QLab 2 video playback issue - mp4 glitch?

621 views
Skip to first unread message

Brendan Patrick Hogan

unread,
Jan 24, 2014, 11:41:25 AM1/24/14
to ql...@googlegroups.com
Hi all,

I have taken over for the video programmer on a show running a large number of video cues in mp4 format (H.264 codec). I am not a video designer, so I'm reaching out to the group with this problem: 

I have noticed an issue in which longer video files seem to start playback at the last ½ second of the file, then proceed to the start of the file, and from that point, playback behavior is normal. I have checked the files in Quicktime, VLC player and other third party players to verify that frame 1 of the video is in fact correct and the odd behavior - *starting at the end* - only occurs when the media is being pushed by QLab. The issue may be caused by forcing these older machines (see notes below) to decompress large video files with limited available resources. I have read that mp4 is a supported format, but that the decompression can run riot on the GPUs of older machines. Here is our setup:

Primary machine: iMac 24" 6,1 / 2.33 GHz Intel Core 2 Duo / 3GB 667 MHz DDR2 RAM / 250GB internal HDD / 250GB external media HDD (USB 3.0 enclosure --> USB 2.0 port) / Mini DVI to VGA out to the projector

Secondary machine: Mac Mini 2,1 / 2.0 GHz Intel Core 2 Duo / 2.5GB 667 MHz DDR RAM / 160GB internal HDD / 250GB external media HDD (USB 3.0 enclosure --> USB 2.0 port) DVI to VGA out to projector / receives MSC commands from primary over Ethernet. 

I don't recollect the internal HDD drive speeds off the top of my head, but I would guess they are 5400 rpm. I do not believe the theater has modified them in any way other than the additional 1GB of RAM I installed in the iMac.
 
The external drives hold all of the cues for both machines. Initially, the internal HDDs contained the cues, but were quickly getting bogged down when running video cues in addition to sound cues. Moving the cues to an external source helped, and I have installed KeepDriveSpinning to make sure the external HDDs are always up and running as they tend to ignore the OS X power management settings to keep drives awake.

Today we are experimenting with alternative video file formats to see if that makes a difference. In the mean time, I would appreciate any insight you can lend.

thanks much,

brendan patrick hogan

Michael Long

unread,
Jan 25, 2014, 12:38:46 AM1/25/14
to ql...@googlegroups.com
On Friday, January 24, 2014 11:41:25 AM UTC-5, Brendan Patrick Hogan wrote:
I have noticed an issue in which longer video files seem to start playback at the last ½ second of the file, then proceed to the start of the file, and from that point, playback behavior is normal.

Very interesting. I would love to hear what you discover when trying different codecs or formats. Please let us know!

--
Michael Long
Projection Designer - Philadelphia

Brendan Patrick Hogan

unread,
Feb 2, 2014, 11:47:28 PM2/2/14
to ql...@googlegroups.com
After exhaustive testing on other systems (2011 Macbook Pro via QLab v2 and v3, 2009 Mac Pro running v2, etc), I was not able to reproduce the behavior described in the original post. This issue appears to be unique to the machines running the show and I suspect it is a combination of the relatively weak processing power and the compressed nature of the large mp4 files. Rendering the videos in other, larger, formats/codecs did not produce the odd 'start at the end' behavior, but the associated increase in file size produced a marked decrease in overall performance and some occasional crashes. Given that the videos are intentionally choppy and glitchy as a design choice, I was able to disguise the 'start at end' problem by starting the videos at 0.025 seconds into the file, with an instant fade to 100% opacity. This produced a kind of 'snap in' effect that actually worked quite well with the design esthetic of the show and the audience was none the wiser, though it was a pretty frustrating 'good enough' fix to settle on.

I wish I had some exciting new information to report, but it does look like this is all just a condition of large, compressed media being pushed through old, tired machines. If they can make it through a 30 day run, that's good enough for me.
-bph
______________________
Brendan Patrick Hogan
--
--
Change your preferences or unsubscribe here:
http://groups.google.com/group/qlab
 
Follow Figure 53 on Twitter: http://twitter.com/Figure53
 
---
You received this message because you are subscribed to a topic in the Google Groups "QLab" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/qlab/HvIUDq0AUA8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to qlab+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Reply all
Reply to author
Forward
0 new messages