[Joseph Owens]

@Mike.

No cynicism involved, and extremely happy to see Alexis chime in on this. So few of us have living memory of composite baseband analog that we live in danger of reverting to a system of superstitions. “When you believe in things that you don’t understand….” thank you Stevie.

The In-Phase axis is a great reference, no doubt. Back in the day of setting up quad playback, where the matrixing varied from machine to machine and trim pots, sometimes the only reference we could really count on was placing the red “dot” on the 2% box, and everything else would just have to fit where it landed. Too bad. It was the closest to “fleshtone”, whatever that is, and resulted in the fewest client objections. Its actually too bad that the I/Q flags are 120 degrees from reference and that the -I burst is not actually set up to fall on its positive axis. but anyway…

As you say, In-phase and Quadrature are no longer relevant, the same as subcarrier and RS-170A, since the advent of CAV and then 601 and 259M. For those who have outboard scopes that show axes, you will also have to come to terms with the the fact that even when those guidelines are shown, they are not actually in 90 degree quadrature anymore, oweing to the difference in matrix calculations, especially between YIQ (no longer in use) and Y’CbCr, which is now ubiquitous. It does make it rather difficult for the manufacturers to be both mathematically rigorous and sympathetic to the artistic community at the same time.

As to the cyan/red and green/magenta bandwidth-limiting strategy, while it is true that the spatial frequency response of the human eye has that characteristic mentioned, from an engineering standpoint, the reality is that you can “borrow” bandwidth from how the channels are constructed. One upshot of how Y is composed, (mostly green) is that you can really knock it off in the chroma channel… because its already there waiting to be re-constituted when you make Y/I/Q back into R/G/B for display.

There are of course three channels in vestigial-sideband composite analog to cope with the historical 6MHz total bandwidth allotted to terrestrial broadcasting. Early videotape recorders needed to achieve a relative head-to-tape velocity of about 90 feet per second to achieve the response necessary. That’s just electromagnetic physics. Anyway, the three components, which is what you need to re-construct any color signal, whether its R/G/B, or H/S/L or whatever, can have different characteristics depending on the transmission. In Y/I/Q, Y is luminance, and saturation and hue are encoded in the phase and magnitude of the I/Q vectors. NTSC engineers were looking for ways to buy the best response with maximum efficiency. Luminance was pretty much mandated for backwards-compatibility with black & white — and interestingly enough, if you look at a bell-curve of human electromagnetic band response, (we see between, but not including, infrared and ultraviolet), right in the middle is green/yellow. So it should not come as a surprise that NTSC “Y” is a matrix of about 70% green, 20% red and 10% blue (numbers inexact for the mathematically-challenged). The immediate technical bonus is that your main signal can be nearly represented correctly with one channel and it gets the lion’s share of bandwidth. It is also true that humans don’t really need a lot of detail in the color component of vision since we mostly interpret the world in terms of contrast. So the I and Q components can be highly bandwidth limited, since we just don’t need it, and you can re-modulate them onto the top end of the luminance signal since the energy distribution will also allow that. NTSC (Color) is really an astonishing achievement for its time.

So two things: If the flesh-tone area was really THAT important, why wouldn’t it be made into a major axis (maybe adding some kind of dotted brackets for some kind of subjective variation)? and, having skipped over two or three years of sampling and communications theory, Fourier analysis, Nyquist theorems etc, can modern media practitioners survive without knowing and appreciating how some of the tools and methodologies evolved almost naturally as the children of necessity?

When I think about it, however, the Inphase axis and I-Bar (awareness and understanding) may also be a line of demarcation of another type… (we? us? old farts who know how things actually work and they? the young ones who don’t actually care?) Now that’s cynicism.

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

[Dan James] “anyway to fix this”

Its usually a function of overloading the media pool. You have to remember that Resolve is looking at the entire source clip for every event on the timeline, but only showing you the in/out on the cut sequence.

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

Yes, there is a “round trip for FCP” in the drop-down menu in the upper left corner of the Render page.

