xiaowan3 / xy-vsfilter

Automatically exported from code.google.com/p/xy-vsfilter
0 stars 0 forks source link

Crash during border rendering when opening certain videos #181

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago
While I cannot personally reproduce this, we've had a report from user Kado 
@Doom9 about this crash, and tracked it down to being caused by our Insane 
Border Support commits.

The test video he provided can be found linked in his Doom9 post:
http://forum.doom9.org/showpost.php?p=1672228&postcount=339

Original issue reported on code.google.com by cyber.sp...@gmail.com on 8 Mar 2014 at 9:07

Attachments:

GoogleCodeExporter commented 9 years ago
Kado reported that the attached XySubFilter test build resolved his crash with 
this sample.

Compared to BETA2, it reverts the following:
https://github.com/Cyberbeing/xy-VSFilter/commit/fc3b925f403f1b8d854a6067d20e4c4
ed7c5034e
https://github.com/Cyberbeing/xy-VSFilter/commit/0d1b62494d61b53869fbc55aa5ee32b
eee38b60e

Original comment by cyber.sp...@gmail.com on 8 Mar 2014 at 9:11

Attachments:

GoogleCodeExporter commented 9 years ago
I finally figured out a way to reproduce this issue, by pure luck.

The attached Issue181.mkv will crash XySubFilter 3.1.0.682 Beta2 when the 
1280x720 video is resized to exactly 1600x900 output window size. Any other 
window size, and there is no crash. Still unsure what the exact trigger is, 
since if I made even a minor change to any of the remaining variables in this 
script (font choice, font size, fscx, fscy, frz, border size, alignment, etc) I 
would not experience a crash even at 1600x900.

The 'XySubFilter_Test_Build_(Revert_Insane_Border_Support' build I attached 
above resolved the crash, so at least this reconfirms those commits are somehow 
related.

Original comment by cyber.sp...@gmail.com on 14 Mar 2014 at 7:48

Attachments:

GoogleCodeExporter commented 9 years ago
Issue 183 has been merged into this issue.

Original comment by cyber.sp...@gmail.com on 30 Mar 2014 at 2:57

GoogleCodeExporter commented 9 years ago
Issue #183 provide another sample from [Underwater] Mikakunin 11, which 
triggers this crash when displayed at 2560x1440 only. This crash has the same 
cause as the other sample, and is also resolved by the 
'XySubFilter_Test_Build_(Revert_Insane_Border_Support' build I attached above.

Original comment by cyber.sp...@gmail.com on 30 Mar 2014 at 3:08

GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
If you don't mind, I'd like to add that it won't actually crash me at 1600x900 
like it will for you, so there's a difference.

Another thing I've found is that it doesn't seem to be have to be actually 
rendered on the screen to crash me. In the file I uploaded on Issue #183, the 
typesetting doesn't appear until frame 8, but I will crash immediately when I 
hit 2560x1440 (just slowly dragging the window larger), even if the video is 
still paused on frame 0. If I reduce the subtitle queue below 9, it won't crash 
until it's in queue, and still before it's actually rendered on screen.

Though if I had to guess the way you guys have your workflow between madvr and 
xysubfilter set up, I'd say this probably isn't new info, but I figured I'd say 
it anyway.

Original comment by lfp1...@gmail.com on 31 Mar 2014 at 7:46

GoogleCodeExporter commented 9 years ago
The sample in Issue #183 only crashes at 2560x1440.

The sample in Comment #2 only crashes at 1600x900.

The sample which Kado linked on Doom9 only crashes at 1319x742.

These crashes only occur when scaled to a single resolution for each sample 
which reproduces it.

Original comment by cyber.sp...@gmail.com on 1 Apr 2014 at 12:13

GoogleCodeExporter commented 9 years ago
I may have found some interesting things. For starters, here's a new sample 
which I'm going to assume for now makes xysubfilter crash for the same reason 
as this issue.

https://mega.co.nz/#!HcdEBSbY!ESryMyTOPWkX7ZPDBEDnUnajPlBQpsDsICGQTQzq3-Y

But while having another example is nice in itself, this sample is also unique 
in at least one way:

1) It crashes at both 2560x1440 AND 1586x892, making it the first sample (I 
believe) that will crash at two separate resolutions.

2) While playing with this file, I stumbled upon something else that may or may 
not have relevance. It was also only something that could be determined BECAUSE 
it can crash at two resolutions (at least, for me). The thing I found is that 
depending on the playback speed that MPC-HC is set at, it may not crash at a 
given resolution. I haven't tested this at one resolution with any of the other 
samples, but for this one, while it's at 2560x1440, if I set MPC-HC to x.25 
speed, it will not crash. x.50 and of course, x1.0, will crash. However, at 
1586x892, it can be set to x.50 or x.25 and it will not crash. x1.0 will crash, 
of course.

Original comment by bryce...@gmail.com on 2 Apr 2014 at 1:04

GoogleCodeExporter commented 9 years ago
Cyber can you compile XySubFilter_Test_Build_(Revert_Insane_Border_Support) x64 
version?

Original comment by Michal.A...@gmail.com on 23 Apr 2014 at 8:13

GoogleCodeExporter commented 9 years ago
Here you go. Compared to the previous attachment, this also contains the fix 
for Issue #174.

Original comment by cyber.sp...@gmail.com on 24 Apr 2014 at 8:02

Attachments: