transfers to TiVo HD abort after 2-3 minutes

[rscottbailey]

I guess I got too happy after the last commit. ;-)

Transfers start as expected, but end after 2-3 minutes (based on a sample size of 2 different 30-minutes-minus-commercials tests). Nothing sticks out in the Wireshark traces, which is not necessarily saying a lot. Connection between Debian tivohmo server and my TiVos is wired CAT5/CAT6, and of course pyTivo transfers the same files flawlessly. Aaargh.

[wr0ngway]

Two possibilities that I can think of:

• the ffmpeg encode options don't work for the TivoHD - check pytivo debug log to see what it is producing and make the tivohmo match it as a test
• Check the QueryContainer response, to make sure the "SourceSize" field seems reasonable - the tivo uses this to allocate receive buffers, and aborts if it receives more data than allocated

[rscottbailey]

Hmmm. The QueryContainer information looks plausible; the duration is correct (2546 [sec] ~= 42 min as expected) and the SourceSize -- presumably bytes -- is 5463919680 or about 5 GB, which seems on the large size to me considering the source file is a 683 MB, 720p .mkv file.

I'll go investigate what pyTivo is doing on the ffmpeg front, since I vaguely recall stuffing in some bitstream constraints all those years ago when I set it up originally... Thanks for the pointers.

[rscottbailey]

Bummer, I'm hallucinating. Maybe I was doing something different in the past, but I'm using default settings in pyTivo now, and (as you probably know ;) ) those defaults are the same as tivohmo's defaults. I'll go stare at the end of my traces for any other hints, without much expectation of finding anything.