Then in Conform, click on the cut sequence and “Export” — it will create a new XML (unlike Apple COLOR) tiled MyMovie (Resolve).xml, which you then import into your FCP master project. No waiting…. well, a few moments and yerdun.

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

[Margus Voll] “Try selecting all your thumbnails on timeline and then make R clic with mouse and batch copy.
Now try to adjust one shot and third.”
?
I’m assuming this is an attempt to “ungroup” the shots…. but it doesn’t solve the underlying issue which is that Resolve is always grading to the Master source clip timeline… even when it is being displayed as a cut sequence. For absolute independence, it is better to create a “local” version of the grade for that portion of the master clip. This is probably better operational practice, as well, since it does build in a degree of redundancy and grade-construction flexibility. This is similar to separating your various operations inside a grade to separate nodes so that you can share them between clips and click portions of the grade off and on to gauge the effect.

Rendering to unique filename also helps to create entities that can be better managed with browser operations.

To reconnect files with different names, at least in Final Cut, one avenue is to “Locate”, instead of “Search”, and you do have the option to un-check the “Matched” filename when you do manage to navigate to the clip you want.

Getting closer?

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

A POV overview with BMD’s own Dan Moran:
https://www.youtube.com/watch?v=tYa6ddNt_tQ

…and its free.

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

[Brendan Maghran] “Do people usually finish with DPX sequences output from Resolve?”

If they’re “usually” finishing for an output that would require a 10 -(or greater)-bit sequential frame format. If you are delivering a web-stream QT or a DVD, this feels like extreme overkill, when ProRes QT would be far more than enough. And far more compact.

DPX-flow lends itself to film-out and DCP, although for the indie crowd, there are many other considerations.

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

[stig olsen] “When working with the NS button on, you will not be able to produce the black that you want to be scaled later on. It set a limiter before the scaling.”

I hear you. I used to argue against grading with a legalizer in the monitor chain, but in the end, if you are delivering to broadcast, it is the destination, not the trip.

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

[stig olsen] “A regular TV can show black levels under 16. If I understand things correct, everyone working with normal monitoring limit their work. “

The difference is Y’CbCr vs. RGB mapped to SMPTE259M. In broadcast 0-100 IRE, the active portion of the gamut is 64-940, which allows for over- and under-shoot. AKA your “superwhites”, where you are allowed to flirt with above-and-below 0-100 IRE… If you want to send 0-1023 as if it was 64-940, it will get re-mapped to those values, with resulting floating blacks and compressed whites. Its very simple math, but gets a little bit complicated as the gamma coefficient will also change the grey scale interpolation. In that case, a simple linear LUT won’t fix it, even when you try to “tweak”-eyeball the values back to their unscaled rendition.

Everyone working with “normal monitoring” is not limiting their work, they are adhering to international SMPTE/EBU standards. This does not apply to digital cinema workflows involving RGB containers, for example, TIFF, TGA, CIN, DPX…

https://www.tri-sysdesigns.com/Articles/BlackisBlack.html

All of this is extremely obvious with even superficial meter-reading skills…. that is, using scopes of the outboard variety that are not being tainted by a software application’s possibly erroneous assumptions.

Why this may be a problem for a lot of users is likely to be a training deficiency, as very few individuals are exposed to technical fundamentals anymore. Me, I pine for the days of Reed-Solomon, but that is all in the past. Back to the rocking chair…

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

OTT lights are pretty good out of the box.

jPo

You mean “Old Ben”? Ben Kenobi?

[Joseph Owens]

My takeaway from a thread on the 680 in the main Resolve forum is that 600-series is not compatible with Mac — and whether updating to Mountain Lion will get around that or not may be moot, as there was also an opinion expressed that the 600 series is not particularly well-suited for its CUDA characteristics, which are vital in Resolve. I underline vital. Resolve is not the same graphics engine as Adobe / FreddieMercury.

And the official configuration guide can’t really go into the use of flashed cards for various reasons.

jPo

You mean “Old Ben”? Ben Kenobi?