Better Spectrum Anayzer Features -Sound Forge - PC

Subject:Better Spectrum Anayzer Features Posted by: bbrustad Date:12/1/2004 5:01:29 PM I would like to see the Spectrum Analyzer show a higher resolution than 1Hz. Often times I'm looking at notching out harmonics and would like .02 Hz resolution. Also provide peak detection on the SA and allow it to view only 1 channel of a stereo file. I would also like the Q of the paragraphic notch filters to go down to 1/64 octave as a minimum. It would also be nice to have more than 4 filters available. There is a lot of wasted space with the fader bars. This would be nice either as a plug-in or feature. Also, I would like to be able to time align a stereo audio file (move the left track in relation to the right track) down to the sample. At least allow you to insert silence as a sample value instead of a time value. This is very difficult to do in Vegas and Foundry. I'd also like to have the option to keep all my sfk files in a separate folder from the actual wavs. Keeping them hidden helps but doesn't solve the file management issue. It would also be nice if track-at-once could let you specify track breaks based on regions. A debuzz harmonics suppression filter would be nice.
Subject:RE: Better Spectrum Anayzer Features Reply by: musicvid10 Date:12/1/2004 9:02:21 PM Sony, I hope you're listening here. All valid requests for anyone into the science of acoustics as I was in college and as a pro musician. bbrustad, I wouldn't hold out much hope for your last request, because as you know, the harmonics from any nonsinusoidal waveform are unbelievably complex. An interesting parallel discussion going on HERE
Subject:RE: Better Spectrum Anayzer Features Reply by: Sonic Date:12/2/2004 2:51:09 PM Spectrum Analyzer bin resolution is limited by FFT size. 1 Hz is just the label/grid/tip resolution. Bin resolution under 1 Hz would require a very large FFT size. If you wish to view separate channels, you can turn off Single Graph mode. 1/64 octave Q is way beyond the capabilities of the filters in the paragraphic EQ. This would have to be a new effect. Most built-in tools pay attention to the current position format. If you want to insert silence as samples, change the window's position format to samples, then call up insert silence. .sfk files must reside next to their media file in order to be used by other Sony applications. If you don't mind waiting for them to rebuild, you might try the "Delete temporary files on close" pref, which will create new .sfk files in the temp folder and then delete them when they are no longer in use (by forge). J.
Subject:RE: Better Spectrum Anayzer Features Reply by: mpd Date:12/3/2004 7:34:32 AM Caffine is still kicking in, but at a 48kHz sampling rate, 0.02 Hz resolution requires about a 2.4 million point FFT. That is pretty big. It is possible to do a smaller FFT and then interpolate either a local peak location or a local range, though. A 1/64 octave bandstop filter would have some really wicked group delay at the band edges if it was an IIR filter, and a FIR version would have to be pretty long. If you really need these features, you may be better off using Matlab or one of the clones.

Subject:Better Spectrum Anayzer Features

Posted by: bbrustad

Date:12/1/2004 5:01:29 PM

I would like to see the Spectrum Analyzer show a higher resolution than 1Hz. Often times I'm looking at notching out harmonics and would like .02 Hz resolution. Also provide peak detection on the SA and allow it to view only 1 channel of a stereo file.

I would also like the Q of the paragraphic notch filters to go down to 1/64 octave as a minimum. It would also be nice to have more than 4 filters available. There is a lot of wasted space with the fader bars. This would be nice either as a plug-in or feature.

Also, I would like to be able to time align a stereo audio file (move the left track in relation to the right track) down to the sample. At least allow you to insert silence as a sample value instead of a time value. This is very difficult to do in Vegas and Foundry.

I'd also like to have the option to keep all my sfk files in a separate folder from the actual wavs. Keeping them hidden helps but doesn't solve the file management issue.

It would also be nice if track-at-once could let you specify track breaks based on regions.

A debuzz harmonics suppression filter would be nice.

Subject:RE: Better Spectrum Anayzer Features

Reply by: musicvid10

Date:12/1/2004 9:02:21 PM

Sony,
I hope you're listening here. All valid requests for anyone into the science of acoustics as I was in college and as a pro musician.

bbrustad, I wouldn't hold out much hope for your last request, because as you know, the harmonics from any nonsinusoidal waveform are unbelievably complex.

An interesting parallel discussion going on HERE

Subject:RE: Better Spectrum Anayzer Features

Reply by: Sonic

Date:12/2/2004 2:51:09 PM

Spectrum Analyzer bin resolution is limited by FFT size. 1 Hz is just the label/grid/tip resolution. Bin resolution under 1 Hz would require a very large FFT size. If you wish to view separate channels, you can turn off Single Graph mode.

1/64 octave Q is way beyond the capabilities of the filters in the paragraphic EQ. This would have to be a new effect.

Most built-in tools pay attention to the current position format. If you want to insert silence as samples, change the window's position format to samples, then call up insert silence.

.sfk files must reside next to their media file in order to be used by other Sony applications. If you don't mind waiting for them to rebuild, you might try the "Delete temporary files on close" pref, which will create new .sfk files in the temp folder and then delete them when they are no longer in use (by forge).

J.

Subject:RE: Better Spectrum Anayzer Features

Reply by: mpd

Date:12/3/2004 7:34:32 AM

Caffine is still kicking in, but at a 48kHz sampling rate, 0.02 Hz resolution requires about a 2.4 million point FFT. That is pretty big. It is possible to do a smaller FFT and then interpolate either a local peak location or a local range, though.

A 1/64 octave bandstop filter would have some really wicked group delay at the band edges if it was an IIR filter, and a FIR version would have to be pretty long.

If you really need these features, you may be better off using Matlab or one of the clones.