How to get all my plasma + LCD monitors in sync?

[thevoicefactory]

Hi,

In our main room we are about to add an additional plasma and two LCD monitors. To make things easy they all have different delay times. The older plasma has a delay of 16ms, the LCD monitors have a delay of 8ms and the new plasma will have a delay of 12. Unfortunately it has been impossible so far to find matching monitors as there are no delays anymore of 16 ms.

Anyone know a good solution to solve this problem?

Thanks!
Michaël

[Frank Kruse]

first you need this:

www.syncheck.com

and then you have to adjust a master delay in your monitor chain if you allso want to have things in sync that don´t come out of PT.

There is no way of adjusting the sync for various *different* displays at the same time unless you use video-delays per monitor and then adjust a master delay accordingly. pretty pricey equipment these video-delays

frank.

[Don Barto]

Let's see... 1000 milliseconds divided by 29.97 frames-per-second means there are 33 milliseconds per frame. And you're worried about how to synchronize monitors that are lagging behind audio by 1/2 frame, 1/3 frame, and 1/5 frame, respectively? I would think under most situations, Pro Tools itself is also delaying the audio by at least that much.

But forget the math and the notion that we aren't supposed to be able to easily discern subframe lip-synch mismatches between picture and sound...and first just hook it all up and see what it all looks like...and then go for Syncheck to find a good average delay that will accommodate monitor latencies of 1/2 frame, 1/3 frame and 1/5 frame — if that's what it all turns out to be under real world conditions. Actually, those numbers sound small based on what I have seen and heard strolling around in the Big Screen section of Best Buy — many plasmas seem to be several full frames off — some, into the double-digits. Syncheck will tell you a lot about the real world. Just looking at the numbers supposedly involved, it would appear that 12-13 milliseconds (a little less than 1/3 of a frame) might be a good average — or maybe just go for the largest offset: 16 ms — and the rest of the monitors would then lag in the single-digit-millisecond tenth-to-a-quarter-frame range.

As for audio delay lines in the monitor chain (or dialed into your final Pro Tools master fader delay setting), if you are working with QuickTime or DV video in Pro Tools, can't you just dial in the closest appropriate quarter-frame movie offset and let that delay audio to compensate for monitor latency? Or am I mistaken as to what that setting actually does...haven't had to cope with that one yet on the monitor level. Soon.

Don "The Synch Nazi" Barto

[Richard Fairbanks]

can't you just dial in the closest appropriate quarter-frame movie offset and let that delay audio to compensate for monitor latency?

That works great for Pro Tools, but nothing else. Do you watch DVD's from time to time, or does a client want to screen a VHS? Do you ever need to layback your mix to video tape? I do sometimes. The best thing about watching the tape's video as you layback is that you can make that final QC check, that one last effort to be certain your audio is the best it can be. You can't do it unless your audio and video pathways both have the same latency. The way I see it, you must either use only CRTs or you must put delays in the audio monitors.

Fully pro post rooms really need to have eq's in the room monitors anyway, I suggest that you invest in good digital eqs (Parametric or 1/6 octave, whatever you need) that include delays. You can set presets for different eq curves and different delay settings, whatever you need, and cover your behind whenever you need to work outside of Pro Tools.

[Don Barto]

Richard, do the relatively low latency times described above seem typical to you? What I see in actual use tend to be much higher, particularly with plasmas.

[spigots]

I was able to come up with a simple but effective solution; and as stated above, you can only adjust for one delay without considerable investment.

The four issues for me were:

1. Delay to compensate for Video Delay
2. Ability to work with delay in the signal path
3. Ability to view VTR playback with delay in the path
4. A fool proof method (No way to accidentally print with delay in-line).

2 obvious solutions: Buy inline digital delays to place in the monitor path; or solve it with Pro Tools.

By eliminating master faders from our work-flow and using busses, aux returns, delay plug-ins and multiple input output paths, I solved it for almost free. The one thing we bought was Richard's Syncheck Box, a wise investment for anyone doing serious post.

Good Luck,

Frank

[EasyStreet]

Just recently finished a feature using a Dell widescreen LCD for video playback. I found it to have almost 70 milliseconds of delay, not the 12ms response time of the LCD's.

That's more than 2 frames. I have seen plasmas that are more like 5-7 frames off.
If it's a flat screen it will be FRAMES off, not milliseconds - so you will need some delay compensation.

I also solved it using delays in ProTools, but this doesn't fix it for your multiple out-of-sync monitors.

[Richard Fairbanks]

Hi Don. Like you, I think 1/3 and 1/2 frame monitor latency times are probably, well, not correct. Usually, solid state monitors have a visual delay of 1 to 2 frames, and plasmas even more. From what I can see, this is not going to improve anytime soon. Manufacturers are only beginning now to build in digital delays in their internal audio circuits, which do not help us at all but is good for home viewers.

