I have a Zoom H1 and an H2. As many people know, both of them lose sync over time. So what other recorders in that price range and quality are available that maintain perfect sync?

The only way they can maintain "perfect" sync is to have a sync or timecode input. That isn't going to happen in the Zoom H2 price range.

I have a Zoom H1 and an H2. As many people know, both of them lose sync over time. So what other recorders in that price range and quality are available that maintain perfect sync?

There aren't any devices that meet your specifications.

Drift is a fact of life. Perfect doesn't exist. And even if you could find a "perfect" audio recorder, you'd pair it with an imperfect video capture. This situation is exactly why time code (http://en.wikipedia.org/wiki/Timecode) was invented. With time code you have a single master to which all other devices must sync. Works like a charm, but isn't cheap.

Sound Devices 700 series with timecode - but not in your price range.

Cheaper - Fostex FR-2 with timecode board (NB: not the FR-2LE).

Cheaper still - Tascam HD-P2, with a video sync input that can take composite and lock clock to it. But that is, I think, the least expensive device with sync.

With time code you have a single master to which all other devices must sync. Works like a charm, but isn't cheap.

So I guess that single master is a device that generates a timecode and the other devices, both camera and audio recorder, have to be capable of synchronizing with that master device, right? Can that be done wirelessly, or is it always with cables?

So I guess that single master is a device that generates a timecode and the other devices, both camera and audio recorder, have to be capable of synchronizing with that master device, right? Can that be done wirelessly, or is it always with cables? That is done with cables. Timecode in video and file-based audio does not provide a sync reference for speed so while it will provides a single reference point (just like a slate) so you can aligm the files it does nothing to prevent drift over the duration of longer shots To prevent drift you need to slave both the audio and video sample cocks to a common timebase. Your camera must have a sync output or a genlock input. You audio record must have a wordclock intput. The master clock provides the clocking signal to both. It provides a common time signal in two flavours - blackburst or tri-level sync to slave the video camera sample clock to and wordclock to slave the audio device sample clock. You can also get an audio recorder that accepts video from the camera and slaves its sample clock to it, but none of those are even close to the budget limit you expressed. Good examples are the Sound Devices 788t at about $6500 and I believe the NagraVI at about $9 grand. The Tascam HD=P2 is a lot less, $1000, and it does have video sync BUT I've heard that it only works if the recorder is also receving timecode - I haven't been able to verify that but until you check it out thoroughly take the claim of sync to video blackburst with a grain of salt.

You can also get an audio recorder that accepts video from the camera and slaves its sample clock to it, but none of those are even close to the budget limit you expressed. Good examples are the Sound Devices 788t at about $6500 and I believe the NagraVI at about $9 grand.

Ouch. So I guess I'll keep on relying on my good human ear for the time being. I don't know if anyone knows this, but which NLE does a better job at shrinking or stretching audio between Vegas and Edius? I use Edius for everything, but I also have Vegas, so if necessary I can process the audio from the recorder there and then importing it into Edius.

Hi Sebastian,

I've got a Sony PCM-D50 that I use for long recordings of stage shows. I edit in Vegas Pro. The D50 is pretty reliable at drifting 3 frames per hour slower than my Sony EX3. In vegas it is simple to do a very minor stretch to make them sync back up.

-Garrett

Thanks, Garrett. I think it's also a matter of how the different NLEs read the wav files. I imported the files into Vegas 10 to do the stretching to match perfectly, but it doesn't sync in Edius 6. If I sync them at the beginning, after 20 minutes or so they are out of sync by a few frames. But in Vegas they stay sync'ed from start to finish.

