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: Could FLAC (in my case 1.2.1) ever become obsolete? (Read 15071 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Could FLAC (in my case 1.2.1) ever become obsolete?

Reply #25 –
We are talking about whether FLAC would EVER become obsolete...when dealing in such conversations price should not be an issue. If you are going to store something FOREVER, you generally don't worry about price, you worry about fidelity. MP3 was and is great to listen to music on the PC or a small device...but I would never have stored anything in it for the long haul...that's why I have like 50-100 100-place CD-R bundles in my storage that I haven't gotten around to re-copying yet.

That's all I meant about why the hell did you rip to mp3_128.

Could FLAC (in my case 1.2.1) ever become obsolete?

Reply #26 –
Quote from: bmcelvan on
We are talking about whether FLAC would EVER become obsolete...when dealing in such conversations price should not be an issue. If you are going to store something FOREVER, you generally don't worry about price, you worry about fidelity. MP3 was and is great to listen to music on the PC or a small device...but I would never have stored anything in it for the long haul...that's why I have like 50-100 100-place CD-R bundles in my storage that I haven't gotten around to re-copying yet.

That's all I meant about why the hell did you rip to mp3_128.

Hope you have contingency plans for disc rot.

Could FLAC (in my case 1.2.1) ever become obsolete?

Reply #27 –
Quote from: bmcelvan on
that's why I have like 50-100 100-place CD-R bundles in my storage that I haven't gotten around to re-copying yet.


That's among the absolute worst ways to archive music.
 

Could FLAC (in my case 1.2.1) ever become obsolete?

Reply #28 –
Since thread is ressurrected:

Quote from: Beatswork on
Obsolete is not the same as unusable.


This.

Lossless storage can be losslessly converted.  Is FAT obsolete?  Well you can anyway read files stored on a FAT drive, and copy them to something else.
Actually the losslessness is likely to make it obsolete much faster - because you do not need it, you can dispose of it.  We are "never" going to get rid of MP3.  Heck we aren't even getting rid of MP2 or (lossy) WMA.  Maybe MP3PRO and Real Audio.

That said, I would not write off the OP as "paranoid" just because of asking the right questions.  What about http://en.wikipedia.org/wiki/Pentium_FDIV_bug ?  What if some format had a build for that architecture, which worked symmetrically in the sense that everything was losslessly stored provided decoded with the same bug - i.e., decoded signal = original?  I have asked here before, and been told that FLAC recycles quite a bit of code between encoding and decoding - increasing the risk that an error could enter symmetrically. 

But, gut feeling risk assessment: if you one day find your FLAC files unusable, then it would be orders of magnitude less likely to be because you all for sudden find yourself in need of a new computer and an OS with FLAC decoding support is nowhere to be found for cheap - than it being because your storage medium is inaccessible (your drives crashed, your cloud provider went bankrupt, you find your second cloud provider in a jurisdiction taken over by the MAFIAA lobbyists, refusing you access to your music unless you show up in person with the original CDs ...).

Could FLAC (in my case 1.2.1) ever become obsolete?

Reply #29 –
Quote from: Porcus on
What about http://en.wikipedia.org/wiki/Pentium_FDIV_bug ?  What if some format had a build for that architecture, which worked symmetrically in the sense that everything was losslessly stored provided decoded with the same bug - i.e., decoded signal = original?  I have asked here before, and been told that FLAC recycles quite a bit of code between encoding and decoding - increasing the risk that an error could enter symmetrically.

That is extremely unlikely to go unnoticed in a format like FLAC.

1.  Why would you code expecting a bug like that?  As soon as FDIV was discovered Intel offered to replace the affected CPUs.  Workarounds were made, but to not use FDIV.  No one used the broken FDIV on purpose. 

2.  Assuming that we are back in the 90s when the Pentium was the top of the line, there were still lots of 486s, AMDs, Cyrix CPUs, etc which didn't have the bug.  If FDIV had affected FLAC it would have been caught rather quickly when someone tried to decode on a non-Pentium and the checksum failed.

3.  The same thing as 2 would happen if a new CPU came out with a serious bug that affected FLAC.  All the old CPUs would start to fail the checksum. 

Could FLAC (in my case 1.2.1) ever become obsolete?

Reply #30 –
Quote from: Porcus on
That said, I would not write off the OP as "paranoid" just because of asking the right questions.  What about http://en.wikipedia.org/wiki/Pentium_FDIV_bug ?  What if some format had a build for that architecture, which worked symmetrically in the sense that everything was losslessly stored provided decoded with the same bug - i.e., decoded signal = original?  I have asked here before, and been told that FLAC recycles quite a bit of code between encoding and decoding - increasing the risk that an error could enter symmetrically.


Lossless codecs don't generally depend on floating point math being accurate since its only approximate.  Instead they use integer math for decoding, which is exact and identical across all machines always.  FLAC is this way.  The decoder depends only on integer operations and is therefore exact on all types of computers.

Could FLAC (in my case 1.2.1) ever become obsolete?

Reply #31 –
Quote from: washu on
1.  Why would you code expecting a bug like that?  As soon as FDIV was discovered Intel offered to replace the affected CPUs.  Workarounds were made, but to not use FDIV.  No one used the broken FDIV on purpose. 

2.  Assuming that we are back in the 90s when the Pentium was the top of the line, there were still lots of 486s, AMDs, Cyrix CPUs, etc which didn't have the bug.  If FDIV had affected FLAC it would have been caught rather quickly when someone tried to decode on a non-Pentium and the checksum failed.

Probably not. First of all, Intel was relunctant to replace the CPUs, as the use cases in which this bug manifested were quite rare. Wikipedia quotes Byte magazine (the source is no longer available), saying that 1 in every 9 billion divides would suffer from this problem. Therefore, it seems unlikely that faulty files would be traceable to a CPU bug, because it is too rare.

But then again, FLAC only relies on floating-point math for finding the best predictor, while the process of subtracting predictor from the audiodata and coding the residual is a process fully done in integer math indeed.
Music: sounds arranged such that they construct feelings.