7" Video Assist HDR Color shift still an issue? 3.1

The UM4.6K has the Film to Video LUT selected but not applied anywhere and the Dynamic Range is always Film. Display 3D LUT is always Off everywhere. There’s no option in firmware 4.8 to explicitly record a LUT in the file as there is on the BMPCC4K for example.

Clean Feed was set On for the SDI Out; I used both On and Off as mentioned in the previous post just to see if the Zebras on the BMVA12G7 matched the camera as they did.

Rick, just to clarify, the image appearing on your BMD Video Assist is the low contrast image of BMD Film, yes?

Further test shows the red tint although greatly reduced is still there. Unlike before its now less pervasive and can at least see other colours more easily and differentiate what should be white unlike before. It looks like it has boosted the blue and green channel against the already high red channel. I wish another firmware will try and reduce the red or magenta a bit further or include a "Tint" option like they have in their cameras.

Rick, just to clarify, the image appearing on your BMD Video Assist is the low contrast image of BMD Film, yes?

Jamie, yes the image is set to use Dynamic Range Film without LUTs applied.

BMVA12G7 has the 4.6K Film to Video LUT selected, but Display 3D LUT is Off. The UM4.6K has Clean Feed On for 12G SDI. THE 4.6K Film to Video LUT is also selected on the camera but Display 3D LUT is also Off. Zebras are identical.

This is the way to ensure the camera and the BMVA12G7 are independently calculating their respective Zebras based on the clean feed.

The UM4.6K LCD has Zebras On and Clean Feed Off so I can see the camera Zebras. For the Front SDI going to the BMVA, I can have Clean Feed On (or Off and all other camera options Off), so I’m using the BMVA to generate Zebras and they look the same as the camera LCD Zebras and the BMVA12G7 Zebras. Same sensitivity to changes in my iris.

I was surprised to see False Colour variations on the BMVA12G7 which is peculiar because the Zebras are consistent. Sometimes I want to see False Colour while shooting outdoor events under changing light conditions; I have used the camera LCD or the BMVA with False Colour (to maintain pink and green skin) a few times that way but I’ll use the BMVA12G7 Zebras for clipping while the larger and brighter image is good for framing and focus.

It’s been a learning experience going through these options and it also reminded me what a capable camera the UM4.6K is while I’ve mostly shot with the BMPCC4K in the last year. So glad the older camera is functioning very well for the ways I use it: 4608x1920 and 3840x2160 with the 1.33x-65 Anamorphot for client shoots. Only option I have to record CinemaDNG lossless or ProRes 444 XQ for special occasions. Knock on magnesium (but gently, eh?).

To do the tests today and have both cameras feeding HDMI and SDI into the BMVA12G7 and monitor either camera with one button to select the HDMI or SDI feed is nice. I’m thinking if the music festival happens this year I could use both cameras for different angles and monitor both via the single BMVA12G7 since SDI can run a good distance. Let the UM4.6K and my assistant capture more of the stage and shoot with the BMPCC4K to handle the performers’ closeups.

Ah, so false color is looking different while Zebras aren't. That is the crucial clue here which, I think, cracks the case.

Rick, I'm fairly sure the difference between what you're seeing and what Mark and I are seeing with our setups is because:

1) In Gen3 (and Gen4) of BMD film on the original 4.6K, the log curve isn't nearly as flat as Gen5 in the UMP 12K that Mark and I are both looking at.
2) The Zebras on the Video Assist operate in large increments, they don't clearly show differences less than 10%.
3) The false color has smaller increments, especially at the top so differences there do show up.

Your camera and Video Assist are working exactly the same as ours — the zebras you're seeing are reading the log signal as any other monitor would rather than reading any kind of direct metadata from the camera, it's just that the Gen 3/4 curve is not so flat. So, the zebras on your Video Assist will look similar enough to those on the camera LCD but when viewing the false color on the Video Assist it's clearly different from the false color on the camera LCD. While, for Mark and I who are viewing BMD Film Gen 5, the curve is so flat that even the Zebras look markedly different from the camera LCD because they don't get triggered until much lower in the scale with that flatter Gen5 curve.

And, that would also explain why Mark remembers it being different before he had the new 12K, as before that he would have been sending out the less flat BMD Film Gen3 or Gen4.

In other words — Nothing has changed about how the camera and Video Assist read signals for false color and zebras. All that has changed is relative shape of the log curve between BMD Film Gen3/4 and BMD Film Gen 5.

