[Peter Larsen[_3_]]

On 20-01-2016 18:48, Tim Sprout wrote:

128 barely tolerable for spoken word.

It is actually less tolerable for spoken word than for music. The
trick to use with soundcloud and similar services is to convert to 32
kHz sample rate 16 bit properly dithered before upload and upload in
full wordlength and leave bit-reduction to them.

Peter Larsen

Thanks for this. To clarify my understanding, use 32 kHz sample rate
during DAW mixdown

No! - stay at whatever your house sample rate is. What you upload to
Soundcloud should be mastered for Soundcloud.

because it is a multiple of 128 kpbs which Soundcloud
uses,

Those are incommensurable concepts. No. You downsample AFTER mastering
to get rid of the treble that perceptual coding is likely to replace
with white noise anyway. By doing that you avoid wasting wordlength on
it and get a cleaner sound with less splatty treble.

convert from the higher internal DAW 24 or 32 bit resolution to
the lower 16 bit resolution Soundcloud uses using dither to add
desirable noise to mask less desirable noise, and upload the full 16 bit
wordlength to Soundcloud?

Unless you master for dvd audio a mastered file should in my
comprehension be 48 or 44 kHz 16 bits. Some may know this better than I
do, corrections will be appreciated if I am wrong.

Tim Sprout

Kind regards

Peter Larsen