Just guessing here, but in the past I have noticed that audio formats that have variable bitrates seem to have more tendency to drift sync than files like wavs with fixed bitrates. Or if the sample rate and bitrate do not match your project settings.

You could try using an audio program to convert the file to a wav first, but I don't know if that would make a difference.

Dave T2

If it's noticeable after 10 seconds, it's probably due to sync drift from the format as DaveT2 says. 29.97/30 or 23.976/24 pulldown error would only put you one frame off around the 30 second mark.
If you use the Elastique method time-stretch on the audio track to make it the same length as the video track does it then maintain sync? If so, I would then render a wav file so you have a stable sample rate. It may still drift in and out. Who knows what encoding was done to the audio before it was put in the MP4 container. Any digital audio that isn't encoded at a constant bit rate can't be decoded with the samples in the correct spot.

Edit: Just as a test, in Vegas 11 I rendered a 30 sec audio segment in .wma (which I NEVER use) once with VBR10 and once with CBR, both set for the lowest quality 44.1 file. Put them in sync on a timeline and inverted the phase of one - lots of audio remains whooshing in and out. And the VBR file was 4 frames shorter. ??

"This make sense to anyone?"

Sure does. The player's internal clock could be off.
Found this with CD players as well. Then there's the ones DJs use...:(

Bob.

The fact that she was lip-syncing to her own recording, and that was the recording you were given, is puzzling. Tends to negate some of the profound (?) advice I can give - nevertheless for what it's worth..

I recently played with a choir recording for someone; turned out that the video was from the early church service and the audio was from the later service. The lesson I learned from that was that no matter how good an individual or group may be, there will be minor variations between performances. The main difference I noticed was the length of pauses between phrases. Only a second maybe, but enough.

As I say, this may not apply to you, unless her lip-syncing was bad. The remedy may help though. I managed to find the quiet spots - pauses between phrases - cut the audio track and slid it in slightly; also used elastique to stretch another segment. By cutting the audio track in a number of places and lining it up with each phrase I came up with a working result.

Good luck!

Thanks for the replies. The lipsync should not be the issue. One, she was very good. Second, I know this because I can watch raw video of her singing to pre recorded music and she's spot on. All the way through she's on. If I look at the audio wave files in Vegas pro, the audio I received in PMEG-4 format is longer than the camcorder audio. I guess I can try to feed the PMEG-4 audio into cd recorder and burn it and a wave and try that but I think I need to have mastered audio burned and saved as a wave file.
I guess the real question is: Is an MPEG-4 file, even though it is compressed going to be time wise longer than a wave file?

Thanks all,

Sync the audio by eye at the beginning -- how far off it is in the end? Two markers will tell you. How long is the program? You haven't posted the video and separate audio properties using MediaInfo, so anything you get here is just guesswork.

Yes, I do snyc by eye/ear. I tend to match up wave files by eye then tweak it to perfect asI can by ear.I sync camcorder audio with mastered audio. I have never seen this issue before. If it is sync'd at begining there is no reason it should be off at end let alone after 5 seconds. It is off about 5 seconds by end of sone. Song is about 4 minutes.
I'm usualy pretty good at syncing I believe. See present work
www.mancaverecordings.com

"You haven't posted the video and separate audio properties using MediaInfo, so anything you get here is just guesswork."

audio is as I said camcorder audio. It is film of a recording studio performance. Girl lipsync'd to her recording. I was of course there and her lipsync was very good, matches up with camcorder audio great. I was given mastered audio by studio on cd. It is 44.1,16 bit of course for cd but it is an MPEG-4. I have only used wave files before. I think as the MPEG-4 is compressed it may have changed the overall lenght?
Video......uh I'll have to check

Thanks

The studio gave you an .M4A file on a data CD? That seems a bit odd.

Regardless, it is virtually impossible that encoding it to an M4A (or MP3 or WMV or AAC) would have necessarily changed the audio sample rate. Compressed does not mean compressing the length.

MediaInfo is a free download from SourceForge.

There's a good chance there is a sample rate mismatch in the source, or you have another audio device open that is hijacking your sound card. Some detailed information is necessary in order to help you track this down. I hope you are not putting the audio from the CD directly on your timeline, but are copying the file(s) to your desktop first.

All speculation about compression and VBR audio aside, there should be very little difference unless one of the conditions mentioned above exists.

FYI, audio streams for video are 48KHz, but Vegas timeline should handle both "if" your audio devices are at the correct sample rate, and not being held hostage.

I would be curious what audio studio would deliver an MPEG-4 audio track. Is there a reason to do this? I would think MP3 at the minimum or an actual audio CD.

