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: FLAC 24/96 to 16/44.1 (Read 5344 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

FLAC 24/96 to 16/44.1

I have 24/96 FLAC tracks and I want to dither it to 16/44.1. Will doing so reduce the SQ signicantly, and will it still be "lossless"?
sin(α) = v sound/v object = Mach No.

FLAC 24/96 to 16/44.1

Reply #1
Depends on if you have 'bat ears' or not. In your particular situation, I would recommend doing the downsampling on a track or two and then doing an abx test.
A highly publicized test already proved that formats like sacd and dvd-a provided no conclusive evidence that such formats provided any audible benefit. Ie, 16 bit has a noise floor below audibility and the 22.05khz cutoff is inaudible to humans.

FLAC 24/96 to 16/44.1

Reply #2
Quote
provided no conclusive evidence that such formats provided any audible benefit
is hardly even in the conceptual ballpark. The tests found that no one could tell any difference. That doesn't even suggest the possibility of "audible benefit" but rather strongly supports the opposite.

As far as "lossless" goes, the answer is definitely no in terms of the data (it is a one way mapping) but yes in terms of anything one can hear.

Is "SQ" a widely defined abbreviation of something or in-group slang?

 

FLAC 24/96 to 16/44.1

Reply #3
If no one could tell the difference, then what WAS the difference? SQ is sound quality, ie is the quality of the sound degraded in any audible way? The tests I alluded to involved exactly such downsampling, to which, not even 'golden ears' could detect any audible deterioration, hence my post stands, you can safely downsample such tracks with no audible deterioration. The tracks will be 'lossless' to human ears. And probably even to your dog's ears.

FLAC 24/96 to 16/44.1

Reply #4
I assume it stands for Sound Quality  EDIT: Mach-X was quicker.

FLAC 24/96 to 16/44.1

Reply #5
Yes, SQ stands for Sound Quality. I'm not actually in favor of 24/96 because 16/44.1 is already superior to my ears, which makes 24/96 nonsense and wastes drive space. I just don't want to encounter any possible degradation. Nonetheless, thanks for the replies, guys.
sin(α) = v sound/v object = Mach No.

FLAC 24/96 to 16/44.1

Reply #6
It isn't something to fight about, and I apologize for any ruffled feathers, but the point of view is coming from the opposite direction. "not even 'golden ears' could detect any audible deterioration" suggests both sounded as "good" but possibly somewhat different. It certainly leaves the question open. "no conclusive evidence that such formats provided any audible benefit" suggests it may be "better" somehow but not undeniably so.

However, since no one could tell any difference, and thus had no idea which was which, the question of "deterioration" or "audible benefit" doesn't enter the discussion. What WAS the difference? There wasn't any.

Actually, as probably everyone knows, if one turned the volume up to the level where pain and damage would immediately result from the signal, a small amount of additional noise is audible between tracks from the added quantization error. It was specifically stated that the conversion was not optimized to put this noise out of the auditory range. However, this noise was at too low a level to be heard under any circumstances that would let one listen to the music.

FLAC 24/96 to 16/44.1

Reply #7
The tracks will be 'lossless' to human ears.

Ignoring the apostrophes and the possible interpretations for their use, lossless refers to a process that is mathematically reversible.  It does not refer to sound quality, though in some instances sound quality can be inferred.

FLAC 24/96 to 16/44.1

Reply #8
Is there an official, published, definition or is that just common usage? I ask in an inquisitive way, not an argumentative one. I suspect it has to means a two way mapping when applied to compression encodings but I don't see that as a necessity in every aspect of audio.

Commercial interests often get some particular definition inserted into regulations, which have the effect of law for various purposes, invariably to give them an advantage over competitors. Civil law itself is chock full of special definitions that are not in accord with common usage, most often to give the advantage to the government over the individual. Does lossless have a rigid limitation in regard to audio in general or just at Hydrogen Audio?

Mathematically reversible, or a two way mapping, is one thing. There are one way mappings with lose nothing of significance by any rational argument.

Record an LP at 24 bits, input adjusted to say peak -1dB . Convert to 16 bits.. That is certainly a one-way mapping in the narrowest definition but not in a functionally useful way. Nothing of the music captured from the LP is lost.

If one intended to do any non-nonsensical post recording processing (nonsensical such as making frivolous adjustments hundreds of times just to increase the accumulated quantization error to the level where it might become audible) one could convert that 16 bits back to 24 bit without altering the signal, do the processing, and end up no worse than if one had not originally converted to 16 bit. If fact one could do quite a bit of processing at 16 bits without being able to tell any difference except through mathematical analysis, but that is a somewhat different point.

If one took the original 24 bit recording and the reconstituted 24 bit file and applied exactly the same reasonable processing ( remove clicks and pops, decrackling, NR, EQ, whatever "remastering" one can manage without the original tracks, etc..), then converted both to 16 bits, one might conceivably end up with bit identical files. While that could be unlikely, they certainly would be perceptibly identical -- so you lost nothing by storing the recording as 16 bit.

FLAC 24/96 to 16/44.1

Reply #9
I am only aware of the distinctions between these terms as they are typically used both on this forum and amongst people who I regard as being in-the-know.

The word typically used for irreversible changes that are inaudible is transparent.