On 09/30/10 09:40, Patrick Shirkey wrote: Are you sure? Where did you get that info from? Just skimming over Stretch.cpp gives me the impression that it's based on Fourier analysis and re-sythesis. http://en.wikipedia.org/wiki/Fourier_analysis The process() loop looks like: { apply_window(); smp2freq(); // Fourier Analysis process_spectrum(); freq2smp(); // Fourier Synth } In layman terms: There's a smart French guy by the name of Joseph F. sitting inside it: If you play him some audio: He thinks: "Hey, this is actually just a few simple sine-waves added together (superpositioned)", he quickly calculates their frequencies and amplitudes and asks "Now, you want to change the duration?" easy: "I'll generate some new sine-waves with these frequencies and amplitudes, how long did you say you want?" (The smart thing about this French guy is that he actually speaks fluent English - Sorry I could not resist :) Yes, but its implementation is not very well done. Indeed, to change the integrated player to output to JACK would be trivial. It currently uses the PortAudio's StreamCallback which is very similar to JACK's process callback. Using it to do "live" timestreching with JACK (jack-in -> jack-out) is AFAICT impossible because: <details> If you feed it N samples (or seconds) of audio you end up with M samples (or seconds) with M > N. One could do a kludge: eg. for a 1:10 time stretch with continuous output: - read 1 sec of audio from the input - ignore 9 secs of the input but I don't think this will be useful. </details> best, robin

Thanks for all the info guys, I will read about FFT and all that. I hope it will answer my wonders as to why there is no vibration in sound. As for JACK support.... would anyone be interested in adding it, if it is trivial? Would love to have it in my audio chain. As for real time stretching, this is not necessary - it does react to parameter change with a slight delay, but it is ok. Louigi.

Excerpts from Robin Gareus's message of 2010-09-30 16:21:27 +0200: I'd be interested. It would likely take me some time though, and maybe some guidance, but it's something I want to learn. I dabbled in C but can only read simple code. I start to study computer science next week, which will give me access to the library at least, to K&R and a bunch of others. The lectures at university are mostly in java but C is required at some point, and I guess it's better to start learning it sooner rather than later, besides, it seems to be needed for all the neat audio stuff ;) So yeah, I guess I should get the code, put it in a git repo and just start hacking...

Folderol wrote: Ageism lives I see. I'm in my mid 40s and I still have a passion for coding. I don't do as much audio coding as I used to but I am a significant contributor to the DDC compiler: http://trac.haskell.org/ddc/ Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/

On Friday, October 01, 2010 04:05:37 pm Folderol did opine: And I have about a decade on you, I'll be 76 Monday. ;-) Sheesh, no respect for the elders here. ;-P -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) A memorandum is written not to inform the reader, but to protect the writer. -- Dean Acheson

On Friday, October 01, 2010 05:12:54 pm Philipp Überbacher did opine: Thanks for the flowers, and I do share a story now and then, based on 45+ years as a broadcast engineer, 25 of them as the Chief Engineer, spread over 2 tv stations and a radio station. But I try to make the story's related apply to the subject of the message here that got my attention. On any mailing list, random data, no matter how good, is often considered as noise. -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) poverty, n.: An unfortunate state that persists as long as anyone lacks anything he would like to have.

On 2 October 2010 05:03, Philipp Überbacher &lt;hollunder(a)lavabit.com&gt; wrote: It's more than just experience; they've been there, done that. Indeed, every generation has a different kind of "experience", which is always unique and invaluable. I wish my grandparents were alive today so I could share with them the technology I grew to love (when I was younger I learnt everything from science to history from them), and in turn receive wisdom I could probably never pick up alone in a hundred years. Even looking at my father, I realise how little I've learnt for the 2+ decades I've been blessed to enjoy what this modern world has to offer. Sometimes young blood runs hot, too hot. Sorry for derailing the topic further. I guess the train is now on a collision course, but hats off to the old-timers from the audio industry! -- GPG/PGP ID: B42DDCAD

On 09/30/10 22:41, Philipp Überbacher wrote: Aww sorry. I saw this too late. The film I intended to watch was too boring so.. git clone git://rg42.org/paulstretch cd paulstretch ./compile_linux_fftw_jack.sh or get a diff: http://rg42.org/gitweb/?p=paulstretch.git;a=commitdiff_plain;hp=upstream;h=… There's still quite a few things to do. besides: it's a quick hack and far from optimal: - It currently only opens the JACK clients if it's playing. (after stop the port-connections are lost) - It auto-connects the jack-ports to the first two available outputs (unless you remove the -DENABLE_AUTOCONNECT_JACK compile option) - it does not do any resampling (but prints a warning if the samplerates mismatch and plays anyway) - it does not add a ringbuffer. (it may cause x-runs) So you can take this as inspiration :) Using a jack_ringbuffer is actually not that hard, have a look at the jack_capture.c example client (comes with the jackd source). Resampling with libsamplerate is not very complicated, although I'm sure you'll have to work out a few things if you do it the first time (it's a good learning experience - google for "secret rabbit code". OTOH I'm not convinced that the paulstrech player needs anyway. It may be a nice add-on feature though. Hint: for both ringbuffer and/or resampling you'll want to add a new function to Player.cpp that basically does the same as void Player::getaudiobuffer(int nsamples, float *out) I suggest getaudiobuffer_channel(int nsamples, int channel, float *out) Contact Paul and Alan Calvin :) Cheers! robin

