Perian Breaks CRAM Support

[anonymous]

With Perian installed, QuickTime can no longer understand videos encoded with CRAM codec. Once I remove the Perian component, QuickTime supports them normally.

[astrange]

Sample?

Any avi file which works with Quicktime already is using an ancient video codec, and CRAM certainly is ancient. It can't be very widespread, but it should be fixable.

[anonymous]

Replying to tick: I've witnessed this problem also, so I've attached a small sample CRAM avi file fragment that does NOT play when Perian is installed (but does play when the Perian component is removed). The attachments in this defect tracking system are only allowed to be 256k, so I made this 256kb fragment by using the unix split command on an original 19mb file. I realize just "split"-ing it will cause problems in practice for players, for instance the QuickTime Player window going blank after playing the first second or so of video, or "access_file: seeking too far" errors in VLC. However, it looks like all the headers and the defect behavior are the same as the original avi file, so the "split" file fragment should be usable for debugging.

[Ritsuka]

QuickTime's cram decoder (aka microsoft video 1) works only if the video's fourcc is "msvc". But actually most avis have "CRAM" as fourcc, and QuickTime's decoder happily ignores the video.

[anonymous]

I'm confused, the defect says that natively QuickTime DOES play this file, right? It's only after installing Perian that it stops working, that is certainly the the behavior we have observed in the past. Anyway, VLC definitely identifies the codec in the attachment as "CRAM". However, the string "MSVC" also just happens to appear in a hexdump of the attachment at 0x70, but I am not sure if that offset parses out to a fourcc header or not. Does anyone know an easy way to get QuickTime Player debug info?

[anonymous]

I can confirm this. MS Video 1 encoded avi files play fine with straight vanilla Quicktime, but if you drop Perian in, they break (white screen).

[mjhsieh]

This is pretty much included in Ticket #123.

[gbooker]

[752] fixes the sample here.