Thank you all for your responses - each was very helpful! I am particularly interested (mostly out of curiosity) in how to "measure" the room with convolution - would I blast some pink noise and then re-record it with a good microphone, and then perform a frequency analysis on that? I am sure I can look this up somewhere. Would this then yield the resonant frequencies in the room?
Kevin
Hallo, Kevin McCoy hat gesagt: // Kevin McCoy wrote:
Thank you all for your responses - each was very helpful! I am particularly interested (mostly out of curiosity) in how to "measure" the room with convolution - would I blast some pink noise and then re-record it with a good microphone, and then perform a frequency analysis on that? I am sure I can look this up somewhere. Would this then yield the resonant frequencies in the room?
This nice paper may be interesting: http://www.kokkinizita.net/linuxaudio/papers/aliki.pdf http://www.kokkinizita.net/linuxaudio/index.html
Frank Barknecht _ ______footils.org_ __goto10.org__
Usually acoustic measurements are done with impulses, AFAIK. An
ideal impulse actually has all frequencies in it, so it's useful for
that kind of thing. Plus it's easy to differentiate between the
initial signal and the room effects just based on time.
.hc
On Feb 8, 2007, at 3:13 PM, Kevin McCoy wrote:
Thank you all for your responses - each was very helpful! I am
particularly interested (mostly out of curiosity) in how to
"measure" the room with convolution - would I blast some pink noise
and then re-record it with a good microphone, and then perform a
frequency analysis on that? I am sure I can look this up
somewhere. Would this then yield the resonant frequencies in the
room?Kevin
--
++++ http://pocketkm.blogspot.com
--
++++ http://pocketkm.blogspot.com _______________________________________________ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
Computer science is no more related to the computer than astronomy is
related to the telescope. -Edsger Dykstra
If I'm correct, an impulse is just a moment where the sample level goes from one to zero, causing a click? I'm interested in doing this still but the things I've found are a little bit technical (explanations, but no practice/application/instruction)... I'll keep looking.
Kevin
On 2/13/07, Hans-Christoph Steiner hans@eds.org wrote:
Usually acoustic measurements are done with impulses, AFAIK. An ideal impulse actually has all frequencies in it, so it's useful for that kind of thing. Plus it's easy to differentiate between the initial signal and the room effects just based on time.
.hc
On Feb 8, 2007, at 3:13 PM, Kevin McCoy wrote:
Thank you all for your responses - each was very helpful! I am particularly interested (mostly out of curiosity) in how to "measure" the room with convolution - would I blast some pink noise and then re-record it with a good microphone, and then perform a frequency analysis on that? I am sure I can look this up somewhere. Would this then yield the resonant frequencies in the room?
Kevin
--
++++ http://pocketkm.blogspot.com
--
++++ http://pocketkm.blogspot.com _______________________________________________ PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Computer science is no more related to the computer than astronomy is related to the telescope. -Edsger Dykstra
I practice recording for convolution by simply clapping in the hand on the particular location. Later on the computer I cut away the inital sound of the clap to have only the tail of reverbration available. May be not that accurate in the scientific way but works well for me.
But convolution can do more than adding room to recordings, I got incredible results by convolving synthetic sounds (which have to have a lot of harmonics to be useful, simular as in vocoding which is close to convolution) with speech as 'impulse'.
Cheers,
Malte
I've used very rough "impulse" recordings to do binaural placement for sounds--as well as pure abstract textures as Malte suggested--with very nice results. See soundfiles and descriptions below from Karosta Project (Latvia, 2002-3) http://karosta.edworks.net/
best, d.
binaural.tunnel.study.1 [w/max.borisov][OGG 2min00sec 1.67Mb] http://www.gracies.org/derek/audio/field.recordings.2/binaural.tunnel.study....
An excerpted acoustic study of an abandoned Soviet aircraft bunker. Max Borisov wears ear-mount binaural microphones, and I test the space with bricks and stones. Max and I found this technique very interesting, as it allows one person to literally "become" a human microphone, and the other to perform for that audience/recorder [many thanks to Aaron Ximm for inspiration in this and other matters]. A clever soul might use these acoustic signatures to place other sounds in the same space virtually...
