversion i i was fortunate enough to spend uh eight months working at range networks which is the company that that was developing that software oh no that was my introduction to software-defined radio um was at that company and yeah um you can run your own cell tower using open source software what way cool yeah yeah that is cool um so that was range networks you said yeah

Derek Kozel: yeah range networks cool um okay so the thing that i like about this is that so you were talking about you know if we're just taking a high level view again right so you have a radio say it's an rtl sdr right that's that little computer chip or that uh it's a tv tuner that got you know basically they started pulling the firmware out of it to make it more accessible but it had a debug mode yeah which was great yeah that's awesome so assuming that's the that's the thing so people can visualize this thing someone wrote a driver for that and if people go and listen to that show where we talked to the beaglebone guys a couple uh episodes ago that's probably a good reference for what that means right now it's a data stream that goes in your computer the thing i like about it is you really can't break things you know what i mean like if i was going to go and open up one of those boat anchors you were talking about earlier right so say you have a radio you open it up assuming you don't shock yourself on a high voltage plate or anything like that if you wanted to change the operation of it aside from what the dot knobs are on the front you had to actually go and like change wires around in this case it seems like you can just do stuff to it and yeah the data might go out of range but it it won't it won't fundamentally break anything yeah that's completely correct you just

Chris Gammell: like play around and mess stuff up right yeah and it makes for a great environment to learn in because your signal of course doesn't actually even have to come from hardware uh there are you know virtual signal blocks uh where you can say i want a sine wave that is at this frequency this amplitude and and throw that into your into your flow graph which is this uh total graph that you're

Derek Kozel: producing with all these connections and blocks yeah and um are there like internet sources of just like data streams like that too where you just like it's already tied in so if you had like a radio tower a local repeater that pumps it out that can be pulled into a new radio so yes and no there's nothing

Chris Gammell: stopping this and there are there is the ability to receive a signal stream uh over the internet you know via tcp or udp but do i actually know of any public sources that are streaming iq data that the uh fundamental representation of the the signal i i there are definitely people doing it but i don't know of anybody who's doing it into gnu radio uh as an open source project i guess it would be a lot of data too huh raw data so if you look at the rtl for instance the maximum sample rate you're pretty much running those out is 2.4 to 2.8 ish mega samples per second and so that's 2.8 million samples per second and the front end the adcs on the um rtlsdr are only eight bits and so every sample ends up being uh a pair of two bytes uh which represent the signal at that point and i'm going to completely gloss over the fact that this is a complex mathematical signal uh because that's that's math that anybody who is interested join the new radio mailing list and send me an email i will happily you know point you toward towards resources and you always have a list of links yes i think and so i keep in track so i can throw you a few um all right great nice introductory resources to complex signals so we're ending up with 2.8 million sets of two bytes so right so five megabytes per second yeah right which isn't that bad i mean it's depends on your connection to be honest sure sure yeah

Derek Kozel: especially you're streaming it right because like you know it's streaming is like yeah okay udp you might

Chris Gammell: get the stream but you might drop half the stuff so yeah and it depends on what your signal is how much

Derek Kozel: that ends up mattering right right yeah you start doing yeah you drop some of the stuff in the middle and your frequency still your frequency plots go all haywire because now you're missing a bunch of the

Chris Gammell: data so yeah so there's there are there are definitely some other sdrs out there where they have really nice streaming services um but yeah bandwidth is definitely a consideration there's a lot you can do to try and optimize it a lot of the time what you'll end up seeing is like if you go to these web sdr applications many of them will show you a graphical plot of what the spectrum looks like being updated 10 times a second 30 times a second something like that that's not very much data at all and then you can select you know like a three kilohertz section of that to to demodulate into an audio um signal and then they'll send you the audio signal or some of them will send you the iq samples that represent just that little portion of the spectrum and now you're talking 600 kilobits or 600 kilobytes a second and that's much more tractable for a streaming service right well i think the main thing to think

Derek Kozel: too is that i mean we're at a stage where like these things are cheap i mean like the the rtl sdrs are 20 bucks hacker fs are 300 bucks the lime sdr the new ones are what 100 bucks i mean like so people would just get those too right i mean like just as simple to just buy one yeah and and i if

Chris Gammell: you take one thing away from listening to this podcast is you should absolutely go and buy an rtl sdr they are available for under five bucks five bucks oh i thought they were 20 okay i mean so the the really nice ones uh that i enjoy which have the metal case an sma connector a stable reference oscillator um those are 15 maybe 20 but if if you go on ebay yeah five bucks

Derek Kozel: wow okay i think people can stomach 20 but if not five bucks great yeah okay so like pretty much no

Chris Gammell: excuse at that point there really is no excuse and one of the other great things is you can hand out like a dozen of these to a classroom and and that's a really that's totally affordable you know you can just factor that into the cost of running a course and now people are able to look at real signals on on the air and they're able to sit there with this free software which works on windows mac linux and they can just plug and play through you know what they wanted what they want to do with the signal you know it's not did you hear my sigh did you hear my sigh does it work on all operating systems uh so which operating system do you run and how much of a pain was it to install right i know

Derek Kozel: you hear this crap all the time so i had trouble with my mac um but i will i will fast forward through all of the pain that i had with it and say uh people should just there is a binary actually for mac

Chris Gammell: and just do that um so software stuff uh yeah and uh we we can talk more about gnu radio's future a little later on okay that sounds good yeah it's a bright and wonderful future got it um okay so

Derek Kozel: let's just let's gloss over that for the moment and say installation goes swimmingly uh everything's fine and now what so now someone has a srti sdr and they're plugged into their usb port they have

Chris Gammell: gnu radio open what are they doing then so almost certainly the first thing that anybody wants to do and will do is uh just put the output of your source block straight into a frequency sync which is going to give you this spectrum plot the traditional waterfall and there you know you'll type in the specific values for like you know you want to go look at 2.4 gigahertz the ism band there or 433 and look at the the band there and all of a sudden you start seeing your thermostat is transmitting you're doing this on a laptop near your car and you realize pressure monitors there's tire pressure

Derek Kozel: monitors which are fantastic and all have a unique id yeah yeah um jared from the uh the heck rf and the what do they call that thing the the plug the hat on top of it um port-a-pack yes uh yeah he he shows that when he walks around he's just like hey new car driving by it's creepy um but which is really

Chris Gammell: interesting i mean these are not things that people think about you don't think about all the things that are around you that are transmitting yep because it's not something you can't reach out and touch it but now for five dollars and free software you can be staring at it and so you see this little squiggly line go by and you're like ah there was a signal something was here something was here right now the question really does end up being how do you go from there to something meaningful and and that is i will readily admit a a non-trivial problem if you just want to receive like an analog broadcast fm signal you know 104.3 you can do that and that's pretty straightforward and there are tutorials on the gnu radio website there's tutorials located other you know elsewhere on the internet and i've been using uh osman's hacker f1 the sdr1 i was just about to say his his video series and and uh set of notes are probably about as good as it gets at the moment um there are very few things very few resources elsewhere that will take you step by step through the process of there's now now you have numbers how do you get to actually receiving something right yeah no and and i mean like and so

