x264 encoding stalls before end without any error message

[nautilus7]

Hi, I've run into the same problem twice the last 2 days. I'm trying to encode "The usual Suspects" Blu-ray (MPEG-2) to 720p using x264 r1259.

I 'm using a nVidia GeForce 8400GS and Neuron2's NV tools to decode the stream.

The first time the encoding "stalled" during the 2nd pass, 30-40 frames before the end without any error message. Second time it stalled 250 frames before the end, but on the 1st pass.

The command line is the following:

C:\Encoding\x264>x264 --pass 1 --bitrate 4390 --b-adapt 2 --bframes 7 --ref 12 -
-deblock -3:-3 --partitions all --direct auto --me umh --merange 32 --subme 10 -
-aq-strength 1.0 --trellis 2 --psy-rd 0.8:0.2 --rc-lookahead 60 --no-fast-pskip
--level 4.1 "c:\encoding\input.avs" --output NUL --slow-firstpass
avis [info]: 1280x544 @ 23.98 fps (152585 frames)
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 Cache64
x264 [info]: profile High, level 4.1
[99.8%] 152340/152585 frames, 1.41 fps, 4297.01 kb/s, eta 0:02:53

At this point nothing happens... CPU is @ 0%... Don't know if it's a decoding or encoding related problem. Any advice is welcome.

My system is win7 x64, 3GB ram, Core 2 Duo E6600, GeForce 8400GS. I'm using 32 bit version of all programs (x264, avisynth, etc.)

[GZZ]

its not a x264 issue. Its your frameserver that stopped servering frames to x264 and then it has nothing to do and idle. Try play your avs script in mediaplayer classic and skip to the end, last few min before the end and then let it play. If playback stops before it reach the end, your video properly have errors.

Edit:
Just notice you are using DGIndexNV tool, then your CUVIDServer.exe might have crashed or has been killed or freezed. I have seen that before, but its long time ago. Are you using the latest driver 190.62 and DGIndexNV Beta 10 ?

[nautilus7]

The script is working just fine in mpc. I am using latest drivers (190.62) and beta 9. The updated beta 9 with a newer version of the server. Beta 10 has some changes concerning mobile cards.

Anyway, the server is still running (it hasn't crashed). It looks like it has just stopped serving frames.

[Revgen]

Are you using MeGUI?

DGDecodeNV has had problems with recent versions.

[nautilus7]

No, just the cmd and feed it with a pretty plain avs script.

[neuron2]

Do you happen to have the beta 9 without the server change to try?

I will get this BluRay and try to duplicate the issue. Is it region 1?

Also, please give me your exact script.

[nautilus7]

Yes, it's region 1. Don't think it is available in other regions anyway. My avs script is:

DGSource("usual.dgm", deinterlace=0, use_top_field=true, resize_w=1280, resize_h=544)
LSFmod(strength=90, Smode=5, Smethod=3)

Cropping is done by your application.

No, i don't have beta 9 w/o the server change. Also i am testing using beta 10 now. Will report tomorrow.

EDIT: just recalled something. I initally did the 1st pass with beta 9 w/o the server change and then i updated to beta 9 w/ the server change. 1st ok, but 2nd pass stalled 30 frames before end. Then i did 1st pass again (i changed my x264 settings) and the encoding stalled again. So, it's something caused by the server changes i would say.

[woah!]

i get this reported problem above aswell, mainly with sources needing deinterlacing/resizing in DGSource. i also get watchdog.sys crashes randomly, and i even have a new gcard now and still happens. ever since nvidia driver updated to 190.38 onwards (i believe anyways)


i have no issues at all with bluray progressive sources. only when needing deint/resizing do i get issues and i cant track it down... heres a readout of a crash when running a 1080i tv source through beta10 using this script:

LoadPlugin("D:\NV10\DGDecodeNV.dll")
DGSource("H:\1.dgm" ,deinterlace=2 ,resize_w=1024 ,resize_h=576)



On Tue 9/22/2009 3:43:22 AM your computer crashed
This was likely caused by the following module: nv4_disp.dll
Bugcheck code: 0x100000EA (0x88B42890, 0x88DE39F0, 0xF78C6CBC, 0x1)
Error: THREAD_STUCK_IN_DEVICE_DRIVER_M
file path: H:\WINDOWS\system32\nv4_disp.dll
product: NVIDIA Compatible Windows 2000 Display driver, Version 190.62
company: NVIDIA Corporation
description: NVIDIA Compatible Windows 2000 Display driver, Version 190.62

[neuron2]

I will notify Nvidia about this. I will also remove the timeout in the CUVIDserver until this is resolved.

[woah!]

maybe my other gcard is ok then heh... looks like sli 260gtx for gaming then ...

[neuron2]

I've re-upped NV beta 10 with the server timeout removed.

[nautilus7]

Once again, same thing happened with beta 10 (the initial version). It stopped at 99,8% (230 frames before end).

I am trying again with the updated beta 10 now.

[neuron2]

OK, I will await your results before buying the BD.

[menlvd]

latest beta of nvtools same things stall
[99.9%] 150630/150720 frames, 2.62 fps, 9128.19 kb/s, eta 0:00:34
sys wxp x64 nvdrivers 191.03

[nautilus7]

I am trying again with the updated beta 10 now.Success! Everything went fine with this version. :)

[neuron2]

Darn, I just bought the BD.

Still, I can use it to investigate what is happening with the server timeout, assuming I can duplicate the issue with the timeout enabled.

[nautilus7]

You said you will wait for my results... My cpu takes a while to encode the whole thing... Sorry.

[neuron2]

Don't worry, it was on sale and it's a cool flick anyway. Thank you for your test results.

[neuron2]

Actually, now I'm glad to have that BD, as it has exposed some bugs in random access. For example, if the sequence header carries quant matrices, they were ignored (code based on Nvidia sample code). Also, in some cases the CUVIDServer, when opening a new video, can return a frame from the previous loaded video.

I'll fix these.

[nautilus7]

Ok, that's nice to hear! I guess you 've also reproduced the bug i had in the first place, right?

(why this thread is still here and not in the proper category?)

[nautilus7]

Don't know if it's of much interest now, but my encode finished just fine. Thanks!