Recording Engineering & Production: Making records better, not just louder and cheaper. The brightest minds from the creative and scientific world of audio discuss the elements of audio production that interest you most.
I use the Lynx Two AD/DA converter. It is optional to turn on dithering when recording, but not recommended to do so if recording at 24bit. Should one expect any audible difference between dithering and not dithering when recording at 24 bit? On my brief experiments, listening with headphone, I thought the dithered recording sounded better, but it is difficult to be sure without doing repeated, random, blind tests. Is there a definitive answer to this?
Thanks,
Patrick Bower
Barry Hufker Messages: 8100 Registered: October 2004 Location: between heaven and hell
Platinum Member
While I suppose one could dither at any point, the main point of dithering is to "hold on" to bits that would otherwise be lost as one goes from a higher bit depth to a lower one.
I never dither when recording straight to 24 bits. I dither when going from 24 to 16. Nuendo, the program I use, probably dithers from its 32 floating point processing when exporting to 24 bits.
I also don't use noise-shaping,which may be part of the dither you're using. This could make the recording "sound better" by reducing noise in the middle frequencies. I just don't think noise-shaping is benign and so makes an unpleasant audible difference.
You don't need to dither the output of the ADC during recording at 24 bits if the noise of the ADC itself is greater than around -140dB ref full scale - because the noise of the converter is providing more than adequate dither anyway. Since as far as I know no converter has this level of SNR you are safe to turn the dither off when going 24bit.
However modern converters are better than 16bit (96dBFS SNR) therefore if you are outputting (or storing) only 16bits then dither will be required. This should be conventional 'white or pink' dither and noise shaping dither should be avoided if possible.
As has been posted - noise shaping dither lowers the perceived noise of the dither by pushing it into freq ranges you don't hear so readily. This actually causes the peak noise level to increase. This means that the signal is more fragile to further changes (cos anything that happens to it might put this higher noise level back into bands you can hear again) and therefore it's less tolerant of processing down line. So personally I would never use noise shaping dither on any signal during recording. Noise shaping should only be done (only if really needed to reduce perceived dither noise below -93dBFS at 16bits) at the very last stage of the production process - perhaps at the mastering stage
Members
Search
FAQ
Register
Login
Home
