"Pre-mix" Dithering vs "Post-mix" Dithering - 96khz>44.1Khz

[The Silent Coup]

Hi all,

I often carry out basic mix work on my laptop at home on audio that has been recorded at the studio. My laptop set up runs at 44.1khz, while the sessions recorded at the studio are at 96khz. Both setups run on Cubase 5)

In order to work on my laptop, i have to convert/dither the audio down to 44.1khz (sometimes it is possible to work without converting the files if there arent many of them, as I use the ASIO4ALL codec with an Alesis io2 breakout box. However, i imagine what i am hearing is not "true" 96khz?)

It is possible for me to copy across the Cubase session and mix settings from my laptop to the studio computer, and run a mixdown from there, thus still running everything in the 96khz environment.

What I would like to know is whether it makes any "difference" (apologies for the vagueness of the term!) to the quality of final mix if the audio is dithered down prior to mixing at 44.1khz, as opposed to mixing at 96khz and dithering down the final mix.

As a final question, is there any difference in how a DAW treats audio at different frequency rates? (e.g. more headroom at 96khz?)

Thanks

[The Elf]

A fair bit of misunderstanding going on here, hence your confusion.

Dithering has nothing to do with sample rate. You don't 'dither down' between sample rates.

If you can run a Cubase Project at 96kHz then you can mix down at 96kHz. When you create your final mix file you can specify the output file's sample rate.

If you are recording at 24-bit, when you render the final mix file to 16-bit it is then that you would apply dither, because the word length is reducing. There have been a number of SOS articles that explain much of this in far more detail, and it is worth you doing a search or three to get a good, basic understanding.

Given you confusion I might question why you would choose to record at 96kHz? And then I would ask what word length you are recording at?

But Cubase 5?!?!? Good grief!

[Hugh Robjohns]

In order to work on my laptop, i have to convert/dither the audio down to 44.1khz ....

As the Elf has highlighted, there does seem to be some confusion here -- at least in the terminology if not the actual technicalities -- and we should sort that out first to avoid more confusion when others try to help.

Changing the sample rate is called sample rate conversion and, when raising the rate, oversampling or up sampling, or decimation or downsampling when lowering the rate. Changing the word length (say from 24 bits to 16) is where dithering comes in. Dithering is a means of retaining the wanted audio information when using a shorter word lengths, albeit with a higher noise floor. Sample rate conversion and word length reduction are entirely separate and unrelated functions.

It is possible for me to copy across the Cubase session and mix settings from my laptop to the studio computer, and run a mixdown from there, thus still running everything in the 96khz environment.

With the caveat that I'm not a cubase user, I'd say yes. In principle, your mix settings and automation are just instructions which are applied to the actual audio tracks in the project. Provided that the audio you work on in your laptop is in the same place on the timeline, across the project tracks, and have the same file names, the mix instructions should translate across perfectly.

The exact same concept is employed routinely in off-line video and film editing, where the edit decisions are made with low res video for convenience and speed, while the hi-res pictures are subsequently cut automatically according to the off-line instructions -- a process known as 'conforming'.

What I would like to know is whether it makes any "difference" (apologies for the vagueness of the term!) to the quality of final mix if the audio is dithered down prior to mixing at 44.1khz, as opposed to mixing at 96khz and dithering down the final mix.

If you take the actual audio file produced by your 44.1k laptop mix, it will have a lesser 'quality' than the 96k material (a reduced bandwidth, anyway!)... But if you take the mix automation from your 'offline' 44.1 session and import it into the 96k session to 'conform' the original 96k material, you will have the same 'quality' as if you had mixed in the 96k sessions directly -- because you are, but with pre-defined mix instructions instead of live commands.

As a final question, is there any difference in how a DAW treats audio at different frequency rates? (e.g. more headroom at 96khz?)

No, none at all. Headroom is related only to the audio level of the audio, and how much space you have left between the signal peaks and 0dBFS (clipping). Audio recorded at 96k has, in theory, audio content running up to around 48kHz (if your mics can actually capture anything that high, or your other sources generate it!), while audio at 44.1 has a bandwidth that stops at around 22kHz -- which is still rather higher than almost everyone can hear anyway. But in either case, the DAW treats both the same.

Hope that helps

H