QUOTE(Mr_Rabid_Teddybear @ Feb 14 2004, 12:42 PM)
But I couldn't get it to work with 3.90.3, ran into same problems as mentioned above, wondered if lame had a pipe problem, tried MPxchange, Multi Frontend, UniversalFront & Frontah as frontends, tested with 3.90.2, 3.90.3, 3.92, 3.93.1 & 3.95.1 for lame versions. Ofcourse the problem was using 1.95e for mppdec. When changing to mppdec 1.93j (after finally finding this thread) it worked with all these lame versions, and I can use 3.90.3 or 3.95.1 as I like.
Intresting observation though: mppdec 1.95e had a broken pipe with all these lame versions, exept 3.95.1 (latest rarewares compile) with which it seems to work allright.... Wonder why?
(by the way, off topic: have "--alt-preset *whatever*" now become "--preset *whatever*" in 3.95.1? Is it the same?)
Furthermore: If latest mppdec versions do have a broken pipe --- why isn't a new bugfix version released (it's a long time since last version came out), why use an older version as permanent (?) problemsolving, and why isn't this problem mentioned on Case's website, with a link to download 1.93j....?
It is just another bug of Lame.
Lame don't read the data which mppdec writes.
It is the 4 GByte bug of Lame. The length of the initial RIFF header is set to
MAX_UINT to enable processing of files between 2 GByte...4 GByte for pipe
processing (older versions used MAX_INT, this limits pipes to
3 hour 6 min 24 sec, the movie "Das Boot" is some minutes longer).
Lame interprets these length as negative number and does then a lot of crazy
stuff.
I have here 5 or 6 recordings with 3:06:24...6:12:48 hours recording time.
For lengths larger than 6:12:48, 64 bit support is necessary.