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: 24/96 to 16/48 - Resample and Dither Order (Read 5464 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

24/96 to 16/48 - Resample and Dither Order

I've never really worked with high resolution audio, but recently tried something from HDTracks out of curiosity to see if the masters they get are any better than what I already own.

I have no intention of leaving the files at 24/96 for playback, so went about reading about and using SoX to do the conversions.  While tinkering, the thread question came to mind.  I tried changing the order (resample to 48KHz then dither to 16 bit, dither to 16 bit then resample to 48KHz) and there was no audible difference.  That's unsurprising to say the least. 

I did note that SoX does not support high quality noise-shaped dither at 96KHz.  For that reason alone I suspect I'd be best served by resampling to 48KHz before dithering to 16 bit.  But there are many people here far more knowledgeable than I on this subject, so I was wondering if there is any preferred order or "best practice" to doing something like this.

24/96 to 16/48 - Resample and Dither Order

Reply #2 –
To minimize quantization errors, do everything possible before reducing bit depth. Apply dither and reduce to 16 bits as the last operation.

24/96 to 16/48 - Resample and Dither Order

Reply #3 –
Thank you.  That makes sense and is exactly what I wanted to know.

24/96 to 16/48 - Resample and Dither Order

Reply #5 –
As an alternative, how about trying lossyFLAC instead of dithering into 16bit?
Probably typical average of wasted bits will be 8 or so even at very high quality setting of lossyWAV, and the result will be effectively compressed as if it were 16bit, but you still have theological 24bit dynamic range.

24/96 to 16/48 - Resample and Dither Order

Reply #7 –
Quote from: skamp on
What religion are you referring to? 

 
I didn't mean "unlikely" by that wording. It WILL actually have wider dynamic range unless resulting wasted_bits is CONSTANTLY greater than 8.
I might have meant "indistinguishable to human ear", though. Engilish is difficult to me.
 

24/96 to 16/48 - Resample and Dither Order

Reply #8 –
Quote from: nu774 on
As an alternative, how about trying lossyFLAC instead of dithering into 16bit?
Probably typical average of wasted bits will be 8 or so even at very high quality setting of lossyWAV, and the result will be effectively compressed as if it were 16bit, but you still have theological 24bit dynamic range.

That's an idea.  I've played with lossyWAV before and thought it was pretty neat.  I should tinker with it on these files.  That being said, I think lossyWAV would be more lossy than just converting to 16bit with SoX.  I could be wrong about that, but I don't think I am.  I doubt I'd hear any difference, but I'm just trying to get these files to be more or less like the rest of my CD collection on my NAS.  Maximizing compression isn't the primary goal, otherwise I'd use lossyWAV on all my stuff.