Gearslutz

Gearslutz (https://www.gearslutz.com/board/)
-   So Much Gear, So Little Time (https://www.gearslutz.com/board/so-much-gear-so-little-time/)
-   -   Ultrasonic (48 kHZ+) signal boost to lower noise floor? (https://www.gearslutz.com/board/so-much-gear-so-little-time/1311917-ultrasonic-48-khz-signal-boost-lower-noise-floor.html)

davorfilms 23rd May 2020 04:22 PM

Ultrasonic (48 kHZ+) signal boost to lower noise floor?
 
Since most ADCs seem to use that technology (Delta-Sigma or similar sounding name) which raises the noise floor in ultrasonic frequencies, which is clearly visible in 192 kHz recording mode, and given that the intensity of ultrasonic frequencies in most audio content is negligible, I was wondering if it might be possible to effectively reduce the ultrasonic noise floor by boosting it by some 30-60 dB in the analogue domain before capturing the signal.

Is there any gear out there that can do this and do it without degrading the signal to a point where this whole endeavor becomes useless (read: in good quality)?

If not, how hard would it be to build something like this? Say, some kind of passive circuit which does this.

Then after recording, this EQ could be reversed, leading to a nice low ultrasonic noise floor.

UnderTow 23rd May 2020 04:39 PM

Quote:

Originally Posted by davorfilms (Post 14756206)
Since most ADCs seem to use that technology (Delta-Sigma or similar sounding name) which raises the noise floor in ultrasonic frequencies, which is clearly visible in 192 kHz recording mode, and given that the intensity of ultrasonic frequencies in most audio content is negligible, I was wondering if it might be possible to effectively reduce the ultrasonic noise floor by boosting it by some 30-60 dB in the analogue domain before capturing the signal.

Is there any gear out there that can do this and do it without degrading the signal to a point where this whole endeavor becomes useless (read: in good quality)?

If not, how hard would it be to build something like this? Say, some kind of passive circuit which does this.

Then after recording, this EQ could be reversed, leading to a nice low ultrasonic noise floor.

The Ultra-sonic noise floor is intentional. It is noise shaped dither. It needs to be there to avoid quantization distortion. The dither is pushed into the ultra-sonic range because you can't hear it there. :)


Alistair

davorfilms 23rd May 2020 04:47 PM

Quote:

Originally Posted by UnderTow (Post 14756237)
The Ultra-sonic noise floor is intentional. It is noise shaped dither. It needs to be there to avoid quantization distortion. The dither is pushed into the ultra-sonic range because you can't hear it there. :)


Alistair

I understand, but if I push high frequency content first and then lower it after the AD conversion, I should get a lower effective noise floor while profiting from the dithering anyway. Of course I would get less dynamic range, but since ultrasonic content seems to be pretty low anyway usually, that wouldn't be much an issue.