I've had the good fortune of speaking with several people working within some very big companies, all trying to come to grips with this. I am glad that the companies are aware of the potential sync problems, but I am really saddened that twice in the last three weeks I have heard them proclaim, proudly, they are within the ITU recommendations for sound and image synchronization error.

There are a couple of ITU recommendations. The most recent is ITUR BT.1359-1. It suggests that for average viewers, audio sync is not distracting to them as long as it is within a window of 40ms ahead of image to 125ms behind image. This recommendation is univerally disavowed by any reputable broadcaster or audio person but is being cited as good enough by the manufacturers and software designers who have no interest in doing better.

Thankfully, many people realize this recommendation is total BS.

[Don Barto]

I would suspect the flat panel video monitor latencies we are all now learning to cope with are even a much bigger can or worms than many of us would care to imagine. And I seriously doubt that we in this field will accomplish much if we take video monitor manufacture's specs as gospel — after all, to most of the industry, Audio is just Video's dirty little friend...

It seems as though numbers like the 16 ms. delay spec quoted above might apply either to the time it takes for the actual plasma display pixels to excite and start giving off light...or for the scaling circuitry built into the monitor to processes the input video signal and rebuild it so it can be displayed given the monitor's native pixel configuration. But based on what I have observed wandering casually through the Large Screen department at Best Buy, I seriously doubt that 16 ms. is enough time for both to happen. What do the specs really mean?

Add to that the fact that the "delay" we see in plasma displays isn't totally a delay in the picture appearing — it's also a delay in the picture going away. Because (get this...) the way plasma displays work, the brighter the image (or componentent of the image) the longer it lingers on the screen. On a CRT, "brighter" just means applying more electricity; on a plasma, "brighter" means longer pulse width so the image has more time to build in optical intensity. So if you were to find an objective way to measure the lag in a typical lipsynch image on a given monitor, the lag might actually be meaurably longer for a bright flash of light image — like an explosion, or something similar that might get a hard sound effect. Very troubling.

And add to that the likelyhood that different scalling processing needs may actually yield more or less delay than others. For instance, to prepare a 24p 16 x 9 HD image for display on a typical 42" plasma HD monitor may require much different processing — and a correspondingly different overall video monitoring latency — than to show a 29.97 interlaced 4 x 3 Standard definition image. There could also be differences based on which video input is used, too. And could motion or the amount of detail also have a bearing? So having time delay in your audio monitoring chain may not be a set-it-and-forget-it proposition. I can't honestly say at this point.

I would like to see video moitors that actually take all of this into account and either provide about a dozen channels of audio time delay that track along accurately with what the monitor is being used for, or settle on some standardized, easilly attainable delay that all flat screen monitors would exhibit — sort of like "projection synch" in the film world.

To wrap it up, remember the first time you ever saw a live or videotaped video image that was out of synch...and when you complained to the editor or the video engineer, he or she said: "you're crazy, how could it be out of synch!"

Maybe I'm just obsessive about picture/sound synchronization...that and playing in tune.

dB

[Tiago Silva]

The 16ms number given is, as far as I know, related to how fast the pixels can change color and not to overall delay. That "latency" metric serves as an indication of how "blurry" the panel will look with things moving across the screen.

What matters to us is how much time the receiving interface and the image scaler take to put the image into the LCD panel. If the interface is DVI or HDMI, it most certainly has a buffer, and if the interface is analog (D-SUB, S-Video, composite -gasp!- or components), it has to be digitized. Then, if the input image has a different resolution or aspect ratio than the native resolution of the panel, this digitized image has to be scaled, resampled and filtered. Sometimes scalers also do noise-reduction, line-doubling, de-flickering and other shenanigans. All that adds to the delay, but I've NEVER, EVER seen any review mention how much total delay a flat-screen TV had. Of course, I'm not into home theater and my monitor is a CRT, so I haven't read much.

So, to minimize the problem, feed the panels with DVI, disable as much of the image processing as possible (even stretching an anamorphic picture will add delay), to the point of not resizing the image at all, and avoid plasmas like the plague.

Something I have always wondered: doesn't delaying the audio to match the video interfere with mixing? Meaning, isn't there a disconnection between what you do and what you hear? Wouldn't it be better to start the picture earlier and then delay the sync signal to ProTools (like, delay the LTC coming out of the video player?). Then ProTools would sync LATER than picture, with the desired effect of compensating for video latency without adding audio latency. Of course it wouldn't work if you were inserting audio under the video, but then if you had the delay on a patch, you'd repatch the delay between the audio return from the recorder and the monitor out.