[Joseph Owens]

If you search the forum, there are a few threads about programmable game controllers, and of course there is a Resolve layout available for the “Keypad” iPad app.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

[William Edwards] “Should I be outputting for people legal broadcast outputs when they are going to bring that into Final Cut, “

FCP is a bit dumb about that — it is expecting Y’CbCr for most formats, and in Rec709 that would be the linear legal video setting. Full scale is for RGB formats — dpx and so on, which FCP has trouble with. ProRes 4444 confuses the issue because it can contain either RGB or Y’CbCr and there is no flag that would clue in the application as to scaling.

AVID does allow for deliberately scaling values on input.

These errors are exceptionally obvious for online operators looking at outboard scopes, but that would be extremely rare for passing around clips for VFX, etc. You might get a comment back like “seems a bit dark”… or conversely “washed out”… if you hear that…. panic.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

Legal broadcast for luma/chroma is a bit of a can of worms when you are working in an RGB environment.

Tolerance is 2 IRE around 0/100 for luminance, ie -2 to 102 is the gamut limit for an iron-clad deliverable, and 20 IRE is the combined over and under for luma plus chroma ( that is -20 to +120) when you are measuring it in composite baseband. Which you probably can’t do if you are depending on the Resolve scopes, since they don’t even display luminance, let alone simulated composite.

Basically all you can do with RGB scopes is a wild guess, so building precision LUTs is a waste of time if not a deception. The final matrixed composite value cannot be estimated looking at an RGB parade. You can, sort of, set up an approximate envelop using the CbCr vectroscope representation… by not letting the chroma go outside the 75% saturation dots — but — this approach doesn’t even qualify as crude, and a colorist will never be able to push a look with this kind of blinkers applied. Where the parade evaluation method really falls down is in the blacks. Saturated reds and blues near 0 IRE (police car markers are a good example) routinely are outside broadcast legal and are otherwise difficult to detect, not like bright yellows, particularly the values that graphics artists almost invariably choose. (100% R plus 100% G with 0% B would be the one that would guarantee trouble, even though it looks legal on the parade.)

The only way to accurately evaluate broadcast legal is with a display/monitor/scope system that actually does the math. Gamma-adjusted Y’CbCr is not a linear transform. You only have to mistakenly confuse the full-scale / linear video settings once to find that out.

Tektronix Diamond and Videotek Iris displays are good examples, and they can even be programmed to flag the timecodes and overages. In the absence of a dedicated hardware legalizer, that is making adjustments based on composite baseband, an operator’s only option is to fix the OOGs (out-of-gamut) on the spot with grade qualifiers. And to massage those values until the scope stops squawking.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

“Sony” covers a lot of territory, so not sure what you might be referring to.

There is a certain amount of ambiguity among some codecs, so be careful with your assumptions. ProRes 4444, for example can be either RGB or Y’CbCr, and isn’t flagged in any way to delineate. Anyone involved in a multi-way conversation between FCP, Avid, Resolve, and VFX applications will find out quick that ‘full-range’ and ‘legal’ are interpreted in different ways.

Simply, is the requirement for RGB 0-1023 or Y’CbCr 709 64-940? That is the difference.

As far as ‘deliverables’ go… I have a binder full of them. American voters might get the jibe.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

Typically, long-GOP formats like mpg and H264 seem to cause a lot of audio sync problems, so one strategy you should consider is a meet-in-the-middle intra codec like one of the ProRes flavours. While you’re at it, the GoPro source media could be massaged so that it plays nice with the rest of the team. A conversion to 23.98 might be possible if it was consistently shot with a 23.98 camera rate recorded with 2:3 field cadence over the 60i frame sequence. If not, and the footage really is either 5994i or 60P (but not 24 with Advanced Pulldown), you will be into a more-or-less frame blended or optical-flow re-master. Depending on the quality you want to achieve, this can take either a fair amount of time, or look soft and gooey.

Three different cameras in three different formats? How is any of this going to match? I knew the money had gone out of music videos, but this kind of shocks even me.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

[Christian Stoehr] “Thanks, what are your system specs, I am just running A 2008 Mac with a GTX570 and GTX120,”

Its a late MacPro3,1 with Quadro4000 UI and 3x flashed GTX480’s in a Desktop Xpander. I access source media through a number of interfaces anywhere from FW800 Rugged LaCie drives and a 1.5GBs fibre-optic RAID. Its configured to Blackmagic specs otherwise, running under OS10.6.8, so its not completely orthodox.

Its always safer to divide and conquer, and the reel-by-reel approach does take away a lot of the potential for adventure. It makes a ton of sense if you are prepping a project for film-out; 20 minutes is a standard interval for that, but unless its for archiving purposes, DCP seems to be taking over, for better or worse, and will want the whole project as was implied in one of the previous posts.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

Version 8 would suffer the dreaded Out of Memory warning if the Media Pool got too big, I haven’t seen it in V9 and just finished a 72 minute feature all-in-one-timeline. Burning the BluRays as this is written.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

Among telecine colorists, correcting for colour temperature between stocks and filtration was one of the most basic things asked to do. No knobs were ever provided for 3200 <-> 6500, generally because the dye layers are too non-linear in response across the exposure range for a single control.

In your case, you will have to compensate gain for diminishing exposure, which will affect overall dynamic contrast, and you will also have to start biasing toward blue/cyan, in whatever way your application accomplishes that in its processing matrix. As golden hour progresses, the amount of correction will also vary. Depending on whether your digital camera was matching peak whites or not, you may find that your final compensating correction is more pronounced in the “mids” or gamma range than for an overall gain change.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

Revising an AAF-imported timeline isn’t a very robust workflow yet, as it will lead to other limitations on the return trip.

This will sound like a klugey workaround, but I’m going to suggest making the VFX revision in your AVID editing application and exporting a new AAF. Import that new AAF into the same Resolve project as your original, and the new timeline should open and automatically link to your already established grades. That would be the beauty of the Master Timeline.

jPo

“I always pass on free advice — its never of any use to me” Oscar Wilde.

[Joseph Owens]

DPI is irrelevant.

The absolute square pixel count is, and 720×480 is correct for 4×3 SD NTSC DVD. Create 853×480 for 16×9.
PAL is different — 720×576 for 4:3 and 1024×576 for 16:9

jpO

“I always pass on free advice — its never of any use to me” Oscar Wilde.