Derek Kozel: i mean we're kind of dancing around the education thing too but this is like when i watched that video i was like this this should be required viewing in every signals class anywhere this is like exactly what i wanted for getting started visualizing it making it contextual about like why why the hell do you even care about the math before you learn the math right and like it's perfect because it it shows a lot of that those those dsp principles that are baked right into the gnu radio too so i think that's and a lot of the tutorials do that it's just this my the first one i watched was mike's so um and people can go listen to mike osman on this show i mean god he's been on here what six seven times now so uh yeah if they want to they can get a lot of osman so uh so yeah i mean i think it's just a great introduction to concepts you know and and practical input or practical applications of them you know yeah and part of what i'm sorry one more thing about that tutorial specifically uh the thing that mike doesn't show and i think is the that i actually had to google for and nobody else talked about it when you're connecting blocks in gnu radio you don't click and drag you click once on the output you click once on the input and just draws an arrow so if you're trying it and you get frustrated you heard it here first folks yeah yeah and that's great feedback um because once you start using

Chris Gammell: software and i've been using it for three plus years now yeah you don't think about that

Derek Kozel: yeah yeah it's just beginner's mind stuff you know like it's just like you gotta you gotta

Chris Gammell: you know yeah you know you're playing the obvious stuff absolutely absolutely and what i did quite recently back in november i believe the very end of november was going to the local university here cardiff university and do a two and a half hour tutorial and we were using gnu radio we set it up and we weren't using any hardware for the first most of it and we were sitting there with all of these built-in uh the basic lego blocks you know let's make a sine wave let's add it to something else let's look at it in the time domain let's look at it in the frequency domain let's output it to a speaker and we were able to go from there's a thing called a sine wave you can sample it what does sampling mean you know you're pulling out regularly spaced time samples from it we're going to you know plot it on an oscilloscope trace on a time domain we're going to now add two of these together you can look at it on a frequency plot and see there's two spikes you know there's power at two different energy at two different frequencies and we just walked our way through that fm modulation am modulation we didn't really get around a phase two and a half hours is only so long but right especially if you're starting

Derek Kozel: from scratch without like the the other knowledge of the of the uh you know the background right yeah

Chris Gammell: and i was doing this much more as a introduction to gnu radio than an introduction to signals and systems these students for the most part had already had at least an introductory course but the feedback that i got afterwards was that being able to sit there with little sliders and change the amplitude of you know the am uh carrier signal the amplitude of the am modulating signal they were exploring the concepts of like depth of modulation um the deviation of an fm signal you know what how much spectrum are you occupying what's your occupied bandwidth these were concepts that they knew they understood and they could express in a mathematical realm but they hadn't been able to sit there and hands-on see all of this in such a a

Derek Kozel: visual and intuitive fashion right and and the manipulation of it too i feel like this is kind of this is kind of like the equivalent like arduino blinky thing right where you're changing the frequency of it you're like i did this thing it just draws you in because it's like i'm actually interacting with the system instead of just

Chris Gammell: viewing it right it's a doing versus watching yes and that's something that i would like to see happen a lot more and you see some classes use you know software like matlab where you are in control of this and you are looking at plots yeah right this this isn't an it's not a continuous experience it's a right edit compile execute loop you know which which programmers will be really familiar with you're not but that that'll never beat the ability to just change a slider and watch the signal change

Derek Kozel: in a streaming architecture right well and the fundamental stuff like i remember doing the matlab stuff where you're like plotting like you can do the the dots that are like in digital signal processing we actually have the the dots uh plotted and all those things too and it's like all right yeah i guess it changes some things but it's still very tied to the math versus like what you're talking about which is tied to the the real world signal and that's the important thing in my mind yeah yeah and of course

Chris Gammell: here's where you start wandering into the realm of educators and they'll be the first people to say that people learn in different ways and a lot of the time you want people to learn in multiple ways and that that's how you get the best understanding you gain the intuitive understanding by you know building these blocks together and going oh okay i get this and then you go back and you work well in my mind you go back and you work your way through the math and you go i saw an addition block in my flow graph what does that actually mean what are we actually doing here what does how does this work

Derek Kozel: and then multiplying frequencies together why would i do that right like that that's a pretty important

Chris Gammell: thing you know like yeah and and on on a first basis you're just like i completely don't get that you know why would we multiply them what's the outcome and then you look at it and you go oh my signal is now at a much my same signal that i had at a low frequency is at a higher frequency and you're like oh that's cool um for all sorts of reasons that again you start running into textbooks and and

Derek Kozel: stuff about right propagation and stuff right i feel like some of this stuff is even like naming things like i know it sounds really dumb but like when i was when i was getting started bandwidth was a word right and when you are in an sdr thing and you're actually adjusting bandwidth you're like oh the width and you're like you're like closing in this window and opening it smaller and wider and you're like oh like literally band width like this is a band and i'm changing the width like it's so dumb but like that is such a important concept that that at the beginning it's just a word you know and so tying it to that that concept is really important in my mind yeah yeah uh you know i am selecting a

Chris Gammell: range of frequencies this is the band that i'm looking at right it goes from you know my little side here to the side here and i can see that if i'm side bands yeah oh i mean you know just visually looking at the signal um is so powerful and being able to then do opera so one of the cool things that you can do in gnu radio is you can have multiple uh signals plotted on the same plot or in in multiple plots and so you can look at your signal at different stages of the the demodulation or the process you know the digital signal processing process here's your input here's after you add them together here's what happens when you filter and you can sit there and select them and go oh if i slide this one it changes here but now i'm filtering too much and so some of my signal is getting lost right i need to adjust my filters my bandwidth is wrong or you know i'm getting great sound but my signal is five times wider than it was before now i understand why fm you know audio sounds good but not fantastic because there's only so much space to fit all bunch of channels in so you start understanding the trade-offs that are made every day in in real world radio systems yep yep i know it's not

Derek Kozel: str in this case but i always like that um that um animated gif that ben krasna made too of like the record have you seen that slow motion oh the stop motion gif thing yeah it shows like a needle in a track you're like holy crap that's an audio track you know like it's like oh there's like the signal right there you know yeah yeah um so cool okay so i mean i mean and i think that like fm radio every like well i mean say everyone grew up with that but we got some young listeners now uh i mean like radio is pretty pretty you know a lot of people still understand that it's a thing that comes over the airwaves and it you know there it turns into audio and that's how you receive it similarly more and more people especially listeners this program they understand the you know the bluetooth and the lauras and the you know the 802.15.4s of the world or you know they're operating these bands that are you know there's signals there so how do people trans uh not transfer but like how do how do you move up the the you know the in the levels how do you level up in sdr world sure um and here's where i

Chris Gammell: wish we were doing a video blog because you would have seen me very excitedly stabbing the air with my finger here because this is absolutely the question how do you get from you know these basic analog fundamentals up to a real world a real world system and yeah everything's gone digital these days

Derek Kozel: pretty much and and maybe what does that mean too like so so we're talking about everything being in sdr so it is digital in that way but you're talking about what but the actual the actual

