#385 Gross Bugs Found

[Steph Kerman]

I have encountered a number of repeatable and non-repeatable failure modes in both the encoding and decoding processes, in which the program either self-aborts during encoding or runs to completion, producing an error message and a FLAC file which does not decode properly. This has occurred with a significant number of different WAV files, probably more than 1 in 10 . The problem WAV files were produced by Audacity and have been kept. The problem FLAC files have also been kept.

Repeatable cases produce identical defective FLAC files each time. Non-repeatable cases may fail the first time and succeed on a second attempt. I don't believe insufficient disk space is a factor. The program is reading from and writing to a drive other than C: which has tens of GB of space. The space on C: is small enough that most likely it would cause them all to fail if the program needed space on C: for work files. Most likely if insufficient space were the cause the program would not produce identical defective files days apart.

The details exceed what can practically be reported through this system. If anyone wants the details and the original WAV files for analysis I can provide them. However the WAV files are typically about 3.5GB, too large to conveniently transmit even by uploading to a server. That would take a considerable amount of time even over a cable TV HSIA connection. Perhaps DVD media through the mails is more practical.

Because I don't trust the encoding verification process I have been decoding each FLAC file and comparing the result byte-for-byte with the original and then deleting both WAVs if they matched.

Contact me directly if you wish to correct the problems.

[Erik]

Two questions:

• Does the problem still exist with FLAC version 1.3.0 available from https://xipg.org/flac ?
• How do I contact you to get real details of this bugs you are seeing?

I will close this in a month if I don't get a response.