Video playback with Audio. Sync error?

[SourceChild]

I was advised to play the audio back on a new layer and ether use sync id or trigger both the layers at the same time.

I'm not sure who advised you of this but there is only one advantage to placing audio on a separate file. That reason is to be able to adjust the inframe or outframe amounts of the video file to adjust the offset of frames between layers.

This is an advanced technique that I have suggested to people in the past as a workaround but this is not a solution that someone should use when creating new shows.

If you must play audio with video, follow Simon's statement and render them together as one file, verifying they are synced in post before loading them to the Catalyst.

[pip]

Just found that running the audio from the same layer causes the playback to jitter and a 'spike' in the mov 1 play reading in the stats hud...

Attached video: http://www.youtube.com/watch?v=wkmajfu0wdA

This does not happen when using an non-audio play mode. I take it this could be a graphics card problem? Will look at replacing the cards tomorrow.

[samsc]

Just found that running the audio from the same layer causes the playback to jitter and a 'spike' in the mov 1 play reading in the stats hud...

Attached video: http://www.youtube.com/watch?v=wkmajfu0wdA

This does not happen when using an non-audio play mode. I take it this could be a graphics card problem? Will look at replacing the cards tomorrow.

'spiking' is normal - this is audio being buffered

[SourceChild]

Just a note about the FPS counter and audio playback....

I am currently doing a show through South America and we are running content that is 29.97.

Yes, I am running content on Catalyst and no, I did not want 29.97 but we are also using the content on playback decks as well so 30fps was not an option.

The content has to run frame accurate with the timecode of the show because the content was created exactly to the music of the set. In fact, we placed the music on the content as reference.

Since the show timecode exceeds the total allowable length of time that a Catalyst frame counter can recognize, I have had to program the presets to be triggered from the SAM cuelist. Basically meaning the cuelist starts the file and then it free runs until the next SAM cue fades it out.

I know this is a bit of a tangent but my point is simple.

In order to make the files play frame accurate (roughly) to the timecode, I needed to use the audio playback mode because it will use the audio as a time base reference for playing the files.

Yes, I drop frames this way and my whole team is aware of it. However, when the vocalist is shown in the content, lip syncing the same lyrics she's actually singing then the audience is more likely to notice voice slip than a few missed frames.

Not the correct way to do things, but since day to day we might have 50hz or 60hz, power and PAL or NTSC, it's the only way to cover all bases.

Take these thoughts as insight but try not to use this method if a better alternative is available.

[pip]

Thank you both for your comments. In reply to Simon the mov specs are as attached...

[pip]

Ok, quick update. I have now installed Catalyst 4.40 onto two of my MacPro and my office i7 MacBook Pro and tried each with the file i am experiencing problems with.

When playing the file in mode 70, all the macs play with a stutter so this narrows it down i guess to the way the video has been rendered in post..?

Or i have three bad macs!

[SvenCRen]

Okay, I have a few thoughts on this subject. I have had these problems on many occasions and have yet to find a solid solution for it.

Here is what I understand:

1. Running the video with an audio play mode when they are merged together in the same file can cause playback to be less than smooth. This is because as Todd mentions, the playback rate is driven by the audio portion of the file for sync, not the video. The Mac OS will always treat the audio as a priority.

2. Running the audio and video on separate files on different layers will potentially yield the best result for smooth video playback but then this is where we run into the "lip sync" problem. I typically run the audio from a different server dedicated just for that purpose.

Here is what I have discovered since:

1. Running the audio and video independent of each other works wonderfully if the media was created at a specific frame rate and your display is running perfectly at that rate. ie: When I have the media at 30fps and connected to monitors that are at a perfect 60Hz refresh rate, I can achieve perfect sync.

2. 60Hz is not always that. On most systems (Spider, Encore, Etc.) or even large displays and projectors, the true refresh rate will be 59.94. Even when your display settings are set to 60Hz in both the system settings as well as Catalyst, your output frequency will be adjusted to that of the display your connected to thanks to EDID management through DVI. This is easily verifiable when you look in the Stats window in the HUD.

Here is my plan to fix the problem next time around:

1. Use a DVI detective to capture the EDID from the monitors that work and force the video card to stay at those settings when connected to other systems.

2. Continue to run audio and video independently so that the video playback will be as smooth as possible.

3. Be more cognizant of the refresh rates that the Catalyst is capable of.

The playback pro doesn't seem to have this problem. I am really curious to know what they do to correct this and provide solid, clean playback. That is a Mac based system too for anyone that doesn't already know and is responsible for taking over when Catalyst has failed me in the past with this.

There are a couple of products on the market that are actually designed to deal specifically with this problem. The DVI parrot is one that comes to mind that is designed to "prevent display misunderstandings" as their literature says.

[pip]

Was up last night testing different mov files (photo jpeg, AIC... 25/30fps) on different versions of catalyst (4.10 - 4.40) running 10.5.8 + 10.6.7! Then got some sleep.

Im still finding dropped frames when using the audio play mode as mentioned in this thread. Looking at the stats its the middle 'pink' bar on the layers play graph that spikes (as mention before) causes the dropped frames and stutter.. be good to learn what this represents?

I also split video/audio on another layer and

• a) triggered both play setting at the same time... video was one second behind by the end of the clip
• b) synced them by local ID
• c) ran Catalyst on another MacPro (no dongle) running just the audio and remote synced the two together. MacPro with the DV dongle was the master.

I did notice the screens i have been using are fixed to 60Hz and the cinema display 59.XX.

Will be using a TH2G with LX41 and Lighthouse R7 LED come the show in a weeks time!

Not looked into Playback Pro yet, will do, but did run the 25/30FPS versions on Qlab and that showed no problem. Least i have a backup, but need the functions of Catalyst.

[pip]

If anyone wants to see if its the content i personally am using... i have cropped a section if it would help??

http://bradeng.com/tmp/001%20Test%20Cropped.zip

Screen recording... Note at 15FPS!

http://bradeng.com/tmp/scr_rec.mov

[samsc]

what screens and what screen sizes?
what catalyst playback frame rate?
screen 1 div 2? screen 2 div 2

-

'spiking' is normal

the usual problem is not drop frames at all - but screen refresh rates that do not match the audio/video frame rates
the 'visual' stutter isnt what people think - there isnt any stuttering - the screen just isnt displaying the image because the movie frames display times dont match screen refresh times

when doing audio - you have to work differently -

the frame rate of movie needs to match frame rate of screens -

triple heads are particularily obnoxious for this

-

screens dont run at 60Hz or exactly 75Hz -

Was up last night testing different mov files (photo jpeg, AIC... 25/30fps) on different versions of catalyst (4.10 - 4.40) running 10.5.8 + 10.6.7! Then got some sleep.

Im still finding dropped frames when using the audio play mode as mentioned in this thread. Looking at the stats its the middle 'pink' bar on the layers play graph that spikes (as mention before) causes the dropped frames and stutter.. be good to learn what this represents?

I also split video/audio on another layer and

• a) triggered both play setting at the same time... video was one second behind by the end of the clip
• b) synced them by local ID
• c) ran Catalyst on another MacPro (no dongle) running just the audio and remote synced the two together. MacPro with the DV dongle was the master.

I did notice the screens i have been using are fixed to 60Hz and the cinema display 59.XX.

Will be using a TH2G with LX41 and Lighthouse R7 LED come the show in a weeks time!

Not looked into Playback Pro yet, will do, but did run the 25/30FPS versions on Qlab and that showed no problem. Least i have a backup, but need the functions of Catalyst.