Chris Gammell: i mean okay here's where we can you know sit back and get very relaxed and say oh but everything is analog on the air analog man no no no no hey no no it's quantum man you know it's all it's all digital it's quantum let's let's pull that back a little bit and try and try and be practical right let's trim the neck beard when yeah everything over the air is analog you know you're you're looking at a continuous a continuum of values you know your signal could be at any voltage at any minute and inevitably it'll end up being all the voltages all the voltages eventually um but in a digital signal you are quantizing that you're you're limiting the states which you're going to say this particular voltage at this moment you know this representation of the signal this is going to be a symbol and i am going to say that this symbol means a binary one and this one means a binary zero and so the most simple version of this is on off keying which is a form of amplitude modulation um and it's is there a signal or not and if there is it's a one or sometimes a zero um interestingly sorry side tangent but i love this one uh old analog tv signals uh it was actually you would see white on the screen when the signal was high not low because any noise that you get is always going to add power and so if you add power maybe i've gotten that backwards but if you added noise power it would go black um so i i did get that backwards when i first said that sorry so that a a lot of energy would actually be a black signal it was inverted from what you might expect and that was because it's a lot better to look at an image that has random black spots than random white spots right um so yeah sometimes these things end up float but anyways on off keying amplitude shifting um yeah on off keying is kind of like what uh

Derek Kozel: i mean that's that is what well speaking of tv that's what a remote is right so like you key the um the infrared led goes just on and off and then you do that at certain frequencies right

Chris Gammell: absolutely yeah but at a certain at a certain symbol period right so 38 kilohertz whatever the frequency is yeah the you know you're hearkening back we call it the baud rate um symbol period same thing uh or symbol symbol frequency same thing and yeah and of course this is you know this is not just your tv remote it's all you know pretty much all garage door openers older than a few years ago and probably still a bunch of them i are all using this because it's really easy right exactly cheap and what do you know really cheap yeah it's really easy for a receiver to tell is there a signal or not yeah right um and so it's very cheap very easy and it's a great place to start um you know if you want to take a real world signal and try and decode it an on off keyed signal is really easy another really easy one is frequency shift keying um it's like fsk frequency shift keying two is just your signal is bouncing between two carrier frequencies um and so if you look at things like the amateur radio aprs network um uses a really simple form of of frequency shift keying they've got some subtleties in it about how they exactly implement it but fundamentally you're

Derek Kozel: bouncing between two tones oh interesting okay so and what what kind of like uh what what is the range

Chris Gammell: of tone uh what frequencies roughly are those at so they use uh 1200 hertz and 2200 hertz uh if my memory serves i so it's pretty low low frequency relative huh well so uh those are the two tones referenced from a carrier frequency um so the the baseband okay the baseband tones um because they wanted to be able to support it over um standard fm broadcast uh radios basically okay so let's

Derek Kozel: explain that in real in real terms here too right so if people are looking at if it's a one megahertz carrier right then it would be the the signal would be bouncing between 1.0012 and 1.0012 and 1.0022 well not point but times 10 to the 6th right so like that's that's the idea though it's it's these big high frequency signals but then they they shift a little bit in the actual um changes right yeah and

Chris Gammell: and so the entire reason and and so this is a great great learning opportunity um the entire reason you do that is so that you can put your particular signal using this modulation mode uh wherever you want in the spectrum so i can be transmitting it at you know uh 444 megahertz the next person can be 444 0.1 as their carrier frequency right we do the bandwidth of this is very small right it's only

Derek Kozel: like what three kilohertz total right you don't need a ton of bandwidth yeah and i mean technically

Chris Gammell: if you're using an sdr you can pull this out in in even less than that because the two tones are actually only spaced one kilohertz apart right yeah but yeah it's it's super small super simple and they run it super slow uh 1200 baud and and so yeah so that's the other that's the other thing

Derek Kozel: that always caught me up too right so it's switching between these two tones but how fast you switch

Chris Gammell: between them is then the baud rate like you're talking about right yeah and so the faster you do that the more data you can carry because you're transmitting more symbols per second uh but it also means that you are sending less total energy per symbol uh if you're transmitting at the same power so if you're transmitting at five watts and all of a sudden you and your symbol is you know one second long you're sending one watt for one second we could convert that into joules but hey i'm not even gonna try uh you send it at two symbols per second you're transmitting only half the amount of energy it's a lot easier for noise to interfere with that your your signal isn't going to go as far um under the same conditions because noise will eventually swamp it a lot earlier yeah

Derek Kozel: right right and noise is just you know uh it's just randomness basically entering into the system right and your neighbor's toaster

Chris Gammell: yeah um so yeah you you get noise everywhere it's always going to degrade your signal and it usually ends up being your signal to noise ratio how high your signal power is compared to the noise floor tends to be the defining aspect of how far your signal is receivable and do interpretable do you know you can do a valid demodulation on it right and so these are the

Derek Kozel: things where oh crap i forget who we're talking to about this the other day but we were talking about um so if you start adding in so these so this is a digital signal right so this is you know there's a zero that something that represents zero something represents one in these cases right and then how often you're changing it is the baud rate and then you could start to make it even crazier right where you start adding in codes that then have to so you only it's not now is it a zero or a one but is the string of zero and ones in order that's your actual zero right and then there's a one

Chris Gammell: yeah so you're adding an error correction error correction yes yeah and so there's tons and tons of different ways to do this and looping back around can your radio has an entire library of functions that implement this so you can drop one of these once you've gotten the signal demodulated to the point where you're getting a string of ones and zeros highs and lows or frequency one frequency two yep uh you can drop down one of these blocks and say uh i know that they were doing uh an 8 10 forward error correction with this algorithm with these settings because you read the manual you made the transmitter whoa whoa we don't do that around here um somebody told you on a forum somewhere uh you're just guessing widely yeah yeah oh yeah or or you've run um i forget if if uh mike osman's code did this or just check summing um but he was doing intelligent brute forcing i think it was i think it was intelligent it was at least brute forcing of the check summing and it would return out your check check summing algorithm is running on the first 24 bits producing a four bit value with you know this particular polynomial as it's as its base mathematical right input um

Derek Kozel: and that's how you could check to see if it's a valid code matchup right yeah whether you had any

Chris Gammell: bit errors in there right you know to see if the check check sum matches yeah uh and a forward error correction is basically it's not a fancy form of a check sum but it supplies enough additional information that you can correct some number of errors uh right yeah right and this is i mean you can

Derek Kozel: keep layering this stuff too right so like at this point we're talking about okay you're turning this thing on and off right and you know you see these two frequencies switching back and forth and now maybe you have a 10 bit or 24 bit string like you mentioned that it has to match again to be a one right so now that's a one right you're sending 24 characters by turning this thing on and off on and off on and off 24 characters now you say yeah i'm pretty sure that's a one but now you have to build if you actually want to transmit text if you're going to do that sort of thing you have to send eight of those right so eight times 24 whatever that is uh 100 something 96 i don't know uh uh and 186 sorry bad math uh so that would be then a single character right and then you could apply encryption on top of the characters too right you could have an encrypted message that you also then or maybe you're not supposed to send encrypted but you could also encode so that you're sending these characters but they represent some other characters right yeah and i'll throw the

Chris Gammell: i'm not a lawyer but i in amateur radio yeah yeah no encryption right you're not supposed to on the ism bands yeah sure i don't think there's any limitation there um well actually i'm quite positive there isn't because bluetooth ends up encrypted all the time and that's not licensed

