32 bits & 96.0 kHz .AIFF conversion to 24 bits & 96.0 kHz .FLAC [SOLVED]

Hello everybody,

I read the manual. I googled terms. I searched the forums. I read through everything.
I am still confused.

I just have this 32 bits, 96.0 kHz .AIFF release. It’s 3,5GB big.

I did a macro conversion to 24 bits, 96.0 kHz .FLAC.

After the conversion, I had a 1,38GB big folder of .FLAC files.

“Design:
Audio sources encoded to FLAC are typically reduced to 50–70% of their original size, similar to other lossless formats, though the final size depends on the density and amplitude of the music being compressed, and, with some music, the file size can be reduced by as much as 80%.[4]”

I zoomed in in various parts and compared both versions on the sample dots and on the spectral viewer. I saw no difference.
I listened to both versions in various parts. Parts with high sounds, parts with low sounds. And especially when a song faded in and faded out
at maximum volume. I heard no difference.

Would be dithering required? I read that it would make a difference. But so far I don’t know how to spot any difference.
Am I missing something here?

You wouldn’t hear a difference (in a proper-scientific-blind [u]ABX test[/u]) if you reduced it to “CD quality” (16/44.1) :wink: It’s often impossible (or very difficult) to reliably hear a difference between an original and a good-quality MP3 copy!

As you probably know, FLAC is lossless but it doesn’t support 32-bits so the conversion to 24-bits is “mathematically lossy”.

Would be dithering required? I read that it would make a difference. But so far I don’t know how to spot any difference.
Am I missing something here?

The theory is that you should dither whenever you downsample, and since you’re going from 32 to 24-bits you should dither.

But as a practical matter, you can’t hear dither (or the effects of dither) at 16-bits under normal listening conditions, and at 24-bits it’s doesn’t matter at all. And as you may know, dither is added noise that’s supposed to sound better than quantization noise.

…At 8-bits you can hear quantization noise so I assume dither helps but I never tried it because I virtually never work with 8-bit audio. I have heard the quantization noise… It’s like “regular noise” in-that it’s most noticeable when the “signal” is quiet. But it “rides on top” of the signal so unlike regular-everyday noise it goes-away with pure-silence.

Also if you recorded form an analog source, the analog noise is often greater than the quantization noise or dither noise (assuming 16-bits or better). In that case it’s “self dithered” and the slight-noise added by dither won’t make a difference.




BTW - It’s unlikely that your DAC supports 32-bits. And from what I’ve read, most 24-bit DACS are only accurate to about 20-bits.

Dither is likely to significantly increase the size of FLAC files, but for 24-bit FLAC the benefits of dither are extremely small (some would say “insignificant”). For 24-bit FLAC format, it is arguably not worth the additional file size to use dither. I’d suggest that you export everything twice; once with dither, and once without - listen to both and compare the file sizes, then decide which version to keep.

Thank you guys for getting back so fast.

This kinda verified what I thought.

It’s mainly important to use it from 24-Bit to 16-Bit. Also funny to read about the 8-Bit part. Never thought about dithering stuff like that haha.

I know that I won’t be able to hear this, but I rather have it in this quality than 320 kbp/s. Even though I know that 99% of all people, including me
kinda hear almost no difference @ 192 kbp/s and above.

That is it from my side. I don’t know if threads get closed when everything is done. If so: I got my answers and I am fully satisfied.

Even 16-bit FLAC is more accurate than 320 kbps MP3. (16-bit FLAC, like 16-bit WAV, is CD quality).


Super. I’ll close this topic.