PDA

View Full Version : NewB Live Switching Question



cccvt4
04-24-2006, 08:43 PM
We have been attempting to switch multiple live camera's with text (lyrics) overlaid. Text is generated on a separate computer (Easy Worship) and input through iVGA. The Chroma key results are poor due to the spill around the characters AND we can not seem to switch the live video underneath.

DSK shows the iVGA as a source (under the no Alpha sources) but this seems pointless since it just shows the iVGA image. Like I said we are new at this… any suggestions appreciated. Oh – VT4.6C :help:

billmi
04-25-2006, 07:20 AM
There is another thread on this topic, if you search for it that goes into more detail but in a nutshell....

If you are outputting to a projector, there is another keyer (it's luma, not chroma) in the setup for VGA output. If you do the key there, it leaves the DSK free, and it doesn't downscale your iVGA video to video res for the key.

appsyscons
05-02-2006, 01:35 PM
We are a church in SW Florida, doing exactly what you want, using VT for about 6 weeks now.

We have our song & power Point presentation software (we use Song Show Plus) running on the PC that runs VT. We have 2 video boards, each with 2 screens on them. Screen 1 & 3 are for VT. Screen 2 is running SSP and screen 4 is the output of SSP. Using VGA capture, we grab the screen 2 stuff, key out black in the utility setup for VGA capture, and then use this as input to the DSK. We switch between camera 1 and camera 2, while leaving what ever is coming from SSP on the lower third of the projected output, through the DSK. Works perfect, no color spill at all!

The only problem we have is the delay in the video. We are using Canon XL H1 cameras with genlock from an external black burst generator, and seeing, 6 frames of delay in VT. We are expecting 2 frames of delay in VT with the genlock. Our system is in TX today for diagnostics, and hopefully repair.

Cineman
05-02-2006, 08:07 PM
The only problem we have is the delay in the video. We are using Canon XL H1 cameras with genlock from an external black burst generator, and seeing, 6 frames of delay in VT. We are expecting 2 frames of delay in VT with the genlock. Our system is in TX today for diagnostics, and hopefully repair.
This is an easy one to solve. You don't have the cameras genlocked to the correct source. Toss the black burst generator. It is absolutely doing you no good. Instead, genlock your cameras to the VT output.

Not that familiar with the XL H1 beyond the fact that they have the genlock connector. Check your manual for an advanced setting for vertical sync. Go there and advance the vertical setting until you get a glitch (sync roll, picture tear). Then back up one click.

If you are getting a measured six frames of delay now, you should halve that to end up with three.

Nes Gurley

jsanfilippo
05-03-2006, 10:30 PM
This is an easy one to solve. You don't have the cameras genlocked to the correct source. Toss the black burst generator. It is absolutely doing you no good. Instead, genlock your cameras to the VT output.

Not that familiar with the XL H1 beyond the fact that they have the genlock connector. Check your manual for an advanced setting for vertical sync. Go there and advance the vertical setting until you get a glitch (sync roll, picture tear). Then back up one click.

If you are getting a measured six frames of delay now, you should halve that to end up with three.

Nes Gurley

My DSR 390s, with D50 CCUs, also seem to not be locked properly, cause I certainly don't have the suggested 2 frames. And I tossed my BBG a couple months ago... but I didn't follow the vertical setting step.

Let me make sure I understand.... when you say "genlock to the vt output".... you take any of the composite program out feeds, and send that into the "ref in" on your CCU?

And I need to adust vertical on each CCU... NOT on the VT, right?

Jim Capillo
05-04-2006, 06:19 AM
And I need to adust vertical on each CCU... NOT on the VT, right?

That's correct.

If you were to bring the VT into a system that already has BB (such as mine is here), you simply loop BB to the VT from either the last piece of equipment or a BB DA. Then you would time the VT to the system.

Either way should work.

The only thing I'm wondering is if your processing of the DSK signal through another graphics board is introducing a delay. Typically, anytime you process a signal through another piece of equipment, a delay (might be small or large) is introduced. I don't know enough about your setup to give you a definitive answer, though.

Jim Capillo
05-04-2006, 06:32 AM
HERE (http://jehoo.netian.com/tech%20brief/brief%201/Tektronix/Ntsc/genlock.html) is a pretty thorough explanation of genlock and how to genlock a studio facility.

Cineman
05-04-2006, 11:36 AM
My DSR 390s, with D50 CCUs, also seem to not be locked properly, cause I certainly don't have the suggested 2 frames.
First of all Jamie, I have never read anyone from NewTek on these forums that have suggested 2 Frames as an absolute. My recall is that it has always been two or three frames. I think the mathematical determination indicates as low as two, but I have never found a single instance of that.

For my initial testing, I only read the results to the frame. I later found that I could read results to a field. I have found 2.5 frames. As I know it, the one less field means nothing positive, as the VT will then wait on the next vertical interval to maintain proper field dominance.


And I tossed my BBG a couple months ago...
Thanks for including that. It allows me to know that the VT is your system, which I believe is best.


but I didn't follow the vertical setting step.
Don't think many have. I also don't know how many of all the camera possibilities have that capability. I have good expectations that your DSR 390s do since the EVW 300s do.


when you say "genlock to the vt output".... you take any of the composite program out feeds, and send that into the "ref in" on your CCU?
No. I did not mean that. I believe that would actually add delay, but perhaps wash in the end. The camera would still be sending video according to its internal clock instead of the VT's. The CCU would then delay that until the proper vertical interval, and then pass it as a genlocked signal to the VT. It is hard to know if that hold delay could be more than the VT's TBC genlock time.