Derek Kozel: right i'm just trying to get like the stack up of all these things because i think it's also i mean so like i think is what like from my personal experience it's been like so so why would i do this versus just buying a radio buying a key and then you know doing morse code right and it's like okay all those things we just mentioned try doing that with a morse code key so i saw i saw a great i i

Chris Gammell: finally started using twitter um about a year ago how dare you and i follow a very limited number of people and they are all talking about um they are all talking about radio so maybe that means i have a very small world view but uh at least on twitter now you just know what you like you know that that's oh i do like it and your filter is set very strong stringently somebody somebody was posting about the recent spacex launch and how they were streaming back video using it was either five or ten watt transmitter from you know five times the orbital distance of the moon and getting great video uh-huh and they were sitting here going man i wish that amateur radio could do this you know between two points five miles apart on the surface of the planet why can't we do this what are what are we missing and you know i i sit and bite my tongue a little bit because the answer is we're missing 40 years worth of communications theory and and and then i laugh a little bit because we're not even you know it's just uh what's uh william gibson's phrase the future is here it's just not evenly distributed exactly yep exactly um there are absolutely people in the amateur radio realm who are using digital video broadcast dvb which is a published standard for digital video broadcasting unsurprisingly and they're using this right now to send a gorgeous video from point a to point b right your new radio has full implementation other people have chosen to do it with hardware chips that are doing the the modulation and demodulation all inside their own package uh-huh um so the only limitation is really

Derek Kozel: that people are choosing not to dive into that piece of it right that there's there there is there are resources but it's just where you know they got to go find them right they have to find it they have

Chris Gammell: to be able to run it they have to understand it the examples of how to you know there aren't a lot of tutorials that show you how to do a end end setup with you know a low cost affordable transmit capable uh software defined radio to be able to do this and and part of that is that transmit capable low cost uh sdrs were totally non-existent 15 years ago right exactly yep right so just catching up to the technology right now as well yeah and and to to you know i i love this i love this about my my you know small universe that i exist in it keeps looping back around matt eddis when he founded eddis research was staring at this problem and saying i'm gonna fix this because nobody else seems to uh that's me putting a little bit of words into his mouth but he's gonna he was gonna go and address this problem right and he was gonna do this by building open source hardware with open source interfaces targeted at putting you know what at the time was really fundamentally good tools into the hands of radio amateurs whether it was actually an amateur radio operator or just somebody who was interested in this and wanted to let operate an unlicensed spectrum and so he created the universal radio hardware peripheral uh peripheral uh usrp so i've butchered it somewhere in there um i think you've

Derek Kozel: missed the s i probably missed the s uh radio system sure sure we'll say system we never we never use

Chris Gammell: the full phrase uh marketing usrp is very convenient yeah marketing is currently staring at me um so yeah he he made the usrp1 and put it out there and it immediately started to fall into the hands of really interesting people like ballant sieber for instance who you know was getting this finding software plugging together systems and receiving you know uh location beaconing from airplanes and stuff like that and this was just this wasn't possible before and now it was and now that you had this standardized interface you could start writing more and more complicated software and you know that was one of the fueling things a in the early gnu radio was this hardware started to become available it all started adding together and you had right place right time for a lot of this stuff and a lot of hard work um from the the software contributors yeah yeah yeah and we should i mean we should mention

Derek Kozel: too that so like so we've been talking about um we've been talking about gnu radio which is python based but then what we're really talking about and all this visualization is actually gnu gnu radio companion which basically has these blocks which instantiates python code which then talks down to the drivers which then talks to the radio right so it is it is software underneath but most what most

Chris Gammell: people interact with at first is the companion yes and so the companion is written in python and the majority of gnu radio blocks have a python layer on them you know user interface api on them but most of them uh actually i think this was a hard and fast rule invariably all of the core blocks are written in c++ oh okay cool um we need the speed and and certainly python is getting a lot faster right it's actually scary and impressive how fast python has gotten but um the ability i mean we are talking about potentially massive sample rates here in some of the more commercial or high-end research systems and the c++ gives you that that additional access to the to the architecture that lets you

Derek Kozel: really optimize things down um yeah i mean it's it's abstraction in general right yeah like you can yeah you get it you get access to the registers you can do all that stuff but then yeah you want to have a slightly more friendly software on top and python python fits that bill i think yeah um the one the thing i wanted to highlight here too is that so you mentioned you know edis kind of starting the usrp and stuff like that but gnu radio is kind of making its way into the being of the

Chris Gammell: de facto as well as far as you told me so it is it is massively adopted and and i should take a step back and and properly don my hat now i a couple weeks ago the person who had been maintaining the project and and really doing a phenomenal job and a phenomenal amount of work on it jonathan corgan decided that it was it was time for him to take a step into a different role uh and so now he's he's a you know huge tech resource for the project but he said he wasn't going to do the maintainership portion anymore and so uh ben hilburn who was the the project organizer a project leader went out into the community and started tapping people and saying are you interested in becoming more involved you're it you're it yeah yeah that is definitely how it how it initially felt but um he came to me and said you know i'm putting together a group of people to be uh project officers i think is the the term he used you know very generically we want to put a group of people into a position where they have enough you know blessed autonomy and and authority to to go off and and drive parts of the project that they're interested in um so that's you know permission to you know contribute code in and and really push push things that you feel are important it's an open source project we get contributions from all sorts of people um many of them phenomenally valuable contributions but there haven't been that many people contributing to like the underlying architecture of new radio um usually when people make contributions there it's because their own software at work there are a huge number of businesses that run this um or they're a phd student or in a research lab and and they need some change they'll implement the change and and then uh go back to their day jobs and so there was a need to have more people contributing to the to the main core and have a you know a large enough group of people that all of these contributions could get reviewed and could get put in all the all the project management side of things and um so as a what are you doing as of two weeks ago i am one of those officers nice uh and i i haven't figured out where i think i can make the best contribution in this new role that's that's something i need to figure out i so people in other open source projects should

Derek Kozel: definitely ping derrick on twitter and be like i can help you oh we yeah that would be that would be

Chris Gammell: great um and actually i'm going to i'm going to out you chris here a little bit um because part of what makes the open source community is so great is that there is so much uh openness in them and the the ability to get people who have a lot of experience in an area um ask them for advice and and get it uh we were sitting at dinner at fosdom uh two weeks ago and martin brown who is the my manager and the guinea radio um community leader was talking to chris and it was all about how do we move ki cad forward in terms of expanding the community and what does that even mean right and and it seemed

Derek Kozel: like you guys remember last week i was talking about i want to talk to people about manufacturing that may have come directly from this conversation with martin and derek so yeah and and the kai cad uh developers as well so yeah uh i think that's a that's a big piece you know and this is my first time in a role like this um well it's tough too i mean like so there's just the normal human problems right so there's the human problems but then there's also the like the you know the general conflict in any you know inside a company even you have engineering versus product versus you know users versus support versus whoever right and it's like even in a company it's tough and now it's like we're not going to pay anyone uh the team is completely distributed um there's there's competing uh competing goals there's competing uh you know outside interests and it's like boy that is that is a recipe for disaster uh so yeah it's it's the open source ecosystem is

Chris Gammell: so varied and and does have so many interesting attributes to it i love it

Derek Kozel: unabashedly and if you go to fosdham too you see like interesting individuals and a lot of club

