Vast differences between exposure tools on VA & Ursa 4.6k

Hi, Does anyone know why I'm getting different exposure reading from the Video assist 4k and the Ursa Mini 4.6k?

No LUT on each

Zebras both set to same amount

Simple, false color on ursa is on sensor data/ raw data 12bit range , va4k read sdi out 10bit of ursa which is a bit shorter range than original data.
Is not only for ursa, I remember that also in my old production camera 4K I had similar problem but less pronounced be cause it have less dinamic range than ursa.


Inviato dal mio iPhone utilizzando Tapatalk

Simple, false color on ursa is on sensor data/ raw data 12bit range , va4k read sdi out 10bit of ursa which is a bit shorter range than original data.
Is not only for ursa, I remember that also in my old production camera 4K I had similar problem but less pronounced be cause it have less dinamic range than ursa.


Inviato dal mio iPhone utilizzando Tapatalk

Ahh ok thanks. Which one is correct then? Based on image attached

Actually shouldn’t 50% grey etc be the same what ever bit depth? 8, 10, 12 etc?

Simple, false color on ursa is on sensor data/ raw data 12bit range , va4k read sdi out 10bit of ursa which is a bit shorter range than original data.
Is not only for ursa, I remember that also in my old production camera 4K I had similar problem but less pronounced be cause it have less dinamic range than ursa.


Inviato dal mio iPhone utilizzando Tapatalk

Ahh ok thanks. Which one is correct then? Based on image attached

Camera give you info sensors about raw recording, va4k about eventual prores recording.

At same time if you work on edge of red burning highlights, you can recover with raw 12bit, and you cannot recover with 10 bit prores.

This is The reason that you can setup zebra not only at100% but also lower value.

18% of gray is the same on 8-10-12 but, but ...
we not read a 8-10-12bit data, we have a 16 bit linear data sensor with a translation to 12bit/10bit log curve which not put in the same position the same lighting value.

The small difference is in the log curve conversion that move a bit the gray point measured.




Inviato dal mio iPhone utilizzando Tapatalk

Can the Ursa output be set to video/709 without using a lut? At first glance it looked like a superwhite issue which might explain a stop of difference at the top. However mid grey on one is white on the other so there is a gross level mismatch. LUT in the VA?

I was shooting in ProRes 422 for a client yesterday, in a conference room with big windows.

Of course, highlights are going to be blown, so I mostly try to protect the subject's skintone/exposure. However, it really had me wishing for a more nuanced False Color scale.

I'm not a fan of zebras, just visually. They're so intense it's hard to get a good idea of what the frame looks like. I'd much rather pop False Color on real quick, rather than go back through the menus to turn Zebras on/off (already have my Function buttons programmed to other things).

However, the False Color's red does not necessarilly mean something is blown out. That can be difficult to use as a reference for what's blown out and what isn't. Is there any reason why there is not a specific color that is 100% loss of highlight detail (or shadow detail, for that matter) rather than the red starting at 95%?

In scenes with inherent extreme ranges from dark to bright, at 12-15 stops dynamic range, you will clip or crush somewhere. Perfection is a goal, but it can be unattainable without control of the light. In other words, perhaps difficult not to clip.

You have some control where the red colour displays. Red does not mean clipping absolutely. No red does not mean you are safe absolutely. That’s because one channel or colour can clip without red appearing. Experience is your best guide.

If you want to get closer to absolute, set your camera to 100%. Frank Glencairn normally uses 95%. I was 85%, but settled on 90% in the last year. Having the threshold lower than 100% helps prevent the situation where one channel clips without the OS letting you know. Many times I’ve had some red when shooting but not in post.

A combination of a lower setting for the red threshold and the fact that you will see yellow before red is reasonably nuanced. I’m always confident that yellow is great and a little red is likely good too. A field of red is likely bad and you need to alter the aperture or shutter angle.

Apologies I can’t demonstrate the differences on an external monitor as I only have the BMVF. I solely rely on the BMVF to show False Colour as I use the monitor to control framing on a shoot. When I have checked False Colour on the monitor and the BMVF, they look the same.

Sent from my iPhone using Tapatalk

Can the Ursa output be set to video/709 without using a lut? At first glance it looked like a superwhite issue which might explain a stop of difference at the top. However mid grey on one is white on the other so there is a gross level mismatch. LUT in the VA?

You just record in raw with video profile, internally record raw log and external output is video (rec709) output. But if you work with prores, no. Don’t forget that if you put a rec709 lut only on sdi put the contrast difference is so great that you can see a real different picture. Anyway check in monitor menu setup of ursa, you can output out on Sdi and see log on internal monitor.


Inviato dal mio iPhone utilizzando Tapatalk