Dither question

[TranscendingM]

Originally Posted by iz_thewiz

If I record into my MPC 1000 at 16/44.1 (as this is the only resolution available), should I keep my DAW session at the same resolution for printing from the MPC? I don't think there would be any logical benefit recording into the DAW at a higher resolution than what is available on the MPC (i.e from 16 bit to 24 bit).

Also, is dithering necessary at 16 bit sessions before the final bounce out?

"It is not necessary to expand the wordlength of the session before mixing: the sound can never get more resolved than what was originally encoded."

Bob Katz - 'Mastering Audio'

Have to be careful how you interpret that - and I'm assuming you quoted correctly. So if it is correct, this can be misinterpreted to mean while actual mixing you don't need increased word length where in fact you do. Even volume adjustments start to increase word lengths and potential for rounding errors. So for recording (hence "before mixing") one can argue it is not necessary but in mixing where there is processing, that's a different story.

As for dithering, again, if there was any processing you've increased word and you should dither.

[wado1942]

Essentially, if you can only record 16-bit signals, that's what you should capture. The DAW itself and any mixes you do should be at least 24-bit. Software like Sony Vegas is a bit tricky, because you can set it to capture 16-bit, but will use a 16-bit buss or set it to capture 24-bit and will use a 24-bit buss. Most software will let you chose those settings independently.

[huejahfink]

Do you mean that you are sampling into the MPC at 16bit, composing and then recording the analogue output into your DAW? If so record your MPC at 24bit, maintain that depth for any subsequent editing in the DAW, and then only dither down to 16bit as a final stage for a CD or for any digital distributors etc who require 16bit files. Keep a 24 bit version as well though, use that to send to any external mastering places if necessary.

[iz_thewiz]

Originally Posted by huejahfink

Do you mean that you are sampling into the MPC at 16bit, composing and then recording the analogue output into your DAW?

Correct. My MPC only supports 16/44.1
I'll use 24bit sessions as the general verdict here is that its preferable, even when using a 16bit source.

[huejahfink]

Just the fact that you are recording from analogue outputs is good enough reason to go 24bit in my opinion....

I think Bob's comment is with regard to the idea of expanding the bit depth digitally from an existing 16bit file to 24bit. He is trying to dispel the 'myth' that it would automatically increase the quality. It doesn't. However, any subsequent mixing or processing benefits from greater depth. This is why most DAWs now operate at 32bit float for internal calculations, regardless of the fact that it will not be the original recorded depth or target depth.

In your case, any 16bit sounds are being converted to analogue signal on their way out of your MPC, so you ideally want to recapture that with as much detail as possible. I would say the same even if your machine only recorded samples at 8 or 12 bit and put out an analogue signal.

[Dene]

Originally Posted by huejahfink

Just the fact that you are recording from analogue outputs is good enough reason to go 24bit in my opinion....

I think Bob's comment is with regard to the idea of expanding the bit depth digitally from an existing 16bit file to 24bit. He is trying to dispel the 'myth' that it would automatically increase the quality. It doesn't. However, any subsequent mixing or processing benefits from greater depth. This is why most DAWs now operate at 32bit float for internal calculations, regardless of the fact that it will not be the original recorded depth or target depth.

In your case, any 16bit sounds are being converted to analogue signal on their way out of your MPC, so you ideally want to recapture that with as much detail as possible. I would say the same even if your machine only recorded samples at 8 or 12 bit and put out an analogue signal.

+1
exactly what I think as well...

[Squidly]

Exactly as above. Any processing (eq, compression, volumes changes) would benefit from a higher bit depth and sample rate.