• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Lossless file conversion question

PRL

Member
Joined
May 25, 2020
Messages
23
Likes
11
I've been reading the current MQA thread and I am having some trouble understanding how any argument can be logically formed that something is identical to the original when it clearly is not.

I didn't want to be seen as derailing or being off-topic, so I have posted my reality-check question here.

If you have any original file as a .wav, DSD, or whatever, and you convert the original file to a lossless file format, such as FLAC, I am correct in thinking that you can then convert this file back to it's original format and it will be identical digitally to the original file and not just indistinguishable from the original when compared as an analogue file?

Lossless means digitally lossless, doesn't it?
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,380
Likes
1,311
FLAC does not support DSD and the conversion will not be lossless, as long as the FLAC file's encoded sample rate is not identical to the original DSD file (at least 2.8224MHz).

If one needs an efficient mathmatically lossless DSD compression format, use WavPack.
https://www.wavpack.com/
 

voodooless

Major Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
2,371
Likes
3,542
Location
Netherlands
If you have any original file as a .wav, DSD, or whatever, and you convert the original file to a lossless file format, such as FLAC, I am correct in thinking that you can then convert this file back to it's original format and it will be identical digitally to the original file and not just indistinguishable from the original when compared as an analogue file?

Wav, yes, DSD, no. DSD is no PCM data, so cannot be compressed in a FLAC container. Well, you can theoretically put in any data in a FLAC container maskaraded as a PCM file (in the MQA thread I've made the example of a Word file), but it just would not compress very well.

Lossless means digitally lossless, doesn't it?

That depends on which marketing department uses the term ;)
 

abdo123

Major Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
4,184
Likes
3,736
Location
Brussels, Belgium
why did you make the thread if you have your answer already?

MQA reduces dynamic range in high frequencies (because no real recordings need 24-bit of resolution at 40KHz) so you would have the same exact content (allegedly) with significantly reduced file size.
 

voodooless

Major Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
2,371
Likes
3,542
Location
Netherlands
Well, you can theoretically put in any data in a FLAC container maskaraded as a PCM file (in the MQA thread I've made the example of a Word file), but it just would not compress very well.

Essentially this is what MQA partially does, just a bit more clever. The upper bits of each sample contain the PCM encoded audio below the subsonic, and the lower bits is noise encoded data consisting of some metadata and the compressed version of the subsonic content.
 

PierreV

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
1,257
Likes
3,852
Lossless means digitally lossless, doesn't it?

Expect constant nitpicking and endless discussions, especially now that MQA has changed its lossless tune to better than lossless.

There is bit losslessness on arbitrary amounts of data, on storage or in transfer. That applies, clearly, to any digital technology.
If you had an Intel based automated nuclear missile launch system in the 80ies, flipping one single bit in those opcodes might have launched WWIII
01110100
01110101
More trivially, that type of bit flip also made a copy-protected game run or not run at all. Or a partition not to be bootable. Or a colored led to light up, etc...

But we aren't talking about that in most cases in audio.

We are concerned by the losslessness of a signal of limited bandwidth as covered by the sampling theorem. That is what was considered _lossless_ around here until very recently :) If one properly samples a band-limited signal there is no ambiguity: there can only be one signal that results in that sample (but many eventually time-shifted samples that can represent the same signal). In theory, the CD format has the abilities of our biological audio system covered in terms of frequency response. There are practical issues to that mathematical rigor, such as filters, noise in the process, and so on. This means that having a bit of margin doesn't hurt.
As far as dynamic range is concerned, one can argue that the CD format doesn't cover that as well as it does cover the frequency range and that this can be made obvious by amplifying what should be silence by a significant amount.

That is why quite a few people would have loved 48kHz (essentially more margin for filters) and 24 bits (more dynamic range) to be on the "safer" side.

I find the current discussions about what lossless means/is to be a very slippery slope... People seem willing to say almost anything to win an argument.
 
OP
P

PRL

Member
Joined
May 25, 2020
Messages
23
Likes
11
Thanks for the replies. As ever everything can be much more complicated than you think, when you get into the details.

Abdo123, I don't follow your question. I asked a question and people responded with information that I didn't know. I didn't have the answer to my question in advance.

For the record, I don't do MQA and don't plan to, I would just like to know more about how it works/doesn't work.
 
Top Bottom