That is done with cables. Timecode in video and file-based audio does not provide a sync reference for speed so while it will provides a single reference point (just like a slate) so you can aligm the files it does nothing to prevent drift over the duration of longer shots To prevent drift you need to slave both the audio and video sample cocks to a common timebase. Your camera must have a sync output or a genlock input. You audio record must have a wordclock intput. The master clock provides the clocking signal to both. It provides a common time signal in two flavours - blackburst or tri-level sync to slave the video camera sample clock to and wordclock to slave the audio device sample clock. You can also get an audio recorder that accepts video from the camera and slaves its sample clock to it, but none of those are even close to the budget limit you expressed. Good examples are the Sound Devices 788t at about $6500 and I believe the NagraVI at about $9 grand. The Tascam HD=P2 is a lot less, $1000, and it does have video sync BUT I've heard that it only works if the recorder is also receving timecode - I haven't been able to verify that but until you check it out thoroughly take the claim of sync to video blackburst with a grain of salt.

When I said Sound Devices 700 series I was thinking more of the 702T which is about £1,599 +VAT in the UK (Canford Audio).

(Though I use the Nagra VI myself).

When I said Sound Devices 700 series I was thinking more of the 702T which is about £1,599 +VAT in the UK (Canford Audio).

(Though I use the Nagra VI myself).

Was looking for those recorders that have a provision to slave the audio sample clock to incoming video, blackburst or tri-level signals. The 702t/744t series don't have a video sync input, while the 788t does and can accept video blackburst or genlock as an external sync source. The 702/744 do have external timecode input, of course, but according to an email I received from SD a few years ago, they don't slave their audio sample clock rate to the incoming timecode's embedded clock as it's not sufficiently accurate.

Is my memory correct in that the Nagra does sync up to composite video?

IME, "adjusting" (shrinking or stretching) the audio track(s) should be avoided. It is MUCH easier to "pull up" the video to match the audio. Anywhere you make a video transition is a completely seamless opportunity to do a slight correction at that point.

I lay in the audio track (after mixing/editing if that is your work-flow) and use that as the MASTER reference in the video editor timeline. Then I drop in the video clips and align them with the reference audio track. There are commercial products like "PluralEyes" which do this alignment automatically for you. Singular Software - PluralEyes (http://www.singularsoftware.com/pluraleyes.html)

IME, "adjusting" (shrinking or stretching) the audio track(s) should be avoided. It is MUCH easier to "pull up" the video to match the audio. Anywhere you make a video transition is a completely seamless opportunity to do a slight correction at that point.

I lay in the audio track (after mixing/editing if that is your work-flow) and use that as the MASTER reference in the video editor timeline. Then I drop in the video clips and align them with the reference audio track. There are commercial products like "PluralEyes" which do this alignment automatically for you. Singular Software - PluralEyes (http://www.singularsoftware.com/pluraleyes.html)

Richard,

Not sure what you mean by "pull up" the video but I'm curious as to why you would rather play with your video playback rate rather than the audio. Whenever you adjust the frame rate of the video you can experience many more problems with artifacts. You also are forcing the video to rerender every frame. So if you do happen to shoot and capture in a codec that is for final delivery you are needlessly adding to the processing time.

I would agree that if you do have a cut in your video you could cut the audio and make a slight adjustment to the audio track placement. In fact that's one of the reasons why I set up my dance show videos to cut out the dead time between numbers. But for things like ballet performances and plays where you can't make those cuts you have to stretch or shrink either video or audio. In my experience it's better to adjust the audio as it isn't noticeable whereas I have seen some really strange artifacts pop up in the video when trying to adjust by only a few frames over an hour.

-Garrett

One other thing I forgot to mention. Video should only be adjusted by whole frames. In other words, don't adjust a 24 second video shot in 24p by 1.5 seconds. You have to adjust by 1 or 2 seconds or you will see some really unwanted results. You can however adjust the audio by any length without the same nasty effects.

-Garrett

-- Not sure what you mean by "pull up" the video

I mean slide the video clip in the timeline so that it syncs with the audio.

-- I'm curious as to why you would rather play with your video playback rate rather than the audio.

Because most tools I have used leave artifacts in the audio when you diddle with the the timing. Anomalies ranging from mild to horrid. This is not only my experience but that of others who have discussed this before in this and other forums. OTOH, moving the video to match the audio is undetectable by most viewers.

-- Whenever you adjust the frame rate of the video you can experience many more problems with artifacts.

