Scopes won't show signal below zero

Hi all, I wonder why my scopes don't show me signals below 0/above 1023. Even when I switch the view from data to video level scopes. It looks more like a limiting of the signal rather than showing what's actually going on below zero or above 1023. Is this a bug or have I just set something wrong? In my example that Ive attached the parade is set to video level.

Screenshot 2020-12-19 at 16.24.32.png (257.07 KiB) Viewed 1231 times

Some cameras may deliver what is called Superblack or Superwhite. You can use these values by pulling them in with the wheels or a curve, but they are not supposed to be in a technically legal output.

Some cameras may deliver what is called Superblack or Superwhite. You can use these values by pulling them in with the wheels or a curve, but they are not supposed to be in a technically legal output.

That has nothing to do with this problem. In R16, you could see out of range values in the scopes, now you can't. Makes it useless for QC.

The screenshot looks exactly like what I see if I accidentily leave an old home made 64-1023 LUT selected in Output lookup table (Color Management). Clips the bottom off sharply, regardless of video or data level scope.

On the one hand, I have thought the same thing. On the other hand, I end up twisting my mind trying to figure out whether it actually makes sense now in light of the many colorspace/HDR changes in v17, and whether it was strictly accurate before or if it was at times misleading. (Could just be a bug of omission with all of the scope configuration changes they made for v17.)

The BBC's QCTools is free (I think there's a paid version that's inexpensive, with additional options), and it shows graphs over time of min, max, avg, for Y, U, V, etc, that seem like they might be better for QC anyhow. And it would seem to be necessary for QC to occur on a rendered output, not before that because of potential changes that could occur in Delivery options.

It'd be interesting to hear from BMD if the change was purposeful or accidental, or if they're anticipating a more purposeful QC tool that's similar to QCTools.

OK well its also useful to know when I am grading that I am sending the signal out of bounds and by how much. There are other scenarios too that rely on knowing where the signal is BEFORE it is rendered. This NEEDS to be fixed!!

That scope looks to me like the recorded signal has crushed blacks, and the black level has been lifted to show that.

Dear Wayne, I couldn't find those tools on the BBC page, and I suppose you don't mean these do you?
textav.gitbook.io/textav-components/unsorted/qctool

Can you kindly post the link?

Sorry for the derail.

I thought they were from the BBC. Maybe the BBC is just a user of them. I got mine from the Apple App Store, but they originally come from: https://mediaarea.net/QCTools . I'm not a professional QC person, but it does look useful. Sort of like Fairlight's Loudness History except for lots of stuff. Attached is an image from the Apple App Store.

It uses FFMPEG under the hood and they recommend that if you have a lot of video to process you might batch it with FFMPEG and then use QCTools to view the results. On my laptop (8-core), it's analyzing a 1080p24 H.265 MP4 at about twice as fast as realtime.

I've found their MediaInfo app to be very useful for looking at all of the metadata in a video file.

Thanks, found it! I'll give it a try.