cTiVo was just grabbing the first match in the latest 180 bytes from the encoder output, and since my script was only outputting 5 or so bytes every 5 seconds, cTiVo kept getting the first match in the first 180 bytes (0%) and thought it was stalled. I updated my script to print updates every second instead of every 5 seconds.There was a leading space before the regex which I copied from HandBrake, works for Handbrake since the % is part of a long status string but not for my script.HandBrake seems to work fine and it does have a lot of other output that would make this less likely to happen? But I tried changing the regex to (\d+.\d+) % to no avail. Maybe this is just because of cTiVo grabbing the last 180 bytes of the encoder output and it just happens to get cut off after the decimal point. it says 30% even though the script is only at 5%. The very first line of data after "Encoder report:" looks weird. So aside from the apparent bug where cTiVo doesn't kill the encoder when commercial skipping is enabled, the issue at hand seems to be a matter of cTiVo just not being able to match the regex. 02:09:56:027 DL status of Austin City Limits: Beck (Living Room) to Failed (16)
#Ctivo zero byte download#
02:09:56:027 at Encoding, canceled download of Austin City Limits: Beck with progress at 0.000000 with previous check at 08:07:55 +0000 02:09:55:950 City Limits: Beck cleaningup files 02:09:55:950 0.000001 seconds for write data to complete during cancel 02:09:55:950 of Austin City Limits: Beck 02:09:55:950 stalled at 0.0% rescheduling show Austin City Limits: Beck 02:09:55:567 progress with Rx failed for task encoder for show Austin City Limits: Beck
If I turn off commercial skipping, cTiVo comes to terms with the fact it can't figure out the script's progress and kills it: If I download the same video and run archive_tivo.sh on it standalone, I don't see any 0.00% except at the very beginning.ĬTiVo only leaves the script running after failing if commercial skipping is enabled.
In the time since I posted those scripts I found the same thing you did it could output 0% when there were errors in the input, so I increased the tail to 1000 characters and no longer notice it jumping back and forth between 0%. You're right, the -strict -2 is required when encoding aac with the ffmpeg binary bundled with cTiVo. I'm using the latest ffmpeg in homebrew, 3.1.1, though I get the same result using the ffmpeg binary bundled with cTiVo. I haven't come across any issues with it skipping edl sections or failing yet, I've tried it on 3 different OTA broadcast types so far. The option is actually -edl (two hyphens) or -e rather than -edl with a single hyphen like the mencoder option, sorry for the confusion. (ffmpeg should not have continued on by itself.) If so, sounds I've got two things to fix one to reset that timer on each phase during non-simultaneous encoding, and two to actually cancel background tasks when we cancel the subtask within ctivo. So, one workaround is to write "1.00 %" to the output stream immediately upon the start of your script.Īnother to increase the MaxProgressDelay preference to a much bigger number than the default 120.ĭefaults write MaxProgressDelay 1200 start of commercial-ing), then it would immediately fail even though there has been progress since then. If the prior progress check also happened to be in a period where progress was zero (e.g. So, if your script takes a little while to run, it would keep the encoding progress at zero for a little while. Specifically, we reset progress to zero at the launch of each phase, but we don't reset that timer. That works great if the encoding is simultaneous with the downloading, but I think it has a chance of failing if not.
#Ctivo zero byte update#
Because of occasionally crashy/flaker sub-binaries, cTiVo notices if it doesn't get a progress update every two minutes. Then I tried a different video and it failed. awk says "can't open file dl"Ĭ) At least my copy of ffmpeg complained that " experimental aac requires -strict -2"įinally got it to launch, and then it seemed to run fine (although it only seemed to clip the first EDL line), with progress reporting just fine. Trickiest one to figure out was apparently -edl option doesn't work has to be -e.
#Ctivo zero byte how to#
Just reading took a while.ī) It then took me quite a while to create a matching Format, and then figure how to debug, where the output was going etc. serious script-fu some nice tricks in there. So, I've spent a lot of time on this, and I think I've figured it out.įirst a few notes on the script ( I worked with the archive_tivo_one_ffmpeg.sh one)Ī) wow.
0 kommentar(er)