Chris Gammell: mate uh the crossover is very weird i have found a place here in cardiff which will buy me crates of club mate oh so weird i am so happy i still don't get it i still haven't gotten into it but then uh one

Derek Kozel: person i was talking to there they're like you know the difference is i can drink four cups of coffee and i'm a wreck for the day i can drink club mate even though it has a weird taste i can drink club mate and go for 24 hours and i was like all right well that's like more of a that's the best selling point i've heard for it so far because i think it tastes like water that's been sitting in an old boot for a while so and it's like if people don't know what it is it's like it's like a tea-based drink in a glass bottle with a weird logo on it and you see it all these hacker conferences and like people just slurp it down and but yeah you try it you're like uh this is like it's like

Chris Gammell: trying coffee for the first time it's it's gross i was just about to to point fingers and say i bet you can drink four cups of coffee though um yeah i already have today yeah yeah mate is a south american uh plant which has caffeine in it and does have a rather strong flavor um and yeah you brew it into a tea you carbonate it and uh so uh i can't resist um there's a group actually working on making an open recipe for club monte oh yuck yeah but um it's great all right so back to open source communities

Derek Kozel: which have a lot of this stuff in it um yeah it's it is an interesting i mean it's an interesting management problem in the general piece and then you know and then what martin was talking about too is like one of the pieces that we were talking about is kind of the outreach and getting people to use it and and it's interesting from from my viewing of gnu radio because well both of you are edis research employees right and you know it's this software that's deeply embedded in the you know in the fabric of your company basically even though right i mean like so you know let's put it all out there right national instruments is your owner they have software that ties into this too so it is like this shared ecosystem right it's like you can use uh lab view and stuff like that but you don't have to you can also use gnu radio you can use other stuff and it sounds like there's a good balance there

Chris Gammell: yeah yeah and and i should have been clearer at the beginning um edis research isn't a subsidiary it's it's been fully purchased years and years ago long before i joined and so we are we are absolutely an integrated part of national instruments and you know there is absolutely lab view support for our

Derek Kozel: radios and it's great and i love that you guys are like a hardware arm that yeah you can use either one right so like your your thing is like you're making radios that are acceptable and they have these apis that anyone can tie into including you know the company that owns you right i mean like

Chris Gammell: that you know that you are right yeah um and that's ended up being a huge part of the strength of edis's products is that there is the ability to integrate at so many different levels and we get people who one of the things we have is we have a framework where you can drop your own code into the fpga because all of our fpga code is open source nice and so there are people who need to get that that hardware acceleration to do some some digital signal processing and there you go you you can you can do this you can go into the fpga you can put your code there there are people who want to understand some fundamentally deep portion of the hardware you can go onto files.edis.com and pull up the schematic oh really nice yeah yeah um and so i i love the open source environment i love all that it gives you there in terms of flexibility and choice and yeah i'm really really happy to have joined the the new radio uh organization formally yeah and it's like one of those perfect you know

Derek Kozel: public private type of you know like partnerships open source obviously promoting promoting the products promoting the you know the state-of-the-art type stuff and it's i think it's really good yeah so and

Chris Gammell: i i'm very much trying to keep my new radio hat on here uh and if you look at the ecosystem in new radio there's rtlsdr support there's hacker f support there's a group in germany based i should say in germany uh osmocom and they contribute great software which integrates all sorts of radios uh and so you can get you know lime i think can you yeah i shouldn't i shouldn't say this entirely blade rf lime absolutely but i think lime uh put a huge amount of effort into their new radio integration and they made a huge presence at the annual new radio convention um last year and the year before and but they were demoing their hardware there they did their their release of the latest hardware um that that mini uh yeah yeah at the conference last year and it was fantastic it was really well received and it's it's great to see that the new radio ecosystem reaches out to so many companies

Derek Kozel: um and what is this conference you're talking about this is a new radio specific conference

Chris Gammell: sure yeah it is and it is now in its 11th year i believe i should know this given that just yesterday i pushed the uh update with the 2018 uh conference location and dates and which is uh which is september 17th to 20th 21st 21st uh in henderson nevada uh just a suburb of las vegas okay i was like what why would you do that um it's a great it's going to be a great venue uh we were looking at some other options the dates ended up not working out on those for this year but we're hoping to do them in in a future year but we've got a great venue what's been phenomenal to see is this conference started out and i'm not the right person to tell this because i haven't been part of it long enough but it started off i with just one room in a university i think um and a whole bunch of people got together and had a great time looking at software defined radio and new radio and how was all this going to come together and now we're at the point where the conference grew 40 percent last year uh we basically sold out i think if we sold like one more ticket we would have had to start turning people away at the door um and so we went from 250 people to 320 ish uh last year and we have room for 350 375 at least in this new venue nice okay the community is really big we get people from lots of different countries coming we get people from you know the government coming from like government labs from my commercial industry like big companies sending people there uh both to learn about the new radio and to to show the work that they're doing um we get researchers from a whole bunch of different institutions and uh countries coming and they show the work that they're doing in their phds in their masters in their courses in you know their post-grad research and i i love that conference um well good because you're you're kind of helping with it now yeah yeah yeah um so last year last year i was helping i was sitting in the room to organize it uh i did some hands-on stuff at the conference but i didn't i did very little beforehand um and this coming year uh i've taken on a lot bigger role and i'm definitely doing my best to make the conference happen and be as big and successful as it can be well so people listening

Derek Kozel: should they should they think about going i mean like so who who it would should uh someone who's just getting started go or isn't more about like being the ecosystem for a while and and be using it

Chris Gammell: first every year we get people coming who have never used gnu radio um okay which is always uh i i'm gonna go with humbling actually uh rather than impressive uh because it's it's not a free event and it's it's during the work week you know it's a professional level conference style right yeah i mean it's five days i we end up with one main track of talks but we we end up multi-track for at least part of the day on some of the days we have tutorials we have breakout sessions but it was never envisioned as a conference where you would be attending if you had never done this before right right and so it had to become that over time uh and and again i wasn't at the beginning but that that's how i look back on it and so the monday is always a new user day or or has been a new user day in the past we we haven't set the schedule for this next year um and so but that was in the past that's mostly been users who are newer to the software you know you've used it you've received a signal you've connected some blocks you've written some code for it um and you're looking at where do you go next uh but now we're getting a lot of people who come in and have never used it and so we absolutely want to embrace this part of the community and and run some more bare bones stuff i would definitely recommend that people come i think it is a great experience and i think that this year is going to be um a really good opportunity to do it uh travel and hotels to las vegas are really inexpensive that time of year and so because it's hot it's real hot we have air conditioning right right right um you can transmit from inside yep as an organizer i personally guarantee ice at the event got it um cool well let's so let's swing back to the actual radio stuff

Derek Kozel: real quick sure uh uh one of the things i wanted to cover too is like some of those crazy crazy applications you mentioned and i think you're allowed to mention but you you were showing me

Chris Gammell: some pictures are you allowed to talk about that application so i will i will limit what i say to what i have personally seen online um but yeah the the project that you're mentioning uh is ice cat and ice cat uh it's a european research project and if you go onto their website and look through their news reports you'll see that they have chosen for their initial prototype system to um they they did a big contract uh a big um outreach and they they ended up selecting national instruments as the provider of this this prototype system got it got it and um yeah edis edis delivered a customized product to them which was a large multi-channel receiver yeah and so what they're doing is they're they're doing a beam forming setup where their their main project goal was to um transmit and receive energy into the atmosphere and use that to study the composition of the atmosphere at any one time and they've been doing this for decades yeah that's crazy to me i didn't realize

