Glenn Gould De-Vocalizer

[steve_m]

https://www.unpronounceable.com/gould/

[Kassaa]

5500$ what.

And I, for some reason, love the humming of Gould.

Edit: Listened to the partita, absolutely comedic genius 

[steve_m]

3

[henrah]

hahahaahahahahahahahahahahahahahahahahahahahaha

[quantum]

hehehehehe

I like the spectral analysis for the second sample.  Show me a human who sings between 2000 Hz and 3000 Hz.  Must be amzing voice to produce splotches, without anything above or below it. 

[Motrax]

I can attest to the incredible nature of this device. It arrived two weeks ago after a long, anticipatory eleven months, and I now happily listen to the Goldberg Variations without a hint of that terrible, extraneous sound. The inventor of this device is no less than musical genius, able to comprehend sound in a way no other human could. To me, his intellect and impact upon the musical community can be compared with the likes of Bach, Brahms, and Kalkbrenner.

This is my opinion of course, but I daresay it is a difficult one to refute.

- M

[steve jones]

hehehehehe

I like the spectral analysis for the second sample.  Show me a human who sings between 2000 Hz and 3000 Hz.  Must be amzing voice to produce splotches, without anything above or below it. 

Precisely.

Not to mention the high inharmonic partials (ie, the breathy sounds) which would mix with the 'air' of the recording. To remove Gould's vocalisations is absolutely impossible. Theoretically it could be done using phase cancellation, but in reality, no chance.

SJ

[ramseytheii]

Precisely.

Not to mention the high inharmonic partials (ie, the breathy sounds) which would mix with the 'air' of the recording. To remove Gould's vocalisations is absolutely impossible. Theoretically it could be done using phase cancellation, but in reality, no chance.

SJ

Didn't you hear the "after" recording?  It's gone!

Walter Ramsey

[steve jones]

I havent no.

But I do know that what has been described is physically impossible. There is only one way to do this - you would need the original recording, and another recording of the vocal noise which is sample accurate to that in the original. Then you could use phase cancellation to remove the vocal noise from the original recording without damaging the piano part to badly.

Another way would be to use FFT analysis to build a filter tuned specifical to the spectral content of the 'noise'. For this, again you would need to have the noise in isolation so that it would be analyzed.

What I imagine these guys have done is to build a filter with roughly the same spectral content as the noise in the Gould recordings. This could then be applied to attenuate the noise in the original recordings. However, it is not without fault - frequencies overlap, and this results in nasty sideffects. For example, the transient data in the piano attack will probably be damaged.

However, I have yet to hear the demos, so first let me check those before I condemn this technology any further!

SJ

[steve jones]

Haha, I should really check these links before attempted a serious reply!!!

Lol, what can I say, I want one! Scratch that, I NEED one. Where is credit card...

SJ

[steve_m]

w

[steve jones]

Haha, lol 

You should probably understand though, that I work in music technology and hear claims like this day in day out. So when you mentioned that someone was selling some piece of technology (software I assumed) to remove vocals from mixes, I didnt even consider checking the site! I hear people asking me this very question all the time, and the answer is always the same... no chance!

Ofcourse, little did I appreciate the genius of these guys! I wonder if they could make my piano sing 'In da Ghetto'? 

SJ

[ramseytheii]

Haha, I should really check these links before attempted a serious reply!!!

Lol, what can I say, I want one! Scratch that, I NEED one. Where is credit card...

SJ

 

Walter Ramsey
 

[Motrax]

I was wondering why this discussion was so serious!