Hitting a little bit of a wall doing some sinusoid tracking and in an effort to work smarter not harder, I'd like to do some hanging out with someone experienced in MIR. Neural net knowledge a definite plus. This is a pretty fun project so hopefully it won't be too painful. Rates negotiable. Matt EstateSound Com
MIR Consultant needed in NYC
Started by ●September 8, 2016
Reply by ●September 8, 20162016-09-08
On Thu, 08 Sep 2016 09:23:43 -0700, MatthewA wrote:> Hitting a little bit of a wall doing some sinusoid tracking and in an > effort to work smarter not harder, I'd like to do some hanging out with > someone experienced in MIR. Neural net knowledge a definite plus. This > is a pretty fun project so hopefully it won't be too painful. > > Rates negotiable. > > Matt EstateSound ComOK. I was thinking I might possibly be your guy until you said "MIR", which rings no bells (unless you're talking about light with wavelengths around 5um, or Russian space stations). So -- what does MIR stand for in this context? -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!
Reply by ●September 8, 20162016-09-08
Reply by ●September 8, 20162016-09-08
On Thu, 08 Sep 2016 17:12:27 -0500, Tim Wescott <seemywebsite@myfooter.really> wrote:>On Thu, 08 Sep 2016 09:23:43 -0700, MatthewA wrote: > >> Hitting a little bit of a wall doing some sinusoid tracking and in an >> effort to work smarter not harder, I'd like to do some hanging out with >> someone experienced in MIR. Neural net knowledge a definite plus. This >> is a pretty fun project so hopefully it won't be too painful. >> >> Rates negotiable. >> >> Matt EstateSound Com > >OK. I was thinking I might possibly be your guy until you said "MIR", >which rings no bells (unless you're talking about light with wavelengths >around 5um, or Russian space stations). > >So -- what does MIR stand for in this context?I think he means Music Information Retrieval since it appears to touch on neural nets, but that's just a guess since he wasn't clear about it. Could be he's talking about the Russian space station or something similar, though. ;)
Reply by ●September 8, 20162016-09-08
> I think he means Music Information Retrieval since it appears to touch > on neural nets, but that's just a guess since he wasn't clear about > it.Yes exactly and this task should be cake for the DSP engineers on this forum since these sinusoids barely move.
Reply by ●September 8, 20162016-09-08
On 9/8/2016 6:54 PM, MatthewA wrote:> Heh! This music information retrieval! That is, I'm tracking sinusoids in the audible spectrum.You mean like Elmer Fudd? http://www.barbneal.com/wp-content/uploads/elmer01.mp3 Often when people come here asking for help, it's more like this... http://www.barbneal.com/wp-content/uploads/elmer03.mp3 -- Rick C
Reply by ●September 8, 20162016-09-08
On Thu, 8 Sep 2016 16:35:30 -0700 (PDT), MatthewA <matthewaudio@gmail.com> wrote:> >> I think he means Music Information Retrieval since it appears to touch >> on neural nets, but that's just a guess since he wasn't clear about >> it. > >Yes exactly and this task should be cake for the DSP engineers on this forum since these sinusoids barely move.Is it music or isolated sinusoids that barely move? It's getting even more obfuscated.
Reply by ●September 8, 20162016-09-08
> Is it music or isolated sinusoids that barely move?Both: https://soundcloud.com/rajendra-bhagwat/rajendra-heggar-tabla-practice I've got it pretty good so far but I think someone who knows what they're doing can identify these sinewaves much more accurately. I'm looking for the first 6 overtones in the Dayan, (the higher drum). That means ignoring the bayan (bass drum) and filtering out the onsets. As I said, I've done an Okay job already but it can be better.
Reply by ●September 9, 20162016-09-09
On Thursday, September 8, 2016 at 7:35:33 PM UTC-4, MatthewA wrote:> > I think he means Music Information Retrieval since it appears to touch > > on neural nets, but that's just a guess since he wasn't clear about > > it. > > Yes exactly and this task should be cake for the DSP engineers on this forum since these sinusoids barely move.MIR is no piece of cake. note onset detection is hard, especially if the note is not percussive. and note release is even harder. since you're NYC, perhaps this person might interest you: http://www.nyu.edu/classes/bello/MIR_files/1-Introduction.pdf . otherwise, most of the hardcore MIR folks are at the Left Coast or in Montreal (one of the few European cites on the North American continent). rots-o-ruck. r b-j
Reply by ●September 9, 20162016-09-09
On Thu, 08 Sep 2016 16:35:30 -0700, MatthewA wrote:>> I think he means Music Information Retrieval since it appears to touch >> on neural nets, but that's just a guess since he wasn't clear about it. > > Yes exactly and this task should be cake for the DSP engineers on this > forum since these sinusoids barely move.Eh, you mean the many, many sinusoids that are all piled on top of each other and interspersed with just plain noise from things like percussion instruments? Unless you're just looking at flute solos, that is. It doesn't sound easy to me (see, I wasn't your guy!). What's your end goal? To write sheet music from a recording or some such? -- Tim Wescott Control systems, embedded software and circuit design I'm looking for work! See my website if you're interested http://www.wescottdesign.com