Derek Kozel: this is a thing but it sounded like you were saying like it's like radar but for the clouds

Chris Gammell: yeah is that right yeah and and so something that you and and hopefully other listeners have encountered in the past or maybe heard of is um all the weather balloons that get launched sure yeah right a lot of those are um have radio signals on them and they're they're tracking you know the composition of the atmosphere at different altitudes and so it turns out that you can do many things generally different than what the weather balloons are doing um by by blasting radio waves up at it and the same principle as radar uh if you look at you know the signal at different times it'll represent the reflections at different altitudes and it all gets way more mathematical and way more

Derek Kozel: complicated but um right but that's the idea i think that's the other the other thing i want to get to too is that like so this is a very uh complicated system right the pictures you were able to show me like it it looked very complicated very high power all that stuff right but at the end of the day you're you're you're using math to basically you know selectively cancel and add signals together right yeah and and that's how you steer a waveform across these you know basically think of it like a laser right that you're making out of radio waves you know laser right a mazer or i don't even know what a razor uh uh and and basically you're scanning the sky using using these selectively you know canceled what radio waves and then you can do really useful things from that right yeah and so this is a

Chris Gammell: principle which we see used in like uh 4g and 5g cellular uh standards um we see it used on on wi-fi routers when you talk about a mimo system you have multiple in multiple out it's multiple multi-channel receivers uh and transmitters and you're using all of these different properties and some very clever and fancy math um but but fundamentally not all that hard to understand the basics of to steer the energy so that you can get a signal from a particular direction and that may be you know following somebody's cell phone to give them the best coverage or it might be you know doing a radar type thing where you're scanning energy around there was a there was a fantastic demo of this actually where somebody set up a grid an array of leds on a wall with a little little circuit and antenna and the led would light up depending on how much uh rf energy it was detecting on a certain frequency right

Derek Kozel: so it's got like a coil and a rectifier and then that creates a dc voltage that then lights up the led

Chris Gammell: is that the idea that's the idea i i don't know what their particular implementation was it may have just been a little radio a lot of them a lot of these single chip radios will give you um received signal strength indicator indicator rssi um which is just a metric of how much power they're seeing and these folks were using a a mimo beam formed array uh to steer the energy towards different parts of this wall and you could see the leds lighting up with the actual antenna beam pattern yeah and i will try and find that link and send it to you because it is just phenomenally cool

Derek Kozel: yeah that's really yeah that is i mean and and that's i mean that's where a lot of the stuff's going so like um i think that's what greg was doing so uh greg charvat who's been a guest on this show a lot in the past too he worked on that butterfly uh system i think that's what they're they're doing beamforming i think with ultrasound now on chip and that's like that whole new butterfly um

Chris Gammell: ultrasound system thingy so um i haven't heard of that one but i absolutely know that i mean that these are basic physics principles they apply to waves in general and so you can do it with the

Derek Kozel: is the hard part right i mean like yes these are basic physics things but like it feels like the the processing was always the hard part or the the thing that was missing that it is computationally

Chris Gammell: intensive i'll go with that rather than it's particularly complicated math to do it well yeah sure that the math gets more complicated but in in you know 30 seconds or less if you think about it if you have a wave coming from a single point of origin and you have a line of antennas that are you know spaced apart and if you're if your line is not actually you know 90 degrees perpendicular to the direction that the the wave is traveling towards you the signal is going to arrive at each antenna at a different time you know we we see this you know think of an acoustic wave like if somebody uh you know shoots a gun or something like that i you hear the sound depend yeah at different times

Derek Kozel: depending on how far away you are right you have the speed of sound playing into there right all that

Chris Gammell: stuff yeah which interestingly is uh 800 and something thousand times slower than the speed of light um i had to come up with that number for a beam forming example recently and um it yeah it arrives at each antenna at a different time so then if you take those signals that you've received or microphone is in this example right your receiver right um and so if you take that signal on each channel and you delay it by the amount that you think you would need in order to to move that array compensate for that change of time of arrival uh you can calculate out the basically the radio signal will be much stronger when you add them all together uh otherwise the energy gets smeared out over time um and so this is called time of arrival uh or in a radar system time of flight for measuring distances um and so you get this with like acoustic triangulation and stuff like that it's

Derek Kozel: it's all the same principle right um people have heard alicia from embedded fm talk about shot spotter that's the idea where they're trying to figure they have an array of microphones around a city and they're trying to figure out where where locate where a gunshot came from yeah u.s problems hashtag

Chris Gammell: you know i so there was a new radio implementation of shot spotter at this last conference okay um so yes

Derek Kozel: absolutely same same math right same math but but again i want to point out too that like so like my my thing is like thinking of you know people on uh morse keys right like thinking about that as the as the alternative right in the past to correlate you know 50 people to do you know keying keying up radios at different times that was the hard part right well not not the hard part

Chris Gammell: but the it took a lot of stuff to do that right it would have been hard to have ever even imagined

Derek Kozel: doing that as a viable application yeah and so and then and then 20 years ago it was possible but you had to do it with lots of computers and now it's just getting so cheap that you can do it with you know all of the hardware costs are coming down and so you can do it with relatively low cost computational hardware you can do it with relatively low cost front ends relatively low cost adc is all that stuff and basically you're just piling morse law is allowing you to pile more

Chris Gammell: hardware into the solution right yeah yeah absolutely um morse law and also just the changing the changing nature of the problems that we're trying to solve so we see a lot more gpus we see a lot more parallelism yeah and it turns out that for digital signal processing that is that is just a win uh that's absolutely a win and as we look forward into what we what are the problems we need to solve in the future it's how do we make the best use of uh all of this computational power that we have access to um gnu radio uses a library called uh volk uh vector library of optimized kernels which is a basically hand tailored set of code uh where you know you'll say we've got some problem we want to multiply two numbers together but we've got 512 of them in an array and we want to do this to each set of numbers in this array what's the fastest way to do this it ends up depending on what computer you're running on because the newer the processor the more more instructions that you have to do this sort of problem fast um and so this library gives you the ability to say here's our problem go do it as fast as possible and then depending on what machine you're on it'll hand you hand optimized code um to do to do this problem fast that doesn't scale uh it's a fantastic body of work and it's a big part of why the new radio runs as fast as it does but now we have gpus we have fpgas we have custom asics you know custom digital signal processing chips which have been around for a long time and they're but they're more accessible though or what they're blazingly fast at one thing or one one set of things right that's the custom and and if we know that we have access to one of those because we know our you know whatever our problem is is going to involve that at one step how do we how do we end up with gnu radio or whatever framework you're using being able to make the decision to send the code to send the data there for processing without you having to hand select it every time yeah right because the problem is just huge and and right you know and that and that's what it comes down to too

