"Video Level Scopes" not working properly

When "Video Level Scopes" is enabled, the waveform is scaled to legal as expected, but sub-black and super-white values are not shown, even when they are present and visible on an external scope if "Retain sub-black and super-white data" is enable for the SDI output.

Additionally, I like the new options for the waveform scale, but if "Video Level Scopes" is selected together with "Percentage" I would expect the 0-100% range on the graticule to match the legally scaled signal. It currently doesn't.

But overall I am very impressed with all the new additions in the 17 beta. Keep up the great work!

but if "Video Level Scopes" is selected together with "Percentage" I would expect the 0-100% range on the graticule to match the legally scaled signal. It currently doesn't.

Same with mV

Problem with superwhites clipped when the Waveform scale style is set to 'Video Level Scopes' is still there in 17.1 (Windows)
Hopefully just cosmetic, but annoying as you don't see at a glance if there's useful data to recover in the highlights of SDR FHD H264 clips and similar

[important?] update:

When my UltraStudio Mini Monitor is connected, Video levels are OK
When the I/O device is not present, video levels are cropped as described in the OP,...

no IO.png (605.48 KiB) Viewed 1556 times

(above: no I/O devive)
(below: with UltraStudio Mini Monitor connected)

...and this is worse, and suggests it is not just a cosmetic issue [with the waveform monitor] but actually a different interpretation/handling of the levels:
When playing with the contrast in Color (Use S-curve is enabled in the Project settings), the gentle roll-off occurs just below 1023 (hence, in the superwhites area, as expected for this true video-levels clip containing superwhites) when the UltraStudio Mini Monitor I/O is connected, but below 940 when the device is NOT connected: this is not expected, and wouldn't occur if the waveform scope was just clipping the values above 940. Instead it suggests that the same video is handled differently depending on the presence of the I/O device !
Yesterday, I made some basic grading with my laptop without I/O device, and today, while working with the same laptop and the UltraStudio connected, I realize the grades are totally wrong (washed out, lack of contrast, both visually and on the scopes) !!!

Unless there's something fundamental I don't understand with Resolve (I'm quite new to it), this sounds like a bug to me (why would my grades change depending on the I/O device being connected ????)

PS: interesting also to see the ripples mainly in the Red histogram curve, for the "no I/O device connected" case)...why, since *nothing* is different between the two screenshots, except the presence of the I/O device ???

Oh, and this also means I have to revert to 16.2 on my secondary laptop, where Mini Monitor isn't available....cumbersome...

For me levels are clipped no matter what, and this is in DaVinci YRGB colour science too.

I filed a support request for this, but no answer/ack after almost 2 weeks....

I'm not the only one on earth having trouble doing simple tonal range adjustments with video-levels clips when no I/O hw is connected in 17.X, am I ?
How do folks working with "old fashioned" video-level Rec.709 clips work around this ? Even if your output isn't supposed to go beyond the broadcast limit, not being able to "see" the superwhites when doing basic grading isn't very convenient ????

same problem with 17.2, support has ack'ed my request, but not confirmed the issue yet...

BMD Support has responded they could repro the issue. Let's hope this will result in a fix in a not-so-distant future

Fixed in 17.2.2

I'm having a problem seeing anything outside of the set reference levels when using the 'Video Levels' option on scopes. The parade just cuts off at the reference lines. Am I missing something or is this a recurring bug? Thanks.

I'm having a problem seeing anything outside of the set reference levels when using the 'Video Levels' option on scopes. The parade just cuts off at the reference lines. Am I missing something or is this a recurring bug? Thanks.

It happens in RCM projects for some reason, no doubt ACES too