I could be wrong (so maybe i shouldn't chip in), but I think Beno's saying dither a version with no fade at all on the end of the track, then import that dithered track into an audio editor (or tracktion) and create a new version with a fade at the end
Surely that wouldn't audibly quantise, would it?
should I dither ...
-
- KVRAF
- 2417 posts since 17 Jun, 2003
"my gosh it's a friggin hardware"
-
- KVRAF
- 6740 posts since 25 Mar, 2002 from sheffield, england
There is no difference (in terms on the effect on the dither anyway) between a fadeout after the dither but before the render, and a fadeout applied to the rendered file: both will break the dither effect. The rule is: don't do any more processing to the dithered 16 bit file, including applying gain changes.. a fadeout is basically just a gain change. 
When I master a CD I make sure the inidividual tracks have some silence at the start and end, so I can produce a CD with constant dither noise running through the pauses as well as the tracks themselves (I do the same with room ambience when recording acoustic music) as background noise is much more noticeable if it fades to black after each song, and fades back up before the next..
When I master a CD I make sure the inidividual tracks have some silence at the start and end, so I can produce a CD with constant dither noise running through the pauses as well as the tracks themselves (I do the same with room ambience when recording acoustic music) as background noise is much more noticeable if it fades to black after each song, and fades back up before the next..-
- KVRAF
- 2417 posts since 17 Jun, 2003
I think i need to find some time to read that Bob Katz post and maybe some more to get to grips with this properly. Think i'm missing something pretty fundamental ...
"my gosh it's a friggin hardware"
-
- KVRAF
- 6740 posts since 25 Mar, 2002 from sheffield, england
Well, when you import the 16-bit file into an audio editor, it will probably be converted to a 32-bit temp file. If not, when you apply the fade-out, this will be applied to a 32-bit version of the audio (this is what happens if you do it in Tracktion) so the resulting 16-bit file with either be truncated during the fade, or dithered again, which will add another layer of 16-bit dither noise to the whole file, and rather defeat the purpose!chico.co.uk wrote:I could be wrong (so maybe i shouldn't chip in), but I think Beno's saying dither a version with no fade at all on the end of the track, then import that dithered track into an audio editor (or tracktion) and create a new version with a fade at the end
Surely that wouldn't audibly quantise, would it?
-
- KVRian
- 779 posts since 3 Apr, 2003 from UK
I agree with the man with the expensive lug holes. For the sake of a smooth fade, I would fade before dither. The level of the dither is pretty much inaudible anyway, unless your speakers are turned up to ear-buggering volumes.
However, to address the point made earlier in the thread about the end file being MP3, I wouldn't waste too much time worrying about it!
However, to address the point made earlier in the thread about the end file being MP3, I wouldn't waste too much time worrying about it!
-
- KVRAF
- Topic Starter
- 3057 posts since 9 Apr, 2003
It took me a couple passes to get it, but it helps to know at the beginning that LSB = least-significant bit, so the LSB threshold is the voltage threshold required for the LSB to contain a 1 instead of a 0.chico.co.uk wrote:I think i need to find some time to read that Bob Katz post and maybe some more to get to grips with this properly. Think i'm missing something pretty fundamental ...
The random dither noise helps to bump lower signals over that threshold in a way that results in the dithered signal averaging out to the same as the original low-level signal.
So instead of signals below the LSB threshold just being left out (truncated), dithering helps to include them by approximation.
Corrections welcomed!
5 twelve