Dave T2

Musicvid and davetv,

Thanks for help. I typically use 48k 24 bit for video yes. I also don't know why I got audio in MPEG-4? If it were a sample rate issue I would think I would hear a speed issue. Just like if I play back a cd when my apogee is on 48k but its a cd at 44.1. Should not be conflict as I've used same setup,sound card,apogee as with all past projects.

yes audio copied to computer first.I tried it both way for fun, no difference.

"MPEG-4" audio is a generic name; usually means AAC.
AAC is a newer standard than MP3, and is supposedly better in the high freqs.
However, theory and practice are two different things. The AAC encoders out there are generally crap, and are actually worse than LAME, for instance. One exception is the free Nero AAC, which I've started to play with, but finding muxing options for existing video has proved to be a little convoluted.

Sync it as best you can and when you notice it drift out of sync, cut to another angle or b-roll or something. Resync from that point and do the same when/where necessary. Slow-Mo and FX can also disguise it a bit. As will the whole 'filmic' 24p look/feel.

Before you spend much more time trying to sync it, I would contact whoever gave you the music and see if you can get a WAV or CD format of it. It would be higher quality anyway and you would be syncing to a known solid source. I read what Musicvid said about MPEG 4 audio, which would indicate to me this is a compressed format, which for a final music video, I would not want to use an MP3 or other compressed audio format.

Dave T2

DaveT2
Yes that is what I have done, request songs i wave format, 48k 24 bit. That is format I always use. I have no intent in trying to sync. It should take no more than 1 minute to have it perfect. The MPEG-4 is just incorrect..too long.

Jim

Mancaverecordings said:
"If I look at the audio wave files in Vegas pro, the audio I received in PMEG-4 format is longer than the camcorder audio."

You're still not getting it.
I recorded most of the music for a stage show in Vegas. Made it into a CD and handed that to the venue's sound guy. Sounded great btw.

I recorded a stereo feed from the desk into my own recorder and of course the cameras were recording the audio in the venue for sync.

Compared to the original masters made at 48KHz and the CD I'd handed the sound guy at 44.1Khz we still had sync issues. The issue is that many audio players be they CD players or mp3 walkman like devices do not have accurate clocks i.e. 44.1KHz is not exactly 44.1000000 KHz. I've had errors of seconds between the actual CD and how long it was played back and then recorded in a venue.

After much research I found that for this very reason one can buy expensive rack mounted CD players that have a Word Clock input. With those in a venue everything can be locked down, cameras, digital desks, cameras, everything runs from signals run off the one master clock.

One tip, if you're going to get talent to playback and lip sync for a video recording make a version of the track at 48KHz and give them a decent player. For some reason 44.1KHz is considered a bit dodgy compared to 48KHz. 48KHz was chosen as the audio sample rate for video as it's an integer multiple of the base video master clock frequency.

Bob.

Bob,

Thanks for the reply. I do get it however. Like I said, I did not request the audio in MPEG-4, 44.1/16 bit, it's just what I was given. I use 48k,24 bit for projects I do that where I do both audio and video. The question is "will" that MPEG-4 audio @44.1/16 bit cause a snyc issue? I believe the answer is yes. We on the same page? I clock in my studio with good clock. Apogee AD-16X with the Big Ben clock. As I said have never run into this before but then again I have never worked on a project where I did not engineer the music.
Jim

If I had this problem, the things I would do are:

1) start a new project in Vegas or Sound Forge

2) put the foreign file (m4a or whatever) on the timeline.

3) set the project properties to match the foreign file (44.1/16)

4) render to 48kHz/16bits WAV

5) replace the foreign file on the sync project timeline with the newly rendered WAV file.

Does the 48kHz/16 bit WAV file work any better?

"We on the same page?"

Just to be sure, to be sure, did you playout the MPEG-4 recording in your studio while the video was shot in your studio?
What from?

Bob.

Bob,
I went to recording studio where music was recorded. They, the studio engineer, did a playback in room speakers of recording,un mastered yet but mixed of the tune. Singer sang along with tune into mic just for show. I filmed it with 3 hd camcorders. I received cd with audio, same audio she sang to in filming only mastered. The audio was MPEG-4 format. I imported audio into my computer then into Vegas pro 9. In Vegas pro 9 I had one camera clip already there. As the audio was 44.1/16 bit, that is how my Vegas pro 9 session was setup for.
I do know that audio playback was straight from protools into room speakers. I never asked if he was recording in 48k or 44.1 as it did not mater to me at the time. When I do remote gigs I always record 48k for all m multitrack captures. So no, the playback in the studio was not MPEG-4 format. They would have been just wave files. So if I get waved files of songs it should all match up as far as I can see.