I think #2 is the correctly-synced video.
I can’t really tell amongst 2,3 & 5, and the test is made more confusing due to the blur on the moving horizontal bar. Do the dark gray pixels on the advance end of the bar count, or just the light pixels that match the rest of the bar? It’s a complication.
The leading edge of the small dark grey section for the VERTICAL bar.
The main light grey section for the HORIZONTAL bar.
This is based off a video designed to help you sync up your audio and video, so I’d imagine the three separate visual cues are very deliberate. Personally, I find the vertical bar to be the most telling, mainly since (as a musician) I’m used to watching a conductor’s hands bounce up and down off a set ictus.
I’m really not trying to be a hard-ass about this, but when you are asking listeners to distinguish something as little as 40ms apart, it’s supremely critical that you know there are no additional delays or artifacts introduced or involved. “Appear identical” is not sufficient to ensure a valid test when we know that irregular delays are inherent in the Internet data delivery system.
Hmm, I was using the leading edge of the small dark grey section of the horizontal bar. I suppose that means my guesses might be off. That’s my excuse, anyway!
Without reading other responses, I’d say #2 was in sync.
No, I agree. I pulled them back out to double check with my original files, and everything looks as it should. The test isn’t perfect, but the files should be synced correctly.
Looking at the flashing graphic in the top left corner helped the most to see if it was synced to me. The bar going horizontal and the bar dropping vertically just confused me.
I thought 2 was the synced one, too. Others looked out by a little or out by a lot, but I didn’t note them.
I thought 1 was correct, although I see I’m not in the majority.
<comment deleted>
More sciencey stuff with numbers and explanations and malarkey. They should have just polled people on a random message board instead.
I just had a problem with the way the exact moment is indicated. I never actually see the frame when the thing is supposed to hit. It’s always right before it and right after it, with a blur effect. I couldn’t pin down the exact time. It probably would have been better to recreate the original video without any interframe blending.
Then again, it might be my monitor’s LCD refresh rate (which is different from the usual refresh rate. It’s the amount of time it takes a for a pixel to completely disappear once it’s been turned off. There may be another term for it–but it’s generally worse at higher monitor refresh rates.)
The little bit of flashing is probably the only part I can use, but the other stuff is so distracting. Why did they make the leading edge a different color, making it blend in with the blur?
Anyways, I absolutely cannot tell the difference if I’m not very closely scrutinizing the video. Just listening, and I couldn’t tell at all. Trying to analyze it:
Four and five seem way off. One appears to be early, based on the rest. But I can’t really tell between two and three.
I’m surprised at the person above who said that being a musician helped them. I think it hurt me. My brain seems to automatically attempt to make the sound match the visual. When I’m playing, I always feel like I’m perfectly in time with both the conductor and with the sound I hear, even though those don’t happen at the same time.
#4 was the only one I “eliminated for sure”! :smack: I think #4 has audio way early. (FWIW, I voted #5, with #1 also audio early, but this was guessing.)
Audio advance of about(?) 900 msec is definitely noticeable, as I learned watching Morse Episode 16. I almost need to close my eyes when listening to dialog. (What sort of software/operator error screwed up that video?)
I listened/watched all of them twice and while I couldn’t tell which way the delay was, the only one that seemed synched to me was #5 after deliberating between that one and #1. I think I would have been able to tell better if it was video of people talking, though.
FWIW Here’s my guess:
Video 1: -1
Video 2: 0
Video 3: +2
Video 4: -2
Video 5: +1
As a musician who uses computer audio equipment (in particular, live guitar effects and MIDI keyboards), anything over 10ms latency (20ms round trip with software monitoring) is too much - the disconnect between playing and hearing is disconcerting. So most people aim for 5ms input latency. But that is based on me knowing when I struck the note.
Si
Thanks much for that link, an excellent article, and I’ll try to get to your other references.
I think I know where you’re going with this, and if I’m right, I feel your pain and hope the industry can fix this kind of problem. I administer a small cable TV channel, too small to be able to afford outside consultants, and I am constantly finding obscure errors like this sync problem that I have to deal with. Most errors are in the marginally-perceptible range, and I often feel like I’m the only one who notices, but they bother me a lot.
Such errors could happen anywhere in the broadcast chain, from original video files to video server, to modulator, to cable hub, to consumer. Half of the problem is tracking down just where the glitch is, and sometimes it’s in more than one place!