Re: Does Video Processing Always Take Longer Than Audio?
- From: Del Mibbler <mibbler@large>
- Date: Mon, 05 Jan 2009 13:25:21 -0500
dgates <dgates@xxxxxxxxxxxxxx> wrote (in part):
On Sun, 04 Jan 2009 13:16:29 -0500, Del Mibbler <mibbler@large> wrote:
Unfortunately, people are much more sensitive to errors where video
lags audio since that is the reverse of our normal experience. There
are sveral "standards for what is an acceptable range of error, but
one common one is that audio should not lead video by more than 15 ms
or lag it by more than 45 ms.
That's an interesting number. I have a setting on my receiver where I
can adjust for sync problems by adjusting a delay of up to 200ms (at
least, I remember the number being 200, and I think that the unit is
"ms." It goes from 0 to 200 of something).
I've always wondered, if I see some apparent sync problems, what units
I should use to try to adjust. Should I start by adding a full 100?
Or should I start by adding 30? What number is so small that I
shouldn't even bother tweaking any further?
For example, if my first attempt is to add 30 ms, and it seems to help
a little bit, should I then add 10 ms more?
Sure, why not? Or adjust in both directions until you see that the
lip sync is off, noting the delays, then set it in the middle.
One other problem I have is that, while it's easy to see that the
audio isn't quite syncing with the video, it's harder to get my head
clear on which way it's off. I find myself actually pausing and
saying to myself, "Okay. I... SAW the door close... THEN I heard
it... So, that means... the audio is, um... BEHIND. And I've just
adjusted the Lip Sync correction to 60. So next, I should try...
um... " 30? 45? 55? 70? 90?
Since unintended delays are nearly always more for video than for
audio, and since late audio is more acceptable than late video, a "set
it and forget it" solution might be to add a 45 ms delay to the audio.
Then the incoming video could be delayed as much as 60 ms before you'd
notice it.
There's another, more insidious problem. Research has shown that even
when lip sync errors are small enough that most people are not aware
of them, they subconsciously are less trustful of the speaker.
Suppose you're a network that wants to push a particular philosophy
but still appear to be fair and balanced (to coin a phrase). You
could give both sides equal representation but add a small lip sync
error when the opposing side is speaking.
Many years ago I noticed a less subtle approach used on CBS Reports,
the forerunner of 60 Minutes. They gave spokespeople for both sides
of an issue roughly equal time, but the speaker for one side would
make his points calmly and authoritatively while the other would talk
wildly, flap his arms and basically act like a loon. Think George
Will vs. Carrot Top. Or, for Monty Python fans, imagine the Sensible
Party vs. the Very Silly Party. A little lip sync error could slant a
debate between evenly-matched speakers, such as Stephen Colbert and
his Formidable Opponent.
Del Mibbler
.
- References:
- Does Video Processing Always Take Longer Than Audio?
- From: Ricky Jimenez
- Re: Does Video Processing Always Take Longer Than Audio?
- From: Charles Tomaras
- Re: Does Video Processing Always Take Longer Than Audio?
- From: Del Mibbler
- Re: Does Video Processing Always Take Longer Than Audio?
- From: dgates
- Does Video Processing Always Take Longer Than Audio?
- Prev by Date: Re: Networked Media Tanks
- Next by Date: Re: Yes, kids today are that ignorant
- Previous by thread: Re: Does Video Processing Always Take Longer Than Audio?
- Next by thread: Networked Media Tanks
- Index(es):
Relevant Pages
|
