Adobe Audition reduces the file size when saving.

[nick_1903]

Hello.

I am using build 24.0.0.46.

Windows 10 x 64

When I save the FLAC file, it shrinks significantly!

Some information is lost, which means the quality is deteriorating!

The FLAC source file is 16 bit, 44100 hz.

When saving, I also set the FLAC parameters to 16 bits, 44100 hz.

In the program settings, 44100 hz is also set.

For the experiment, I upload only one VST3 plug-in to the Effect Rack (FabFilter Pro-Q 4 (flat default)), enable bypass mode on this plug-in, click Apply, and save it in the same format. The result: more than 5% of the information is lost. The source file is 36.1 MB, and the saved file is 34.2 MB. On large 24-bit files, up to 20% of the information is lost!

Is this a flaw in the program?

How can I fix this?

[SteveG_AudioMasters_]

when you do your FLAC save, what are the Format Settings?

You can see what the options are if you click on the 'change' button.

[nick_1903]

[SteveG_AudioMasters_]

As far as the quality is concerned, have you tried setting the Encoding to 24-bit PCM?

[nick_1903]

Now I noticed that when saving, the bitrate also decreases. This is very bad. Where does the information disappear to? I was just doing an experiment., Format Settings: set the value to 24 bits. When saving, the file has increased, but not all hardware works with 24 bits. When the sample type is set to 24 bits, the file is reduced when saved. I tried using the built-in Adobe Audition plugin (30 EQ bands (+-0db, default setting)). When saving 16 bits with a frequency of 44100 Hz, the file size decreased. How to prevent information loss?

[SteveG_AudioMasters_]

A few things to note: You have to bear in mind that the 'L' in FLAC stands for Lossless. It works by looking for runs of the same number in the code, and if, say, you had nineteen 8's in a row it would code this as something like 8^19, which is rather shorter than 8888888888888888888, and so represents a coding gain - which makes the files smaller. But when it's expanded again it comes back just as it was before encoding. So no information is lost, and you can check that quite easily with Audition. Secondly, the implementation of FLAC in Audition isn't determined by Adobe - it's entirely controlled by LibSndFile, who provide the coder. Is it a complete implementation? I don't think so - I believe that it's capable of more than the LibSndFile encoder is capable of - like I think the standard copes with 32-bit files and faster bit rates.

[nick_1903]

Ok.

I'm loading a LOSSLESS encoded file into Adobe Audition. After saving, the file becomes smaller.

Assuming that Adobe Audition doesn't lose information, does that mean the original file was encoded inefficiently?
Does the encoder built into Adobe Audition encode the file as efficiently as possible?
How can I check with Adobe Audition that no information is lost?

[SteveG_AudioMasters_]

Can't answer the questions about efficiency, but the answer to the check is simple. Open the file you started with, and the FLAC save. Control-A on either of them, and copy the entire file to the clipboard. Open the other file, and with the cursor at the start (or another Ctrl-A) then Mix-Paste (select Overlap-(Mix) and Invert Copied Audio) the clipboard to it. If you end up with a flat line, then the files are identical - you've just subtracted one file from the other.

I just tried this with a FLAC-save of a file and got the flat line, so they appear to be identical.

[nick_1903]

Hello!
I also did a small experiment with several files.
And I always got straight lines.
This means that the source lossless files are encoded less optimally than the encoder built into Adobe Audition does.
Adobe Audition works great!

I have another small question. I have a fairly modern PC (Win 10x64, Intel Core i7-14700K, MSI Pro Z790-A Max, 32444 MB (DDR5 SDRAM), Samsung SSD 980 PRO 1TB). Audition, and all related files, are located on the SSD. The system is configured for maximum performance (no power saving).
It seems to me that I did not make any settings to make the program work faster, because the file processing time, using four VST3 plugins, seems long to me.
For example, a FLAC file of 35.8 MB, 44100 Hz-16 bit, is processed in 59 seconds.
FLAC 118MB 96000Hz-24bit, processed in 2 minutes 35 seconds.
When there are many files, processing takes a long time.
Is it possible to make any settings in the program or system for faster file processing?