Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Question About how Audacity Dithers Exported Audio (Read 1577 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Question About how Audacity Dithers Exported Audio

Hi everyone,

First off, yes, I'm aware that Audacity has its own forums, but I can't register for the site because I never receive the authentication email.

I would like to know exactly how to avoid exported audio from being dithered. Here is the Audacity wiki page explaining how their dithering works...

https://manual.audacityteam.org/man/dither.html

Now I recently edited audio I ripped from a CD (therefore 16 bit / 44100 Hz / Stereo) and encoded to FLAC Level 0 via Exact Audio Copy. I kept the settings the same as in that wiki article (therefore imported at 32 bit / 44100 Hz / Stereo). I edited the audio tracks in various ways...

-Cut and Paste audio from one track to another
-Deleting audio
-Adding a fade out
-Inserting a clip of silence
-Creating a new Stereo track and pasting audio to it, deleting portions, then copying the audio back to the original imported track
-Latency Correction

I then export the track to 16 bit WAV format, to keep it lossless. However, according to this wiki article (If I'm understanding it correctly), because the audio is imported as 32 bit, rather than the 16 bit it actually is, it has to downscale the audio when exporting, therefore dithering it. I figured changing the Default Sample Format to 16 bit would prevent this because the article says this...

Here's the exception to the rule: If you have recorded in 16-bit and are only doing simple editing (cut, delete, paste, trim...) and not doing any processing (amplify, equalize, frequency filter...) then for highest accuracy dither can be set to "none". In this case, because there are no 32-bit operations prior to export there is no benefit to using dither. Exporting a 16-bit track to 16-bit with dither set to "none" will be lossless. The same applies if exporting from a 24-bit track to an uncompressed 24-bit file format with dither disabled.

...but then the wiki article also mentions this...

Downsampling also occurs when processing a 16-bit or 24-bit track because Audacity processes in 32-bit float format, which is then converted (downsampled) back to the 16 or 24-bit track format. Repeated downsampling can be avoided by working with 32-bit float format tracks (default), thus avoiding unnecessary conversion losses. Dithering is never applied within a 32-bit float project because no downsampling occurs.

So I'm confused. I'm only working with 16 bit / 24 bit audio, so what exactly would I have to do to ensure whatever I export to 16 bit WAV remains entirely lossless, with the exception obviously of what I purposely edited?

Re: Question About how Audacity Dithers Exported Audio

Reply #1 –
There is a quality preference where you can set dither to "none".

BTW - Audacity has their own forum...

Re: Question About how Audacity Dithers Exported Audio

Reply #2 –
The manual says that there are 2 kind of edits.
The cut/copy/paste/trim don’t change the sample itself.
As long as you stick to this kind of edits, no need to use float.
The moment you are changing the value of a sample e.g. a filter, a fade-out, there you profit by using float to keep the rounding error (inherent to using integers) down.
As PCM audio is 16 or 24 bit integer, in the end you need to convert from float to integer and apply dither to decorrelate the quantization error from the music.
At 24 bit you might wonder if this make sense as no DAC can resolve -144 dBFS.
At 16 bit  it is a must.
TheWellTemperedComputer.com

Re: Question About how Audacity Dithers Exported Audio

Reply #3 –
The cut/copy/paste/trim don’t change the sample itself.
As long as you stick to this kind of edits, no need to use float.

Alright, thanks for the info. Performing latency correction would also fall under "copy/paste/trim" type of edits correct?

Re: Question About how Audacity Dithers Exported Audio

Reply #5 –
Quote from: Damned536 on
So I'm confused. I'm only working with 16 bit / 24 bit audio, so what exactly would I have to do to ensure whatever I export to 16 bit WAV remains entirely lossless
Your premise is false, because the process of going from a higher bit rate to a lower one unavoidably introduces changes to the resulting audio.

And dithering is specifically designed to minimise the audible impact of that.

Unfortunately that process involving the introduction of "noise" may initially sound ominous and damaging, but my advice is not to concern yourself about it and follow standard practice of dithering.  And while you may be able to detect incredibly subtle differences using different dither algorithms, I believe your focus and time is much better spent elsewhere that can materially improve your results.