Multichannel audio and multiple video file sync

So I have 2 x 1920x1080 files one of which contains 16 channels of audio. The client is unbelievably sync anal and so the sync between each file has to be exact always.

There is no processing going to be done to the video file in cat - I'm literally using it as playback sync device. Audio is going to come out of a motu.

Being a bit of a novice in cat land I wondered if the best approach is to have one QT with the 16 ch audio + a mute QT or to create a full raster output file and just pick out the bits I want to each screen output.

I think probably option 1 (the quickest for delivery) and set the sync id on layer 2 to track layer 1 is the best way to go sync id uses frame number to maintain sync right?

Does anyone have any thoughts of the best way of approaching this ?

Thanks

Jonathan

On the first layer with the track that has audio, you would set the play mode to Play forward audio (or similar) and set that to sync ID 1. On the second layer, you would set the play mode to sync ID 1. The clips will then play identically. Basically frame 1 of layer one will trigger frame 1 of layer 2. So make sure they are the same length. Each layer is still its own file and they can be positioned differently and routed to different mixes etc..

Ryan.

Thanks Ryan,

I have the clip with audio on layer 1, set to sync ID1. Layer 2 has the other clip which is set to play mode 80 Sync to layer 1. For testing purposes both clips have the same visual content which is fast moving with full screen white flash frames.

When I use one of the PL Audio modes, layer 2 does quite a bit of re-sync thought out the clip and it looks like there is about 2 frame difference between them - it is certainly lagging). There are no audible drops that I can hear.

When the play mode is set to a normal play mode then there is maybe one or two re-syncs in a ten minute clip and generally the sync looks much tighter.

In QT the clips are showing a 74mbit/s data rate - they are pro-res LT. I wondered if it was a data rate issue (even though they are coming off an SSD) so I fitted another SSD and have one clip on drive 1 and one on drive 2. This makes no different.

I also made a clip that was 3840 x 1080 to see if that was better - setting layer 2 to user layer 1. Again the audio was fine but the content looked like it was running at 20fps - just a bit sluggish.

Any suggestions?

I'm running on a 12 Core Mac Pro / 16Gb / Radeon 5870 /10.6.8 / Crucial M4 512GB SSD / Cat 4.4_m364

Thanks

Jonathan

Try this.
Render your audio track with small black video (320x240 for example).
Play this file in layer 3 and layer 3 set as sync id 3. Then layer 1 and 2 with video content play in mode sync id 3.
Tom

I would also recommend using AIC instead of pro-res. That may run better.

Ryan

These days ProRes runs very nicely on Macs - I have been using it for years on all my shows. 4:2:2 versus 4:2:0 - no contest for me.

ProRes uses hyper threading which means it performs very nicely on modern Mac Pros. Don't bother with HQ. Nice for Post Production. Not much use for Catalyst as none of our displays/ projectors/ LED walls can render 10 bit.

With respect to Audio synchronisation - it is worth bearing in mind that Catalyst prioritises audio playback over video, and this can create problems when synchronising across multiple machines/ layers. Try and sync your video layers to the audio/ video layer, as others have suggested. Also bear in mind that the performance takes a bit of a nose dive when you use audio playback through Catalyst.

Hugh

I tried running AIC but this really made things splutter - max of 10 fps according to the cat display.

I also tried the small video frame with the audio and then syncing the two full frame video clips to that - didn't work but did seem slightly better.

Perhaps catalyst just cant do it and keep it tight because of the audio priority.

As Hugh already said: ProRes is no problem today - but AIC is no longer satisfying on the quality side.

With multichannel audio I like to have a separate machine just for audio playback (or even a hard disk audio player - triggered by MMC) and add a channel for SMPTE timecode.

This timecode then acts as the sync source for all Catalyst that may need it. If there happens to be a global delay in timing you can easily adjust with the TC sync timing playback controls (found on layer in- and outpoint controls).

Because of individually running gfx cards´ clocks you might see a few retiming frames during playback - I would say like one in three minutes - but that is usually smoother then ID-syncing or free run...

Hugh, is there a big difference in hyper threading between 2009 - 2011 Macpro's and 6, 8 or 12 cores, when 6 cores have a faster processor?

Peter

Quote:

---

Originally Posted by visionmill

When I use one of the PL Audio modes, layer 2 does quite a bit of re-sync thought out the clip and it looks like there is about 2 frame difference between them - it is certainly lagging)

---

Something that was not mentioned is the difference between 29.97 and 30 frames per second.

It is very likely your Video file with Audio is rendered at 29.97fps. Since the audio play mode becomes dominant, it will force 29.97fps. However, the other files (on a cat at 60HZ) will be running at 30fps until a sync is lost and then jump to the correct frame. This will look like stuttering in your playback.

In this situation, re render the file as mention at a very low rez black and have this as a blind layer playing audio (yet with the video set at 30fps). Then your slave layers (same or different machine) can sync to this and as slaves, should sync appropriately.

An interesting point about 29.97 vs 30. Sadly it doesn't apply here as our clips are 25fps running on 50Hz monitors

Quote:

---

Originally Posted by visionmill

An interesting point about 29.97 vs 30. Sadly it doesn't apply here as our clips are 25fps running on 50Hz monitors

---

I didn't necessarily meant that this applied to you. It is something I wanted to mention for the sheer sake that I get this question and have to explain this to people often.

In all cases, the solution of running a separate audio source video file rendered as low rez as possible and syncing all layers to that is the best overall solution.

Peter,

I only use dual quad cores, and have done for many years, so I can't really comment on other machines.

Hugh