How to process music to a monotone?
March 12, 2010 8:17 PM Subscribe
How can I remove the component of pitch (or frequency) from an audio waveform, while maintaining the amplitude? I'm looking for a process analogous to desaturating a colour image to black and white: rendering an audio file literally monotonous but maintaining its rhythm (as changes in volume over time).
I believe Audacity can do that. For freeware it is very powerful.
posted by JJ86 at 8:24 PM on March 12, 2010
posted by JJ86 at 8:24 PM on March 12, 2010
Response by poster: JJ86 — I've had some experience with Audacity so that would be great news. Was there a specific feature you were thinking about? Or were you referring to one of b1tr0t's ideas?
posted by sixswitch at 8:27 PM on March 12, 2010
posted by sixswitch at 8:27 PM on March 12, 2010
As I heard your question, what occurred to me would be to convert the sound to an MP3, and then go into the MP3 and change all the frequency components to be the same, without affecting the amplitude components of them all. But that would probably require custom code; I doubt there's a tool that could do that.
posted by Chocolate Pickle at 8:29 PM on March 12, 2010
posted by Chocolate Pickle at 8:29 PM on March 12, 2010
Can you expand on what you're trying to do? I can't figure out exactly what you're asking.
posted by jjb at 8:31 PM on March 12, 2010
posted by jjb at 8:31 PM on March 12, 2010
Or nevermind, clearly these other guys know what you're asking.
posted by jjb at 8:31 PM on March 12, 2010
posted by jjb at 8:31 PM on March 12, 2010
Response by poster: I want to (veeeery roughly) approximate the effect of creating an unpitched arrangement of a piece of music.
posted by sixswitch at 8:37 PM on March 12, 2010
posted by sixswitch at 8:37 PM on March 12, 2010
Best answer: What you want is an envelope filter applied to... something. A single tone maybe, or white noise or pink noise or something. The "envelope" is what the amplitude looks like, without any frequency information. On it's own it's not going to sound like anything at all.
You could probably do this with a vocoder effect - a vocoder is usually used to apply the amplitude bits from something (your voice) to the frequency bits from something else (an instrument). You'll have to figure out what you want your "instrument" to be (like I said, a single sine wave or maybe white noise or something)
posted by RustyBrooks at 8:45 PM on March 12, 2010 [2 favorites]
You could probably do this with a vocoder effect - a vocoder is usually used to apply the amplitude bits from something (your voice) to the frequency bits from something else (an instrument). You'll have to figure out what you want your "instrument" to be (like I said, a single sine wave or maybe white noise or something)
posted by RustyBrooks at 8:45 PM on March 12, 2010 [2 favorites]
Response by poster: And this is why Audacity drives me nuts! It won't stay open for more than 30 seconds, on Windows or Mac OS. Flargh.
Great ideas, everyone...will try the vocoder first, as soon as I can get Audacity sobered up. If you read this, keep tossing 'em out if you have any others!
posted by sixswitch at 9:35 PM on March 12, 2010
Great ideas, everyone...will try the vocoder first, as soon as I can get Audacity sobered up. If you read this, keep tossing 'em out if you have any others!
posted by sixswitch at 9:35 PM on March 12, 2010
Best answer: +1 b1trot, you want the opposite of a sidechain compressor.
One supposes you could sidechain the sidechained track.
E.g.
Track 1 - original source
Track 2 - sidechain of original source compressing white noise
Track 3 - sidechain of track 2 compressing white noise
So you're inverting the signal twice. You'd have to mess with the compression levels to get something has the effect you want, but this should get you in the ballpark.
posted by MesoFilter at 9:38 PM on March 12, 2010
One supposes you could sidechain the sidechained track.
E.g.
Track 1 - original source
Track 2 - sidechain of original source compressing white noise
Track 3 - sidechain of track 2 compressing white noise
So you're inverting the signal twice. You'd have to mess with the compression levels to get something has the effect you want, but this should get you in the ballpark.
posted by MesoFilter at 9:38 PM on March 12, 2010
Do you have a mac and logic? Logic has a plug in oscillator/filter effect which can do this very easily.
