Hi,
What's the scale of the "amplitude" value that comes packed together with each individual pitch? (i.e., the second value of the output of the third outlet in the help patch)?
If you open the help patch as is and don't touch anything, the example signal (phasor) is being detected ah having an "overall" amplitude of about 90 dB, while the unpacked amplitude of the (only) individual pitch is about 116. How should I interpret this?
thanks m.
The pitch track's "amplitude" is computed as the theoretical loudness of the track, which is computing by adding up the "loudnesses" of all the components, then converting back to dB. "Loudness" is incorrectly estimated as the fourth root of amplitude (should have been the fourth root of power, equal to the square root of amplitude, if one is to beloeve in "sones" as a psychoacoustically correct measure.
So the broghter the sound, the bigger the number (and the louder it should sound relative to its actual signal power). I haven't looked, but if you throw fiddle~ a sinusoid the two amplitude measures should be equal.
By the way, I think sigmund~ works way better than fiddle~ on most classes of input signals :)
Miller
On Thu, Mar 10, 2011 at 06:52:54PM +0100, Matteo Sisti Sette wrote:
Hi,
What's the scale of the "amplitude" value that comes packed together with each individual pitch? (i.e., the second value of the output of the third outlet in the help patch)?
If you open the help patch as is and don't touch anything, the example signal (phasor) is being detected ah having an "overall" amplitude of about 90 dB, while the unpacked amplitude of the (only) individual pitch is about 116. How should I interpret this?
thanks m.
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Hi,
Thank you for the clarifications, and sorry for replying with such a delay to a thread I started myself...
On 03/10/2011 07:31 PM, Miller Puckette wrote:
"Loudness" is incorrectly estimated as the fourth root of amplitude (should have been the fourth root of power, equal to the square root of amplitude,
Ok so I guess that's why the pitch track's amplitude can be greater than the sound's total amplitude.
By the way, I think sigmund~ works way better than fiddle~ on most classes of input signals :)
But correct me if I am wrong: with sigmund~ you can't get the "total" amplitude of a pitch track in a trivial way, can you? You would have to "manually" sum up the track's components' amplitudes, right? (well in a sense that _is_ trivial)
Thanks m.
are you looking for env~? Sigmund has that option too. J
On Mon, Mar 21, 2011 at 8:59 AM, Matteo Sisti Sette matteosistisette@gmail.com wrote:
Hi,
Thank you for the clarifications, and sorry for replying with such a delay to a thread I started myself...
On 03/10/2011 07:31 PM, Miller Puckette wrote:
"Loudness" is incorrectly estimated as the fourth root of amplitude (should have been the fourth root of power, equal to the square root of amplitude,
Ok so I guess that's why the pitch track's amplitude can be greater than the sound's total amplitude.
By the way, I think sigmund~ works way better than fiddle~ on most classes of input signals :)
But correct me if I am wrong: with sigmund~ you can't get the "total" amplitude of a pitch track in a trivial way, can you? You would have to "manually" sum up the track's components' amplitudes, right? (well in a sense that _is_ trivial)
Thanks m.
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
On 03/21/2011 07:32 PM, Jaime Oliver wrote:
are you looking for env~? Sigmund has that option too.
Both env~ and the "env" option of sigmund~ measure the overall amplitude of the signal. Fiddle~ measure the amplitude of the/each detected pitch track only, i.e. more or less the "sum" of the "amplitudes" of the sinusoidal components that make up the pitch track.
If you're following one or more pitches within a noisy signal that makes a big difference.
Doesn't the 'peaks and/or tracks' bit if sigmund~ do that?
On 21 March 2011 19:19, Matteo Sisti Sette matteosistisette@gmail.comwrote:
On 03/21/2011 07:32 PM, Jaime Oliver wrote:
are you looking for env~? Sigmund has that option too.
Both env~ and the "env" option of sigmund~ measure the overall amplitude of the signal. Fiddle~ measure the amplitude of the/each detected pitch track only, i.e. more or less the "sum" of the "amplitudes" of the sinusoidal components that make up the pitch track.
If you're following one or more pitches within a noisy signal that makes a big difference.
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-
J bz -
Jaime Oliver -
Matteo Sisti Sette -
Miller Puckette