Original Formanta EMS-01 Synthesizer improvisation [OGG 2min08sec 1.34Mb] http://www.gracies.org/derek/audio/formanta.original.edit.ogg
Formanta.Drone [Surface.Noise] [OGG 7min10sec 4.92Mb] http://www.gracies.org/derek/audio/formanta.drone.edit.ogg
The first is an excerpt of a live improvisation made on Maxim Borisov's Formanta EMS-01 synthesizer. This is a big, nasty Russian synth that is almost totally useless for melody, but has this amazing noise-driven Low Frequency Oscillator with an array of filters behind it. All sounds on the first track came only from this random sequencer, without laying a finger on the keys. The second track resulted from a digital cross-breeding with surface noise from an old Soviet marching record--instant texture and depth!
That's a sweet concept, and a fun way to collaborate with active field recording. I'll have to try that (giving you guys reference cred of course)!
~Kyle
On 2/13/07, Derek Holzer derek@umatic.nl wrote:
I've used very rough "impulse" recordings to do binaural placement for sounds--as well as pure abstract textures as Malte suggested--with very nice results. See soundfiles and descriptions below from Karosta Project (Latvia, 2002-3) http://karosta.edworks.net/
best, d.
binaural.tunnel.study.1 [w/max.borisov][OGG 2min00sec 1.67Mb] http://www.gracies.org/derek/audio/field.recordings.2/binaural.tunnel.study....
An excerpted acoustic study of an abandoned Soviet aircraft bunker. Max Borisov wears ear-mount binaural microphones, and I test the space with bricks and stones. Max and I found this technique very interesting, as it allows one person to literally "become" a human microphone, and the other to perform for that audience/recorder [many thanks to Aaron Ximm for inspiration in this and other matters]. A clever soul might use these acoustic signatures to place other sounds in the same space virtually...
Original Formanta EMS-01 Synthesizer improvisation [OGG 2min08sec 1.34Mb] http://www.gracies.org/derek/audio/formanta.original.edit.ogg
Formanta.Drone [Surface.Noise] [OGG 7min10sec 4.92Mb] http://www.gracies.org/derek/audio/formanta.drone.edit.ogg
The first is an excerpt of a live improvisation made on Maxim Borisov's Formanta EMS-01 synthesizer. This is a big, nasty Russian synth that is almost totally useless for melody, but has this amazing noise-driven Low Frequency Oscillator with an array of filters behind it. All sounds on the first track came only from this random sequencer, without laying a finger on the keys. The second track resulted from a digital cross-breeding with surface noise from an old Soviet marching record--instant texture and depth!
-- derek holzer ::: http://www.umatic.nl ---Oblique Strategy # 160: "Towards the insignificant"
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
An excerpted acoustic study of an abandoned Soviet aircraft bunker. Max Borisov wears ear-mount binaural microphones, and I test the space with bricks and stones. Max and I found this technique very interesting, as it allows one person to literally "become" a human microphone, and the
That is the good old Kunstkopfaufnahme (dummy head recording)
http://en.wikipedia.org/wiki/Dummy_head_recording
which is used for live-like live recordings
http://de.wikipedia.org/wiki/Binaurale_Tonaufnahme
in your case with a real head. I have to check out your recordings and place some stuff into the bunker :)
Cheers,
Malte Steiner media art + development -www.block4.com-
Hallo, Hans-Christoph Steiner hat gesagt: // Hans-Christoph Steiner wrote:
Usually acoustic measurements are done with impulses, AFAIK. An
ideal impulse actually has all frequencies in it, so it's useful for
that kind of thing. Plus it's easy to differentiate between the
initial signal and the room effects just based on time.
Actually I think, engineers prefer to use frequency sweeps instead of pulses nowadays. The problem with pulses is, that they are hard to get right: You need a clean pulse and a very silent environment:
The main problem with using Dirac pulses in an acoustical measurement is that as a result of their very short duration and finite amplitude, they contain very little energy, and measure- ment accuracy will be limited by the signal to noise ratio of the equipment used and of the system itself. While it is possible to use Dirac pulses reproduced by a loudspeaker in the con- trolled environment of an acoustics laboratory, this is all but infeasible in most real life situa- tions, e.g. for measuring a room or concert hall, where there will always be background noises of some sort.