Derek Kozel: i think like so people listen to this they're they're starting from scratch they don't know it and i think the one of the problems especially like because we're you know to loop it back to the education piece too is like education normally starts you know bottom up like let's talk about physics let's talk about all this stuff right we're talking a lot of the top down thing like let's have an application throw a bunch of computing at it assume the hardware works right you know maybe you buy a usrp or you buy hacker f or a lime or whatever but my opinions usually usually to quote uh you know silicon valley middle out is the best way right pick pick a application pick an application that you know is somewhat simple and it's it's you know it's easy to easy enough to rock and then and then figure out which way you want to go you know do you want to go down to the radio do you want to go up to the computing piece just find something that works to start with though yeah yeah uh so so what is your go-to so if you were going to say to someone they have some coding background maybe a little bit of hardware background they've gone through the uh fm am examples and stuff like that what project would you point them to to for a middle out solution that they should go and try and uh and

Chris Gammell: implement and and learn from i've in the end found that you have to care about the problem i okay because i mean i can point at something and i can say oh there's this really cool data mode that you can demodulate and you can work your way through and it's not that complicated who cares right right

Derek Kozel: you know that's true i'm mentioning at the end it's a cow that's standing in a field right it's

Chris Gammell: like track your cow and and you don't have a cow right i don't i don't have a cow you don't have a cow why do you why would you ever spend the time to learn how to do this um so you need to find your cow that's what we're trying to say here well in the end we might have to find the cat that's trying to break in my door i hear i hear the cat in the background um there's nothing that can be done about that uh but um so what what do i say i i say look at a lot of these simple data protocols and look around your house and say do i have an internet of things connected device um wouldn't it be interesting to look at it a lot of them are unfortunately encrypted so there's only so far you can take it one of the interesting things is if you look at the internet and you look at communications it's a stack you know you you have the basic representation of the signal on the air you can't encrypt that that that's not a meaningful thing you can do things like frequency hopping which make it harder to recover the bits and also have some really big advantages for for transmission you know you can you can make it very resilient against interferers and stuff like

Derek Kozel: that right right um lots of people in the channel messing with your stuff yeah and and so it's a

Chris Gammell: great things like that are a great place to look and say i want to be able to just recover the encrypted bit stream you know and so you can say every time i push this button some blob of rf energy comes out i want to to look at that you can go with that um you can buy your own radios if you have your own radio project or something you know that you're assembling and you're like i want to put an ism band radio in you know one of the texas instruments chips or one of one of whatever you know something which with an open standard that you're you're implementing i want to make a receiver for that you can do that um something that's ended up being very interesting actually to me is uh the number of satellites which exist which can be received with not that much extra work in terms of hardware uh for not that much expense i mean you can buy fairly cheap and you know pre-made antennas and low noise amplifier and a filter and just go at it uh weather satellites they're transmitting digital information it's extremely well understood yeah and and that's a great that's a great one yeah

Derek Kozel: everybody needs to know the weather come on man i was gonna say here is me pointing towards a

Chris Gammell: particular particular thing but i think the weather satellites actually are a great place to look um yeah no that's that's a that's a great one actually yeah yeah and there are projects that have the software already written um they like open satellite project i believe is the name and they have an implementation there are closed source implementations there are you know other ones uh ais is an airplane airplane data tracking system um it transmits more than just where the plane is but it's really cool to see that information just streaming by yeah it's an extremely well understood standard it is not currently encrypted um because it's not considered sensitive information

Derek Kozel: right and that's all i mean you could go on flight radar 24 or whatever that site is and like you can just look at it too but that's just someone's pulling that data from a database but you know that database is populated with it's all coming from somewhere you know well i was about to say a lot of that is

Chris Gammell: actually coming from people at home with their own receivers a lot of them rtl strs really okay yeah yeah um i've got i'm not going to fish around because the entire podcast will just end up being clanks of stuff falling to the ground but i've got a little filter right behind me uh that covers the ais band which is uh 1.09 gigahertz which is right in the band of of what the rtl can support it's a it was not expensive the antenna there are designs which are literally coat hangers that have just been bent in a specific way to make a dipole antenna right right it is not demanding and there are airplanes everywhere in the world these days um so so you know i'll circle back and hedge my bet and say pick something that you're interested in whether it's what you know receiving signals from outer space uh with the weather satellites and or you want to receive planes yeah i started kind of like so i

Derek Kozel: buckled down and bought a hack rf and decided to you know do the the new radio thing because uh because i'm playing around with laura stuff i mean cellular stuff too but like laura is the interface and i talked about it last week too but like the interface is like signal goes up signal goes down signal goes up it's just a chirp you know and it just that's and it's encoded you know it's encoded but it's

Chris Gammell: pretty simple you know so should be decodable yeah it definitely is and depending on on how you want to approach it there's a full there are actually two full full implementations of the modulation and demodulation in gnu radio yeah gr laura right gr laura and gr laura uh unhelpfully really yep that's a yeah that is lack of coordination there yeah okay but yeah one of them one of them will

Derek Kozel: break eventually and um oh one of his matt knight right one of them is matt knight from uh best deal

Chris Gammell: uh best deal yeah i get to meet him that works cool um yep i i don't know the last yeah it's a security firm right yeah it's a security firm based in san francisco um and and they they present every year at gnu radio conference uh they ran the hacking challenge we have an rf hacking challenge oh cool that's i think going to become an annual thing because it's been so popular right and they they kick-started that i think um that's great yeah you know the hacker stuff is

Derek Kozel: interesting too because it i mean it is like it's like what you're talking about it's like it's decoding what is out there right and then it's like what can you do with it i think you know hacking usually has a negative connotation but it's more about like you know how much can you get done with a limited set of time and hardware and everything and and see what you can come up with right and being you know applying a set of rules applying a set of you know decoding whatever's out there and just trying kind of almost fumbling around you know if i could be so bold to say about hacking uh and and and then just seeing what you can get done so i'm i'm looking at the clock and

Chris Gammell: i'm guessing that we're we're not too far from the end just in the interest of listeners being able to finish this you know at some point but um one of the things you have to go too and we could i i could

Derek Kozel: talk about this stuff all day but yeah this is uh you know yeah sorry what one of the things that i

Chris Gammell: feel very strongly about and that it's my guess that in some way my contributions to new radio will end up being based around this feeling is that for a long time the people who were using radios were had to do it themselves you know they had to build the radios themselves and even if you were not totally passionate about electronics design and just wanted to get on the air you had to build your own transmit and receive system and so people gained this really phenomenal and an admirable understanding of the systems that they were using because they were building them you know they were building their own morse code keys they were building their oscillators they were learning how to use capacitors to change the frequency of of a crystal oscillator um on how to build a filter how to build an amplifier all these pieces and and that part of the amateur radio hobby has absolutely stayed strong it's grown we see phenomenal work there it's you know 2018 software is here computers are here and and there are i don't know what you're talking about as i talked

Derek Kozel: you over open source uh you know voip software uh across the distance of thousands of miles the next