On Thu, Sep 30, 2010 at 01:53:44PM +0200, Robin Gareus wrote: The original Fourier Transform as invented by the smart French guy of the same name does operate on continuous (as opposed to sampled) data from -inf to +inf. The 'spectrum' interpretation came later. It was originally a mathematical tool used to find integrals of functions that would be impossible to integrate in closed form, and Fourier himself used it to study the propagation of heat in solids. The DFT (Discrete FT) is the same thing operating on sampled signals. It is usually also limited in time. The FFT (Fast FT) is an algorithm to compute a finite-length DFT very efficiently. The 'spectrum' interpretation is really quite ambiguous. You could take the DFT of e.g. a complete Beethoven symphony. The result is the 'spectrum' and in theory this is fixed over infinite time - the frequencies that are present according to this spectrum are there *all the time*. But that is not how we would perceive the music - we do not hear a constant mash of all frequencies, the spectrum as we hear it changes over time. Ciao, -- FA There are three of them, and Alleline.

On 09/30/10 20:49, Philipp Überbacher wrote: Thanks a lot for this comprehensive history lesson. Back when I was introduced to FT in some Physics lecture I was happy that I was able to use it and completely forgot to check the history :) Probably related to why I favored experimental Physics over Theory. correct. Furthermore there are different kind of windows (here a window refers to a block of audio-samples) and windows can overlap. That's where it gets complicated. The most commonly known window shapes are Rectangular window, Gaussian , Hann- and Hamming windows (the last two are cosine shapes) which allow to avoid discontinuities when processing blocks. Cheers! robin

On Thursday 30 September 2010 23:16:32 fons(a)kokkinizita.net wrote: And if I remember correctly, the minimum frequency you get is double your window-width. So when you do smaller windows like a jack-buffer of 128 samples at 48kHz, you only get down to 187.5 Hz... You can however overcome this if you run a second much bigger window where you only extract the low frequencies. (Unless Fons corrects me on that...) Have fun, Arnold

On Thu, 2010-09-30 at 23:16 +0200, fons(a)kokkinizita.net wrote: You are now (heading towards) argueing that the crescendo in the Beethoven symphony should be heard ahead of time - because we know it is there? Another example: Play two sinewaves, 1 Hz apart. Every second they will cancel out each other completely - even if they undeniably are both present. Just because something exist does not mean that it can be observed at each and every point in time. -- eins, zwei, drei ... tekno tekno?? http://www.youtube.com/watch?v=ZEgbW1FxR78

I understand Fourier invented the Fourier Series "only", anyone knows who generalized it to FT? And yes, I think the FT isn't so hard to understand, and a pretty useful concept. FFT, on the other hand... never really tried. Greetings

On Fri, 1 Oct 2010 03:25:50 +0200 Robin Gareus &lt;robin(a)gareus.org&gt; wrote: mmm, that sounds ambigous to me... what exactly does it mean? "Integral" here is used as synonimous of "solution of the heat-PDE"? What is sure is that Dirichlet was the first to prove a convergence theorem for Fourier series, and for this he had to first give a new (both more precise and more general) definition of function. He gave the famous Dirichlet-function (charateristic of rationals) as an example of a function - with his definition - not satisfying the hypotheses of his theorem. Then Riemann, Cantor and others proved other convergence theorems with other hypotheses. Cantor actually developed his set theory for the need to describe the set of some points regarding Fourier series (something like the admissable discontinuities, can't remember ATM). Also I seem to remember that Riemann's integral sprouted from some problems he had in Fourier analysis. Unfortunately I still know nothing about the history of the FT... renato

On Thu, September 30, 2010 3:47 am, Robin Gareus wrote: I suppose you missed the various windows he has added. Maybe they are just there for show? But nothing wrong with having it jackified for ease of use. Especially if the portaudio code is as you suggest. No doubt someone will take the time if it is a good piece of software otherwise. -- Patrick Shirkey Boost Hardware Ltd.

Excerpts from Louigi Verona's message of 2010-09-30 20:56:31 +0200: Thanks, but I'm not on ubuntu and it's kind of hard to find the sources there...

Excerpts from Robin Gareus's message of 2010-09-30 21:21:49 +0200: Ah, thanks, I somehow had ~2004 in mind, but maybe the 2006 version was the latest last time I checked. Thanks, I'll try it.