This is from Fons Adriaensen's LAC paper: http://lac.zkm.de/2006/papers/lac2006_fons_adriaensen_01.pdf
Fons then explains sine sweeps as an alternative method:
The advantage of using a sweep is that at any time we produce only a single frequency, and any distortion introduced will consist of the har- monics of that frequency only. If we use a rising frequency sweep, the harmonics will be gener- ated ahead of the same frequencies appearing in the signal. So after deconvolution, any dis- tortion will appear as spurious peaks in negative time in the impulse response, and most of it can then be edited out easily.
This method's origin according to Fons is a paper by A. Farina:
Angelo Farina. 2000. Simultanuous measurement of impulse response and distortion with a swept-sine technique. Audio Engineering Society Preprint 5093.
It is implementes in Fons' ALIKI software, that he presented at LAC2006.
Frank Barknecht _ ______footils.org_ __goto10.org__
On 2/13/07, Frank Barknecht fbar@footils.org wrote:
Actually I think, engineers prefer to use frequency sweeps instead of pulses nowadays
How fast of a sweep are you talking about? Seconds or milliseconds? Is this the 50 and 80 ms bit that the paper mentions?
If we use a rising frequency sweep, the harmonics will be gener- ated ahead of the same frequencies appearing in the signal. So after deconvolution, any dis- tortion will appear as spurious peaks in negative time in the impulse response, and most of it can then be edited out easily.
What a great way to use an artifact of the FFT to one's advantage. Awesome!
~Kyle
Hallo!
Usually acoustic measurements are done with impulses, AFAIK. An
ideal impulse actually has all frequencies in it, so it's useful for
that kind of thing. Plus it's easy to differentiate between the
initial signal and the room effects just based on time.Actually I think, engineers prefer to use frequency sweeps instead of pulses nowadays. The problem with pulses is, that they are hard to get right: You need a clean pulse and a very silent environment:
Well, there are also problems with swept sines of course ... ;)
Other methods are e.g.
think there is also an mls example in the iem_tab library. -Least Square Error methods (see system identification example in the adaptive library)
LG Georg
Usually acoustic measurements are done with impulses, AFAIK. An ideal impulse actually has all frequencies in it, so it's useful for that kind of thing. Plus it's easy to differentiate between the initial signal and the room effects just based on time.
.hc
Impulses are okay, but there are better ways. An impulse is difficult to deal with because of signal to noise ratio. A simple look at room reverb as a convolution operator gives a good way to compute the effects of room reverberation using any length of audio signals that cover the audio frequency range (noise will do fine, frequency sweeps, etc...) The principle of maximum length sequence spectrum analysis is that the autocovariance function of a maximum length sequence is an impulse.
k(t) is the room reverberation signal x(t) is the reference signal, (must cover the whole freq range) m(t) is the received signal n(t) is noise
key assumption: the correlation between x(t) and n(t) is zero, on average
given x*k (t) +n(t)=m(t)---->we want to find k(t)
the adjoint of a convolution operator is cross-covariance, a function of time. Cross-covariance is the equivalent of convolution by a time-reversed copy of a signal. Lastly, convolution operators are commutative.
x*k (t) = m(t)-n(t) This equation is handled in the least square error sense by applying the adjoint to each side of the equation, and inverting the adj(X)*X operator on the LHS.
xcov(x(t), x*k(t))= xcov(x(t),m(t))-xcov(x(t),n(t)) xcov(x(t), x(t))*k(t)= xcov(x(t),m(t)) /-the noise cancels out, on average k(t)=xcov(x(t),x(t))^-1 xcov(x(t),m(t))
In the fourier domain, K(f)=(X(f)*conj(X(f)))^-1 (conj(X(f))M(f))
I've been working on a patch to make this work since I've been reading this thread....but I'm having some trouble in Pd lately. The cross-covariance function requires fft and ifft. Capturing the room response has to occur on *large* block sizes (2x the length of the room reverberation you want to recover)
The patch undoubtedly has errors (especially with the syncing of signals). I was unable to test it yet. There are two externals, xcov~ and maxval~ (these are very very standard in terms of comiling) and z~ from zexy and expr~ Anyway, it needs help, and I'm not getting it done fast enough... Chuck