Chris Gammell: time we do this you're going to be on an amateur radio uh system oh it's always my fault okay and and and we'll have uh you know we'll have right exactly yeah but wait we won't because people have started using software and there's uh open source voice codecs that you can use to send digital voice over uh the amateur radio bands over any band but software has been implemented to transmit it over you know readily available uh you know boat boat anchor class traditional designed hardware um there's nothing stopping you from doing that over a software defined radio and now you're getting digital audio over the same frequencies that were used a hundred years ago to send morse code and basic am and then single sideband and so i see the hobby moving but they're most of the time it's people use it's a very very very small group of people writing software that then is distributed and used by hobbyists um by by amateur operators and what's missing is that these end amateur operators we're getting to the point where there is a need for good resources for them to understand sampling theory understand digital signal processing understand digital filters fourier transforms you know i'm not saying that everybody needs to become an electrical engineering you know computer science math genius whiz signal processing person but because that would be a that'd be a losing battle yeah i'm trying to pick an achievable goal here but i think that there should be an understanding of digital signal processing and of software defined radio and and of the tremendous impact and potential that it has for being an educational tool but also of of enhancing the hobby the way you get really clear video terrestrially or in space is you make it digital you add in the forward error correction you you you do all of these things that are well understood in digital communications they're available they're accessible they're available i should say they're not accessible and i see gnu radio as being a phenomenal tool for making it accessible uh and so yeah i i think that is that is your mission you've just you've just stated it okay how do i make this happen um tutorials guides better documentation it turns out that gnu radio has a huge amount of documentation that even after two years of me using it i haven't built the habit of accessing it because it's not right it's not right there in front of me it's not built into the graphical companion yeah that's the thing like

Derek Kozel: our expectations of ui and ux and all that stuff has changed too from the web and from everything like that and you know yeah you could find it if you if you if you dig for it right it's been there uh but yeah things change right it's like it's and it's like some people will dig in and say well we don't want to change to be like that and other people say you know well there are benefits right you do open up the community more if you make some of these changes so yeah it's just people stuff man and and so i i won't

Chris Gammell: spoil uh so the the new maintainer of the project marcus muller uh i know he has grand visions for how to make the so how to get the software into more people's hands more easily um and and he's going to do a lot of work on the packaging of it how you know making the install experience a lot smoother um and and drawing from the resource of all of our great maintainers in the different linux distributions um on windows on mac and and how to make that experience as good as possible for installation and for running it but i think the part where i might be able to contribute well is at the end user actually using the software how do we make them more effective and and have more questions on the mailing list that are i have this tricky digital signal processing problem you know how do we how do we find the solution to this and less of i can't figure out how to connect two blocks together right hey sometimes it's a problem okay which is which is totally fine and and right right you're

Derek Kozel: saying like you're saying have app you're saying have uh stuff so that that question isn't even a problem right either making it more intuitive having a guide that's out there all that stuff right yeah no i think this is just another human problem too it's just like you know until you get past these these initial issues then it like how do you move up to the next layer of abstraction is you make the

Chris Gammell: the the layer below it less of a hassle right yep you don't even think of the fact that your compiler has to generate assembly code because it just does that right well i mean to to put a you know analogy

Derek Kozel: with ki cad or key cad uh you know the um you know you could talk all day about a you know an automatic uh what's a push and shove router and stuff like that when but when people are like i hate how it copies and paste it's like okay well that whole conversation then is like moot right it's like you're not gonna you're not gonna pull more people in because of a feature if you if you're not usable so it is kind of like an evolution of of a project to be a little bit more modern because you already know that it's so useful right i mean you've seen it firsthand yeah yeah and so yeah reducing that

Chris Gammell: learning curve making it making it easier to use easier to learn um all that sort of better resources i think are are a really important part of growing the project to the next level and i mean the the fact that mike put out his tutorial is phenomenal and it's the type of thing that we we love seeing in the community um that's great to see and yeah it would be great to what do you know link to that from the website uh i think it is actually linked to from the wiki but you know it's not it's it's not all that it could be um the wiki and the website um and well this is this is a good lead

Derek Kozel: in too so that we can wrap up uh how do how do people get involved and uh in in the project and get

Chris Gammell: started so uh the new radio.org is the website um just head there and i believe the button is learn more or get started might be get started um and that that will drop you onto a page that has links out to the where to get the installers how to install it you scroll a bit farther down and there's a set of basic tutorials which apparently don't cover how to connect blocks but do connect do cover what blocks to connect to get various effects yeah and you don't yeah it says i'm seeing stuff

Derek Kozel: here for you know remote grc volk introduction like you mentioned that stuff so we'll have we'll have links to a lot of this stuff um you know some of the gooey stuff so i was gonna say you you

Chris Gammell: you can start there but uh yeah i think there is a a set of getting started to i know there are a set of getting started tutorials and they they do start at the level of how do we make a sine wave and how do we view it and right that is that is a good place to start because by the time you're looking at that so much has already gone right yeah right it is the getting the blinky of of the new radio world absolutely absolutely and and if you encounter and so then the follow-on from there is where do you go if you want to learn more and where do you go if you have a problem and we have a mailing list the discuss new radio mailing list um we'll make sure there's a link in the description it's also linked to from um that first getting started page that you land on uh there's a slack channel uh give you radio at slack.org uh to get an invite to that you go to slack at guniradio.org um i because that's how slack works it's also linked to an irc channel they're interchangeable of course um which uh which is extremely active great and i can i can proudly say that it's it's just not quite active enough for me not to be able to read every single message that comes through there so if oh i see if you post to the mailing list i will see your email uh as will basically all the other core developers and a community of thousands of people uh and and that's not to say that you shouldn't post basic introductory questions you absolutely should we love them and there is a there's a big and really wonderful community waiting to answer everything from how do you connect blocks to i am you know a company i work for a company that is operating satellite ground stations right you know i have these three things working and i'm now working on something that could be a phd project if i was still in university what do you think of my approach here we see the entire range and and that's just great cool how could people contact you twitter uh twitter uh at derrick kozell um is a great way i i haven't i am active on twitter uh but i'm not i don't do a lot of conversing there um irc it's a good place to get you i'm sure irc is a phenomenal place if you have if you want to have a conversation that's probably the best place i via email even i don't mind receiving emails uh so that's just uh derrick at bits to volts dot com nice and that's uh that is a good i was going to mention

Derek Kozel: your talk too so so you had done a talk at uh the meetup i used to do out in san francisco the hddg and so i will link that in as well um yeah that was a great talk you did so and it's another

Chris Gammell: introduction to gnu radio it is it is and it's the type of thing the introduction to gnu radio is something that i've been i've been giving talks at various venues i did one at emf camp back in 2011 or something like that 2013 um i've just given one at the university here i'm about to give one in march i keep giving it and it keeps getting a little bit better and a little more useful each time right i and so yeah i i hope that they prove useful all of my previous talks i always post the slides up on my website um and so people can we can include a link to that yeah yeah awesome well

Derek Kozel: derek thank you for joining us uh today we uh this has been great and i highly i do i mean i gave a little bit uh you know a little elbow for the uh the install but once you get it it's actually it is very friendly um and i think that it's uh i'm really excited for the project and to use it as a troubleshooting tool and you know it's just exploring the spectrum is really really interesting

Chris Gammell: um and accessible now i have really enjoyed chatting and and yeah i i sit here thinking of all the other things that we could talk about um but i think we covered the big ones and we can talk again in the future so that's great and and yeah please audience do come try out gnu radio i it works for audio as well if you're into audio processing it's it's a great project it's which is what has made me so passionate about it so hope to see a bunch of you there and yeah look forward to it okay thanks for being on definitely see you see you chris you