HD plugin - audio dropouts at transitions

Post Reply
Joe Linn

HD plugin - audio dropouts at transitions

Post by Joe Linn »

I'm using the JVC GR-HD10 high definition camera.

I render HD videos in MSP. When I view them in the preview window, everything is fine. When I record them back to the camera, there is approximately a 1/2 second audio dropout at some of the transitions. Other transitions are perfectly fine with no dropouts. I've tried to determine what is different between the transitions without audio dropouts and those with, but without success. I thought it was perhaps certain transitions or perhaps the lenth of the transition. Modifying those factors to match the transitions without dropouts doesn't get rid of the dropout.

I'm wondering if it might be some sort of sync issue, since the same sort of dropout occurs between shots when viewing raw footage on the camera.

Has anyone experienced similar problems? Does anyone have a solution?

Thanks!

Joe
Joe Linn

Post by Joe Linn »

I think I might be on to something. I tried rendering the video with the "i frames only" option selected. There were hardly any audio dropouts when I output the file back to the camera.

I wonder if something odd happens with the order of i, b, and p frames at the transitions and it upsets the camera.

Any ideas?

Thanks

Joe
Joe Linn

Post by Joe Linn »

I'm still trying to find a work-around for this problem.

I tried importing the rendered file from MSP into MPEG Edit Studio Pro LE, the software that came with the camera. I thought perhaps if I re-rendered it with that software, it would fix whatever problems existed. I started the render and when it got to the area where the first audio dropout occurs, I got the error message "No sync word (0xfff) found in the MPEG audio stream."

Does anyone have ideas for how I can fix this problem, or at least a work-around that would allow me to produce files without the audio gaps? I am at the 7.2 level of MSP.

Thanks for any help you can provide.

Joe
Brian Conrad
Posts: 4
Joined: Fri Feb 25, 2005 9:59 pm

Post by Brian Conrad »

I am assuming from your message that you are rendering the transport stream (M2T) file with Smart Render off. The camera usually does not like remuxed transport streams if new material has been added so they usually have to be completely rerendered. I believe the latest version of the plug-in turns off SmartRender by default when creating transport streams.

Otherwise have you tried exporting the audio and checking with the Audio Editor for the dropout and if okay save as a WAV file?. Then import only the video from the file and then the WAV file and rerender. That workaround might help.

Packet syncs sometimes occur when the file gets written wrong and examination with a hex editor will often show that the sync word is located later in the packet. A robust engine would note that and check for the sync word elsewhere in the packet and try to recover the data during the render.

Brian
Joe Linn

Post by Joe Linn »

Thanks Brian!

Turning off smart render made a big difference. With it turned off, the audio gaps are reduced to about 1/4 second. It also produced a file that I was able to import into the editor that came with the camera (it failed on the files I had been producing). When I re-rendered with MPEG Edit Studio, the audio gaps went away.

That seems to indicate that with smart render turned on, MSP produces transport streams with errors at some transitions. With smart render turned off, the errors are much less severe and MPEG Edit Studio is able to remove those errors.

Now I'm hoping to find a way to eliminate the need for the second step of running it through MPEG Edit Studio. Perhaps a maintenance release of MSP will address that.

Thanks again! I can now work around my problem.

Joe
Post Reply