The compressor sidechain idea is a decent one, though. But I would do it with a gate/sidechain. The Pgm material with rhythmically open a gate which white noise etc is running through. I think the logic oscillator thing would track much better though. Gates/comps would probably not be high resolution enough, so to speak to render much more than the basic rhythmic feel of the source material.
The line 6 m9/m13 has a freq shift effect which can do this as well.
posted by tremspeed at 10:30 PM on March 12, 2010
The compressor sidechain idea is a decent one, though. But I would do it with a gate/sidechain. The Pgm material with rhythmically open a gate which white noise etc is running through. I think the logic oscillator thing would track much better though. Gates/comps would probably not be high resolution enough, so to speak to render much more than the basic rhythmic feel of the source material.
The line 6 m9/m13 has a freq shift effect which can do this as well.
posted by tremspeed at 10:30 PM on March 12, 2010
Response by poster: tremspeed and MesoFilter: I have just gone through Audacity (crashes on Win & Mac), Kristal Audio (no routing), and Reaper (can't route to the sidechain plugin). I don't have Logic, but I have access to a Mac & PC. Do you guys know of any free sequencers / VST hosts that might be up to this? Or am I missing something on one of the previous (free and therefore sketchy) programs?
posted by sixswitch at 11:04 PM on March 12, 2010
posted by sixswitch at 11:04 PM on March 12, 2010
Sorry, haven't used those.
Bias Peak has something like a frequency shifter effect, try getting a demo version?
posted by tremspeed at 11:24 PM on March 12, 2010
Bias Peak has something like a frequency shifter effect, try getting a demo version?
posted by tremspeed at 11:24 PM on March 12, 2010
Seconding Melodyne.
posted by Blazecock Pileon at 11:29 PM on March 12, 2010
posted by Blazecock Pileon at 11:29 PM on March 12, 2010
Response by poster: Sorry, didn't mean to dis Audacity -- I just keep having bad luck with it. I appreciate the help. I think I will try the Ableton trial...but maybe tomorrow morning!
posted by sixswitch at 11:42 PM on March 12, 2010
posted by sixswitch at 11:42 PM on March 12, 2010
Best answer: If you only want the rhytmical aspect (no tones), use white noise as target for a vocoder, controlled by your source audio. You will end up with no tonal information, only the rhythm, like a ghost. Number of bands controls detail. There is a vocoder in Ableton Live.
You can probably also do this with other means, but this is very simple to do with a vocoder, it is what they do.
If you still want a "monotone" (and a representative monotone from the source), I think you have to use Melodyne Editor, polyphonic mode, detect the most prominent frequency, and either delete all the others or adjust them all to the same. Or you could find the prominent frequency by ear or some other editor, and reinsert this tonal frequency as a single tone in the above mentioned vocoder.
posted by gmm at 12:19 AM on March 13, 2010
You can probably also do this with other means, but this is very simple to do with a vocoder, it is what they do.
If you still want a "monotone" (and a representative monotone from the source), I think you have to use Melodyne Editor, polyphonic mode, detect the most prominent frequency, and either delete all the others or adjust them all to the same. Or you could find the prominent frequency by ear or some other editor, and reinsert this tonal frequency as a single tone in the above mentioned vocoder.
posted by gmm at 12:19 AM on March 13, 2010
Response by poster: Wow, thanks everybody. Got a rough version hacked together using the built-in sidechain compressor in the trial for Ableton. I learned a crapload tonight!
posted by sixswitch at 12:44 AM on March 13, 2010
posted by sixswitch at 12:44 AM on March 13, 2010
I think Praat could probably do this well, though I'm not sure of the exact procedures you'd use.
posted by greatgefilte at 10:35 AM on March 13, 2010
posted by greatgefilte at 10:35 AM on March 13, 2010
Response by poster: Wow. Praat and the stuff at PSL just blew my mind. Killer links, everybody, and thanks for getting me up to speed so fast and sharing your knowledge of what for me is unfamiliar territory.
posted by sixswitch at 11:42 AM on March 13, 2010
posted by sixswitch at 11:42 AM on March 13, 2010
This thread is closed to new comments.
posted by sixswitch at 8:17 PM on March 12, 2010