Question for David T. or N. on IRE levels f/ CF HDV material

[Stephen Armour]

If the the luma level adjust is not done BEFORE Cineforming material, the info in that "above 100 IRE" area is apparently lost? It apparently compresses it beyond any further usefulness.

David T. and/or David N., does CF toss out "above 100 IRE" luminance info? Since it does seem to do that in some quick testing we did, is there any way programatically for you to allow users to set those levels at capture/converson in HDLink, so we can have the option to do luma curves ourselves, instead of automatically saying goodbye to a marginal, but maybe important clip due to lack of user control in this particular area?

Since it seems to do this at conversion time, can you make it so we can chose a higher IRE level if desired, say for a clip with blown out sky, snow scenes or water? Our Sony V1's sometimes capture all the way to 115-120 IRE on some odds shots, but those few shots are totally lost unless we mask and replace the offending areas.

If the choice is between compressing "smaller or faster", as opposed to "including any and all luminance data", we would certainly choose the latter. For lots of our work, this is not a big deal, but as we do most of our productions with available light, it could be a nice way to save a throwaway shot.

[David Newman]

Your testing seems flawed as we don't throw away data above 100IRE. HDLink is a YUV tool so HDV data up to 110 IRE is preserved. Truncation at 100IRE is cg-RGB limitation, and HDLink doesn't use that.

[Tripp Woelfel]

David... In PP CS3 with Aspect in a Cineform project the waveform monitors show no information above 100 IRE. Now the likelihood of me having something set wrong is significantly higher than your chances of winning the Megabucks, but the same clip in an HDV project shows info up to 105 IRE.

If the info is there in the clip, does Aspect chose to ignore it or what?

[Stephen Armour]

Originally Posted by David Newman

Your testing seems flawed as we don't throw away data above 100IRE. HDLink is a YUV tool so HDV data up to 110 IRE is preserved. Truncation at 100IRE is cg-RGB limitation, and HDLink doesn't use that.

Ok, I'm like Tripp. No guarantee of extensive testing here, for sure.

Here's what we did. We captured the same ocean scene two ways. This particular contrasty image has a very bright 110+IRE area of ocean in the upper left of the image, where when captured as a HDV it shows that level (and higher).

When captured/converted to CF and bumped to 1920x1080p, the clip shows cutoff at 100 IRE and the detail we easily recovered (ocean waves and an island) with the regular HDV clip using the luma curve plugin for PP, was totally unavailable with the exact same clip that was Cf'ed.

Maybe our quick test is faulty, but please explain why. We're using the PP CS3's built-in YC waveform to check IRE levels. The reason we looked is because Tripp asked in a PP forum. That got us to checking before just assuming it couldn't happen.

[David Newman]

Through the Adobe SDK the scopes only show 100IRE, nothing above, but if you uses a 32-bit filter (Adobe's or CineForm's Color corrector work) you can bring your highlights down from above 100. We do it all the time, and with other camera that support far more than 110IRE. I'm away for the weekend so you have to get some else to show how to do it -- we demo'd this a lot at NAB so there might a video hosted somewhere.

[Stephen Armour]

Originally Posted by David Newman

Through the Adobe SDK the scopes only show 100IRE, nothing above, but if you uses a 32-bit filter (Adobe's or CineForm's Color corrector work) you can bring your highlights down from above 100. We do it all the time, and with other camera that support far more than 110IRE. I'm away for the weekend so you have to get some else to show how to do it -- we demo'd this a lot at NAB so there might a video hosted somewhere.

David, please re-read my last post. Nothing you just said had anything to do with what I was asking about that I can see. The question is NOT "how to bring down IRE levels in PP"...the question is:... "if CF is doing that automatically when compressing HDV video, resulting in a loss of detail in HDV video above 100 IRE."

Please re-read my post and explain what we're seeing.

Are you suggesting that what are we seeing using the Adobe "YC waveform" in the Program Monitor (with some HDV material showing up to 120 IRE) does not really reflect a "real luma signal" and is really only showing 100 IRE? I find that highly unlikely and very difficult to believe. Plus, it totally discards the fact that the CF'ed material is showing 100 IRE with apparently discarded detail, while the exact same video as regular HDV is showing up to 120 IRE, with luma info still available!

I'm trying to understand this and would appreciate a more comprehensive answer when you have time. If there is an issue with luma levels, we need to know about it and adjust our workflow accordingly.

As it stands, until we can hook up a hardware proc amp to confirm these levels (or negate them), or figure out what we're doing wrong, we're left thinking we need to distrust the software tools in PP CS3!

[David Newman]

I had a brief moment to test. All the data is in the file, as YUV to YUV transcodes do not truncate, I also confirmed that full range RGB decodes (up to 110 IRE) are working in Vegas, but at the moment the Premiere importer is truncating YUV sources (but fine for 444 and RAW.) So someting got switched off, will check when I'm back Tuesday.

[Stephen Armour]

Originally Posted by David Newman

I had a brief moment to test. All the data is in the file, as YUV to YUV transcodes do not truncate, I also confirmed that full range RGB decodes (up to 110 IRE) are working in Vegas, but at the moment the Premiere importer is truncating YUV sources (but fine for 444 and RAW.) So someting got switched off, will check when I'm back Tuesday.

We'll be unable to test further, as we are smack in the middle of packing up to move our whole team and operations 2,500 miles, up to Natal, RN, Brazil's bellybutton pointing at Africa. This just happened to be the last workstation still hooked up and with it's HDDs in it...

Hope it's easy to find and correct. We've never caught/seen that prob before, so like you said, it's probably just something switched off in the last build(s).

[Graham Hickling]

QUOTE: at the moment the Premiere importer is truncating YUV sources ...

Um ... so what's the present conclusion on this discussion - that the latest AspectHD version needs repair?

[David Newman]

Fixed in house for Prospect HD/4K, that will be out as a beta maybe tomorrow. The break happened when we added Active Metadata to YUV sources on Prospect 4K, and prepare for Prospect HD/4K v4.0 changes, so two steps forward and one back to be fixed in a few days.

[Graham Hickling]

I see this is now fixed in Prospect and Neo ... but not yet in Aspect if I am understanding these threads correctly?