Agreed. Which is why I don't adjust the frame rate. And you make my point about artifacts from trying to adjust timing/pitch in general. Simply sliding the video into sync does not involve changing its length or frame rate.

-- You also are forcing the video to rerender every frame. So if you do happen to shoot and capture in a codec that is for final delivery you are needlessly adding to the processing time.

Which is another reason I don't do that. (And again doing it to the audio track forces a similar re-rendering.)

-- I have seen some really strange artifacts pop up in the video when trying to adjust by only a few frames over an hour.

Again, which is why I don't recommend diddling either the video or the audio. You will hear strange artifacts in the audio throughout if you try to stretch or shrink the track.

-- Video should only be adjusted by whole frames

IME, neither video nor audio should ever be "adjusted". The process does damage to both audio or video, but it is much more noticeable in audio than in video.

Any video with a single shot long enough to go out of sync will be deadly dull. Any point where an audio or video edit is made is a perfect opportunity to "pull-up" (or "re-sync") audio and video. This is actually much easier than it sounds. It is also inherently immune from shrink/stretch artifacts in either audio or video.

IME, also audiences are much more forgiving of adding or deleting a single frame of video to pull up the sync than having to put up with audio artifacts throughout the production.

-- You can however adjust the audio by any length without the same nasty effects.

That is most definitely NOT my experience. I would avoid doing that at any cost. Fortunately there are very easy ways of avoiding that.

Adjusting the speed of audio can be done without causing gross artifacts, though people with golden ears will certainly disagree. Stretching audio will, however, cause gross artifacts when you try to change the time and preserve the pitch or when you try to change the pitch and preserve the time.

I would approach it differently, depending on the project.
1) A narrative piece with lots of cuts: Just sync the start of each segment. No processing needed for short clips.
2) A long, continuous speech: I'd stretch the audio as needed. Or, if you have time, cut it up during pauses every few minutes, sync each segment, and blend the background sounds to eliminate gaps.
3) Music video: Don't touch the audio. If you have cuts, it's no problem. If it's a continuous live shot, you can probably add/drop frames when there is little motion without problems. If that's not working (say for a moving steadicam shot), stretch the audio.

The best approach really depends on the situation and the amount of time you have available.

Richard,

In general I would have to disagree with your assessments for the specific types of productions I'm thinking of, single camera continuous running shows that have both dialog and music. I would agree that the best way would be to cut either the audio track or video track and drop the necessary length. I would have to say that I have not experienced the audio anomalies that you've experienced. As Jon noted, if you try to maintain pitch while applying a time stretch you can experience them but I don't think anyone can hear the difference in pitch of stretching three or four frames over an hours time. I would however believe that most people would see a dropped frame or added frame once every 20 or 15 minutes unless you had a cut or black frames to hide them in.

As Jon said, the best solution really depends on the situation. The universally best solution of course is to genlock the audio and video but of course that's a very costly solution.

-Garrett

Adjusting the speed of audio can be done without causing gross artifacts, though people with golden ears will certainly disagree.

Perhaps that is the perspective here. While I certainly don't consider myself "golden ears", I was compelled into getting into video because of the generally deplorable state of audio for video. I've been seriously into video for only around 20 years, but more than twice as long as that in audio. And more than half of my video production (and all of my audio) projects are completely or partially music performances. But even for single-camera "talking head" stuff, I shoot at least a few minutes of audience reaction and other "B-roll" type footage to cover whatever I need to do with the primary footage.

..
The best approach really depends on the situation and the amount of time you have available.

The BEST approach is to avoid the issue all together by using cameras and audio recorders that either a: have good enough clocks to maintain sync over the duration of a typical shot; or b: use cameras and recorders that send and accept external sync so you can use one of the strategies to provide a common timebase to keep everything in order regardless of the shot length.

Very interesting discussion. Now I'm wondering if it would be better to either clip out or duplicate a couple of frames worth of Audio where there are undefined sounds such as applause then do a short cross fade to mask the cuts so that is is not audibly detectable.

-Garrett