How to setup - Reference in/out

Hello

We have four Ursa Broadcasts with fiber converters, wondering how to connect the Reference In/Out to help with our IMAG.

What signal get's connected to the first camera, then I assume the other 3 get daisy chained together and finally you connect to the Atem switcher, then do you send the switcher's signal back to Camera 1 'Ref In' or is that a different input? If there are instructions on how to do it please point me in the right direction.

Also, we have one Ursa Broadcast that's wireless, how/can we loop that into the reference in/out as well?

What's the difference between Reference in/out and Genlock?

Thanks

[Gary Adams]

Hello Martin. These are good questions. For cameras, using a reference signal is the same thing as genlock. It means sync generator lock which is what a reference does. There are several ways to do this but what I would recommend is not using the reference cable at all for the cameras. Set the camera reference to Program video. This will lock the camera to the ATEM program signal. Then in the Line timing to -90 to advance the camera to the input of the ATEM. This will buy you one less frame of latency through the system. The wireless camera cannot be referenced unless you have some sort of wireless system that does feed video or sync back. If you want to use a tri-level sync reference, I would recommend feeding each Studio Fiber Converter separately from the sync generator. This is better than using the reference loop out to the next converter. If there are other devices in your system to be referenced, then you should provide one to the ATEM as well. I hope this helps. Stay safe.

Regards, Gary

-90?

[Gary Adams]

In order to use minimum frame synchronization in the switcher we need to advance the camera ahead of the processing delay. Most ATEMS have processing of about 4 to 10 lines. The Studio Fiber system has about 35 lines each way. That’s about 80 total but to be safe 90 is a good number. Without the fiber system about 15 is good. Set like this the delay through the system from the lens to the ATEM SDI out will be under a frame.

Regards. Gary

and how do you change that setting is that on the sync generator, the switcher or the fiber units?

Thank you, I will try that.

Can you go more than -90 in case latency still exists in the IMAG?

[Gary Adams]

Hello Martin. We are talking lines here not frames. This setting will insure the minimum delay through the system from camera to ATEM SDI out. If it is much lower than this, there will be an additional frame because the camera timing will be later than the ATEM causing the one frame delay. It cannot be used to do something other than that.

Regards, Gary

Gary, I want to time my URSA broadcasts to the router (videohub) so we have less latency switching between them on the router as well.
I have a BM sync generator feeding the Videohub and the ATEM. I use the 12G fiber converters between the router and the cameras, with PGM feeding the video return to the cameras. Is it possible to time to both the router (videohub) and the switcher?

You could only time them both if by some miracle they both had the same amount of resistance in the line. Switchers haven’t required manual “timing” since analog. If the goal is a seamless router switch, then definitely time it at the router. Genlock alone is insufficient for a clean switch, the signals must be timed and even then you’ll probably only have a clean switch between like sources. Differences in anc data can cause jumps even in perfectly timed signals.

Good Luck

[Gary Adams]

There is no latency in a router. It is best to time the sources to the device that needs it which are switchers with frame syncs on inputs. Gone are the days when we used to time to routers down to one nanosecond. Even if perfectly timed, routers without some sort of clean switch design will have a glitch when SDI is switched.

Regards Gary

Care to place a wager? We can set it up old school Twilight Zone style. If I can make 10 clean switches in a row I get a million bucks and a Malibu, if I glitch, you get to cut my finger off. You can pick the finger if I get to pick out the color of the car.

Good Luck

[Gary Adams]

I’ll happily defer to you Howard. Stay safe.

Gary

[Gary Adams]

Actually there is nothing wrong with timing to the router exactly. The timing point should be maybe 11 plus lines ahead of the ATEM so you are ahead of it’s processing delay. You can make your own exact calculations if you want. If you have the test equipment to do this then it will be the best method. Going deeper into timing, it is possible with multiple adjustable genlocking sync generators, the reference to the router and ATEM could be offset. For many this is not necessary but for some larger applications requiring closer timing this might be useful.

Regards. Gary

On top of that, if it's imag do you have an IMAGE pro or scaler, and is that receiving ref?
Most LED wall processors don't have a ref input. absin / ROE / Windvision. MITS used to.
Not sure if they still do.

You will lose frame(s) in this food chain unless you ref them together.
All you can do is limit as much as possible, but there will always be some latency from video to audio.

Hi. Old post, but this the only documentation about timing the cameras.


Genlock from Sync gen (tri-level 50hz) going into atem ref and IMAG scaler genlock (ascender48).

The Ursa G2 broadcast with fiber converters don’t need this genlock? They lock to the pgm return from the atem, which is locked to the genlock right?

The camera’s should be advanced 90 Lines with the fiber converter system.
This way, the cameras should pass through the atem with 0 frame latency and at the same time stay synced with the Ascender48.

If no fiber converter used, then 15 Lines advanced timing should be used.


Have i understood this correct? And is it still best practice to sync the cameras to pgm return and not the ref on the actual camera?