can i resume encoding with real producer after system crash

[wata]

i mean resume encoding from the point before it crash or maybe a few sec earlier

if no, can this feature be implement, i see that producer seem to be writing some temporary file during the analysis

[Sirber]

Producer also use a lot of memory :) Fill a bug report in Helix Community.

[wata]

i notice that producer is using total 250meg too, but still ok for me
but after going thru 18hr and 10% more left to encode, the system crash or something else happen (accidentally pressing restart after some driver/software install), it a pain to start all over again.

[Sirber]

I understand, but life is life :).

[phrentec]

if they can implement a way to make the tmp file readable while producer is writing to it, that would be also great.

tip: to properly stop the encode in producer.exe use ctrl+c .

[ookzDVD]

Is that possible to re-run the 2nd pass only with
the previous 1st pass ?

[Sirber]

Presently, no. I think that most of the 1st pass info is stored in memory.

[ookzDVD]

Originally posted by Sirber
Presently, no. I think that most of the 1st pass info is stored in memory.

I think they should think how to put the 1st pass info into .log file since it will save memory and much more useful for analysis and 2nd pass encoding.

While enabling the EHQ and if I oversize the target size I don't have to restart the both pass again.

-- off topic
Btw, is that possible to "cut" oversize .rmvb file with any utility ?
Sorry for asking this I really newbie in the RV9, I was playing with
XviD :) But after I realize that the XviD is really "b@d" encoding
the dark movies, I start to learn the RV9.

[phrentec]

I'm not sure if this is related but in rmeditor.exe it shows up

log file - the path to the log file.
dump file - the path to the dump file.

what log file? where?

[karl_lillevold]

the 1st pass info is stored in a file, but there is presently no mechanism available to restart just the 2nd pass. This is mainly because the current design of the 2-pass mechanism requires the exact same settings to be used for the 2nd pass. Any variation in the pre-filters applied or bitrate requires the 1st pass info to be re-created. So there is room for improvement here, and one idea would be to have an option to restart just the 2nd pass, with no change in any parameters allowed.

P.S. the seemingly large amount of memory required is mainly due to a memory leak problem in Avisynth / directshow, that results in memory being doubled during the 2nd pass. I tried to figure it out via a discussion in the Avisynth dev forum, but was not able to narrow it down more than a problem with directshow

[midiguy]

That's weird because I only have 128 mb SD RAM (pc 100) and I can run producer okay...

[karl_lillevold]

that's no problem, because all the memory that was allocated by Avisynth for the 1st pass, is no longer used for the 2nd pass, and is then swapped out to your virtual memory file.

[RadicalEd]

Originally posted by ookzDVD
is that possible to "cut" oversize .rmvb file with any utility ?

In the producer directory you'll find rmeditor.exe :0 you can use -s DD:HH:MM:SS.mmm to specify a start time and -e DD:HH:MM:SS.mmm to specify an end time to cut at. To join, you can just set multiple inputs and a single output. It seems to work pretty well.

[Nic]

Probably a bit OT: But I was encoding using producer.exe, when it got 78% through the file and then just seemed to exit back to the C prompt without giving an error message ??
The last warning given was:
Warning: Stereo signal out-of-phase since 31 seconds.
What does that warning mean exactly? Could that have caused the problem?

The machine I was using had 8 gig free for the Temp file (Why does it create the file in the temporary directory before transfering it to the desired folder?) The file I was trying to create should have only come out at 250mb however. (the destination drive had 60gig free).

It's a P4 2.4 with 512mb memory.

Anyone got any clues as to why this might have happened? (it was rv9 at 256kbits vbr, and audio flavour 21) The audio was a good bit longer than the video? Could this have caused problems?

Cheers,
-Nic

[karl_lillevold]

@Nic: Such a problem indicates a crash in a location where Producer's exception handling does not pick it up, i.e. an assembly optimized module. This should never happen, of course, and is therefore rather serious. I will send you my e-mail address so I can hopefully get some more data about your source material, job/audience file etc. This is something i'd really like to look into, and see if I can reproduce.

"stereo signal out-of-phase" should not matter, it's a warning from the audio module that the signal appears to be out of phase, intended mostly for live capture and encoding, where the user might perhaps have switched the polarity of one input. Signals that trigger this occur naturally once in a while, but not too frequently.

Much longer audio than video: this should not matter, but it's not tested much, so I am a little worried this might be the cause of the problem.

Temp location: in producer.pref you can specify your temporary directory. If you set this to be the same drive as the destination drive, there should be no transfer, just a quick re-naming.

[Nic]

Thanks for the interest :) , my email address is nic at nic.dnsalias.com. If you mail me, ill mail you back tomorrow with the d2v file, avs file, rpad etc. I do wonder if it was because the audio was longer than the video (I was using MPEG2Dec3 and and WavSource, maybe there was a problem in AviSynth 2.5 when AudioDubbing audio thats considerably longer than the video ? )

I was trying to convert the Matrix (region 2) to a 264kbit file (should come out at 250mb). Purely because I was playing at work today ;)

If you need anymore detail let me know. Ill still have all the output on the screen (so ill copy and paste that to you as well)

Thanks alot,
-Nic