I just updated to firmware version 3.4.1 and straight away the red or magenta cast was pretty bad, about as bad as it was with the firmware before the previous one before it had the first colour correction. After leaving it on for awhile and switching it off and on the red cast slowly lessen. Not saying the red cast is gone or if it improved over previous version. I don't believe the VA need a warm up to correct the colour cast but it did needed a run-in period with the latest firmware to get settled.

Interesting observation. Thanks for the suggestion.

Going against what I previously said I think the VA does need a few minutes warm up for the red/magenta cast to recede slowly. Its not completely gone and still look a bit warm but certainly look less red/magenta than when start up from cold.

Ah, I posted an new thread without seeing that this thread is pretty recent. Anyway, I just bought the Blackmagic Video Assist 12G 7" and was massively disappointed to see a pink tint on it. I was using the older 3G version of the same monitor and I was extremely happy about that monitor.

Also, like someone mentioned here before, the pink cast only seems to be on the video signal coming from my BMPCC4K. The text and the other graphics that's inherent to the monitor is perfectly fine. God! I hope they solve this. This is a really expensive piece of kit and I really hate to be running around trying to return and alll that hassle.

I have the exact PINK magenta tint to the image, the image doesnt record the tint.

Im in two minds do i just return it and forget about BM and look elsewhere or just know that the imgae is a little rose.

ill get in touch with BM and find out what's going on with this issue. #pinkasF^$%

Roger

I have the exact PINK magenta tint to the image, the image doesnt record the tint.

Im in two minds do i just return it and forget about BM and look elsewhere or just know that the imgae is a little rose.

ill get in touch with BM and find out what's going on with this issue. #pinkasF^$%

Roger

If BMD brought out the same monitor calibration feature it did for the BMPCC4k, 6k and 6k Pro in firmware version 7.3 for the Video Assist then this issue would be solved. I'm waiting for that to happen.

My BMVA12G7 is very close to correct balance when I judge with my eyes. I match the monitor on the BMPCC4K to the colour of the BMVA12G7 with that calibration tool as close as I can. Would be expecting to see that calibration tool appear as a feature on the BMVA12G7 one day. In the meantime, I find my BMVA12G7 to be very good so I feel lucky with that.

I’m using the very good Display LUT for Video now that Gen 5 has arrived on the BMPCC4K; still record BRAW with Dynamic Range Film anyway although when recording ProRes I might record with the LUT for Video embedded.

Applied the latest firmware update 3.4.3 but didn’t notice an improvement. I have calibrated the BMPCC4K display previously using the supplied menu option. The 3.4.3 color calibration change appears to be behind the scene as I didn’t notice any new menu option for the user to control the calibration as on the BMPCC4K. I’d say my BMVA12G7 still appears to be sightly warm compared to the BMPCC4K

Same here. The VA HDR 12G screen is too warm. The magenta cast seem to get less after about 30mins but the warm cast remains. BMD could at least try and give us the screen temperature calibration?
Also could do with a power save option like that seen in the Pocket cameras to save the battery.

Difference (small) between internal and external false color is correct, internal false color work on 12bit signal of sensor, VA work on 10 bit out of hdmi or sdi (ursa), zebra is not so sensible to see difference but false color can see difference.

You can see not clipped signal in camera which is clipped in out signal from camera. Old log curve was optimised for 10 but cannot do miracle if we need more bit, newer gen 5 do better and is optimised for braw 12bit.


Inviato dal mio iPhone utilizzando Tapatalk

Any more news about UM Pro G2 getting Gen 5 colour?
Really puzzling when you consider G2 was announced only in March 2019 while the BMPCC4K was announced a year earlier in April 2018 and it got Gen 5 colour recently also when you consider the latter is suppose to be the cheaper model in BMD camera line up

Using a non HDR Marshall CV355-10X camera, I was also having the pink tint on my 12G 5". Seems that the problem is that the Video Assist fails to detect that this is not an HDR camera (or that the camera incorrectly reports that it is an HDR camera, I don't know).

I installed the latest firmware (3.5), which adds a new option in settings > Setup > page 4 > Record HDR format override. Selecting Rec.709 fixed the problem for me. Saturation was a bit low, I boosted it to 70%, it looks pretty nice now.

I still have some regrets:
- as someone said earlier, the micro BNC is not so practical
- it's a real shame that the Video Assist is unable to charge the batteries, this is the biggest downside IMHO
- it would be great to be able to delete some clips rather than formatting the whole SD card

Still a great tool overall