Barco ICMP image slowed playback

[Marin Zorica]

A great disadvantage for troubleshooting here, is that you have no other video input than ICP and Media Player on the 10S. Which both are integrated in the ICMP and the ICP doesn't help much with refresh rate (even though it should have the alternative checkboard, that wouldn't make much sense here). On its older "brothers", you could have used the DVI inputs to check whether alternative sources have issues as well. In this case, HDMI equals ICMP. Hence, the issue being on HDMI inputs as well only indicates that the issue is probably not with the storage. Yet, there seems to be an induced bottleneck somewhere, and one of the first ideas that came to your mind was to inspect and clean the contacts. And from the two levels of contacts, the "IMB" level would probably be the culprit.
There is an oxymoron here. If it's the power supply, or a solder gone bad, then one needs to question the availability of all voltages necessary. Even more, when voltages in question are not on the projector itself, but the ICMP.
Really interesting!

Yes, that was my logic too......even (just to add) with did test with both 4K and 2K dcp, and same also. And I do not think IMB contact would be problem, but ICP or ICMP, a part where signals are feed from ICP/ICMP to formatters.....that is only (for me) resonal path, because all other video signal processing either DCP or HDMI are done internal in ICMP.

I'm probably out of my depth to continue suggesting anything just being a projectionist that has no ICMP in my booth.

I'd be tempted to verify the problem also occurs with the 3G-SDI inputs too. Probably will since it sounds like some kind of processing bottleneck specific to when the ICMP is installed in the THAT 10s (definitely the weird part).

For forcing the issue to be as noticeable as possible, I would be tempted to display a 4K30fps or 2K60fps translating grid pattern on an input that supports it that rate, or some other moving test pattern where it is easy to spot frame timing issues. Point being to make it work as hard as possible while displaying content where you can easily see the drops. Lacking a pattern, any crawl would be a good place to look.

Maybe trigger an ingest while playback is rolling and see if the problem is gets even more exaggerated?

Being a computer that is still displaying it, but not at the intended frame rate, it might even be something like thermal throttling on the processor?... But your testing of it in a different projector seems to say otherwise, unless the cooling situation is better in that location.

If it was a laptop I'd apply new thermal paste, check temps/airflow, and check the memory. It's certainly not the kind of error that makes me think about bad electrical connections, except maybe at the RAM. Things like taking it off the network too... maybe there is some misbehaving network traffic that is bogging it down. Is there any status page that will show you realtime CPU/Memory utilization? I'm so used to having that in the taskbar on an DCP2K4 at all times.

I could use HDMI to SDI converter and give it a try, ok idea.
This idea with "hard" video is ok, but thing there is no "taskbar" memory viewer as I am familiar, except maybe in logs, which are not real time data.


Also I am suspicious about signal backplane that it could be issue, also, since it happened in all colors, if it would happened in just one I suppose that we would saw color issues on picture too. I was also thinking that there is some kind noise or something that could make ICMP clocks go wrong but I doubt any outside interference could do that...!?

[Ryan Gallagher]

On the clocking thought. You would need a 2nd projector/ICMP and dual projector license to try this... but you could attempt to have it sync with another projector acting as the master, if the issue went away then at least you might have a clue it was something related to the internal clocking. But it seems counter-intuitive that the issue would go away with that ICMP installed elsewhere, as you have tested.

https://www.barco.com/en/support/kno...y-back-dual-2d

I don't know enough about their syncing protocol to know you if could just slap a sync-generator on the input to mimic the master projector.

Edit: according to the relevant error codes in the service manual, apparently the sync cable hosts a form of LTC timecode. Would be a fun experiment sometime to see if you can spoof a master projector timecode with your own LTC software/device, or if there is other proprietary stuff running on that sync cable too it looks for.

https://help.rode.com/hc/en-us/articles/7838106121231-What-is-the-difference-between-SMPTE-Time-code-and-Timecode-LTC