All of the CCUed cameras that I have done is from VT Program out directly to the cameras Genlock In connector. At that point the camera is genlocked to the VT's clock and they will be in sync.


And I need to adust vertical on each CCU... NOT on the VT, right?
I doubt that the CCU can access the cameras advanced menu settings. If not, you would need to do that at the camera.

Nes Gurley

jsanfilippo
05-04-2006, 11:55 AM
Thanks, Jim and Nes, for your thorough replies.

So, Nes, I was going to reply saying that it would seem very strange to me that a CCU isn't designed to "communicate" sync to the camera, but rather, it TBC's the output of the camera. But then, as I think about it, this isn't that strange.

Cameras are designed for BROADCAST, not IMAG, right? So in a broadcast situation, as long as all your pieces are genlocked together, it doesn't matter if they are a few frames behind, as long as EVERYTHING is a few frames behind, right?

I just can't imagine that in order to lock everything down, I've gotten run coax lines BACK to my camera, separate from my CCU cabling, and plug in another cable to my camera. Seems wacky!

Since I synced (or tried) my CCUs to my VT, I saw a slight reduction of latency. But it's still there. I really should measure it and find out exactly how off it is.

I was looking into my CCUs today, and there was Hor. phase and SC Phase adjustment, as you'd expect, but I didn't find any vertical adjustment.

We do have a broadcast engineer in amongst our people - I've asked him to come take alook. The guys that did the install are wizards as cabing and clean install, but they aren't engineers. :P

Jim Capillo
05-04-2006, 02:15 PM
I must preface this response by saying that this is not to challenge anyone's thought, intellect or knowledge on any given matter discussed here. I'm just trying to help out.

Jamie,

Every professional CCU (multi-core ot triax) that I've ever worked with over the years (and it's been more than a few), has had complete control over the camera it is connected to, h-phase, sc phase, iris and setup adjustments included. If your CCU has, as I suspect, a multi-core cable going to the camera, you DO NOT need to run a separate coax to the genlock in to lock up the camera - it is done through the multi-core cable. As a matter of fact, the multi-core might override the genlock in on the camera. You will not be able to adjust a camera by it's CCU adjustments if it doesn't have a genlock reference - it needs BB. The CCU doesn't time base correct its camera in the most correct technical sense, it just brings two or more pieces into sync and allows the cameras to be phased together.

99.9% of the time, the CCU's, switcher and TBC's would be locked to an external reference BB Generator. This may be free running or locked to the switcher or maybe a piece of equipment that does not have genlock capabilities. You may have to feed a non-genlockable piece of equipment thru a TBC that is genlocked to bring it in lockstep. Once all the equipment is genlocked together, then you can adjust h-phase and subcarrier phase, but those adjustments are not causing your delay problem - bad (or missing)genlock could, as well as the projectors in your system.

I will be glad to take this thread off forum and help you out anyway I can.


Good luck !

jsanfilippo
05-04-2006, 02:35 PM
Jim,

Your response is very much appreciated.

I would like to take this offline and talk it through with you.

I'm just wrapping up my day now, and am out of the office tomorrow.

Can you email me ([email protected]) and we can dig a little deeper outside the forum?

My CCUs are multi-core (26 pin). I had always thought that "everything" was transferred down that 26 pin cable, so that I wouldn't need to plug into the camera directly.

I thoroughly believe that my problem is, as you say, bad, or missing genlock (improper settings! I'm not an engineer - I'm a techie/pastor/director!). I know the projectors add a bit... but when I run 1 camera, direct from CCU to projector (skipping the VT) I can percieve no latency. So although there may be some there, there's definately some introduced with the VT. The combination of the two probably makes it perceivable. If I can reduce VT's as much as possible, I presume I'll be happy.

I've gotten no response from the guy here that I asked to come take a look at it - he's hard to nail down. So any assistance you can talk me through is very helpful.

wvp
05-04-2006, 07:55 PM
You might try genlocking everything and using a regular TV connected to the VT output first. after your satisfied with the sync, then connect the projectors and see what the sync is.
The VT (and all computer switchers) do add some delay - it the nature of the beast. Projectors have been known to as well.
Last week I did a gig for the first lady and the guy switching (I was on camera) opted to use a very basic switcher to minimize delay rather than one that would have given him "smooth switching". The cameras were not even synced together much less genlocked!

Jim Capillo
05-04-2006, 08:04 PM
Last week I did a gig for the first lady and the guy switching (I was on camera) opted to use a very basic switcher to minimize delay rather than one that would have given him "smooth switching". The cameras were not even synced together much less genlocked!

Wow... that must've wreaked havoc with the vertical interval..... I have had good luck minimizing delay by using older switchers that don't do A->D D->A processing. If it has any type of squeeze or digital effects, you're looking at delay.

nevmoor
05-04-2006, 08:24 PM
I am enjoying the info about genlock!! This is great stuff! Even though I have done much of this in my daily grind, it is good to have the info brought up to get the brain cells thinking about this again. Most of the time as was stated, for broadcast all you need is to have everything clicking along together without concern for overall latency.

Our Eiki projectors put about 2 frames of delay on our IMAG alone, which is why I don't use the VT to feed IMAG!
Jamie, did you ever get the S-Vid from your CCUs into the VT properly??

wvp
05-05-2006, 06:58 AM
Wow... that must've wreaked havoc with the vertical interval..... I have had good luck minimizing delay by using older switchers that don't do A->D D->A processing. If it has any type of squeeze or digital effects, you're looking at delay.
Oh yeah, most switches included a quick roll of the video. I couldn't believe for an event like that one they would use a "switcher" like that. On the other hand at least when she spoke the delay was almost non-existent!