[Joseph Owens]

[Toby Tomkins] “Would it be better to replace the 580s completely or put the 580s in a Cubix and use all 4? What are the performance implications of mixing cards, do the faster cards have less utilisation once the slower cards max out?”

Probably replace them if you can afford to, because the Cubix pipeline runs at the speed of the slowest card, so otherwise you are throwing away the Titan performance. Using the 580 as GUI plus render may not yield any appreciable difference once the Titans are doing the heavy lifting.

jPo

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

[Joseph Owens]

[Jim Klatt] “let’s just say that there was a 3rd solution. One that allowed for lights to be turned down and windows to be closed when need be to make sure everything was accurate and where you wanted it to be.
I would certainly sacrifice an extra hour per day of doing a final check with the lights down if I knew that I could do my work in an environment where plants could thrive,”

Lovely fantasy, but a case of wanting your cake and eating it, too. Grading with the windows open and the blinds up would be the shortest distance to doing it all over again that I can think of, and probably not just a matter of a little tweak here and there. Your black point, gain and contrast will be — literally — out the window. Especially with current reflective solid-state monitors being the way they are. Its hard enough getting that right even under draconian environmental control.

My plants do very well (maybe even a little bit too well) in my administrative office just off the grade suite, where I retire every once in a while to catch up on correspondence, take a breather, let the machine take care of some business, see what kind of day it is, what the hot topics are for today on the COW… I have an east-facing floor-to-ceiling glass wall, plus skylight, so a tonne of natural daylight… but because it comes in through glass, obviously there is no Vitamin-D stimulating UVA or B. For that, I have to walk up the street, but luckily there is an Italian bakery about 5 minutes away, who make a killer latte for two bucks (way way way better than the Mermaid people), but if you want a cannoli, you better get there early.

Maybe what you are trying to achieve is akin to the Japanese concept of shibui, the studied elegance of apparent casualness, which is anything but, and this is the work environment that you are trying to establish. There are other means to avoid the techno-man-cave dungeon… but if I really wanted to be sipping little fruity drinks with umbrellas on a tropical beach, I wouldn’t be trying to do that in my work place. People get the wrong impression.

jPo

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

[Joseph Owens]

[William Edwards] “You know that look with the black and white footage except for one color,”

Forests have been felled (electronically speaking) discussing what is known as the “Pleasantville” effect. Argue the semantics of “look” vs. “effect” among yourselves… You can spend a few days reading all about it with a simple Google search (3,420,000 hits). Different grade and NLE systems have different approaches. Of course the big challenge after roughing in your secondary qualifier is cleaning up the garbage around it, and hoping that the block compression artifacts aren’t so big that all the edges chatter like St. Elmo’s. Then you are into a tight roto, which should be avoided. Roto in Resolve is cruel and unusual punishment, even after conceding that Roto, in general, is itself cruel and unusual. One word: MochaPro.

To improve qualification in Resolve, we sort of need a chroma-blur node like we used to have with one click in SHAKE, and for sure, the kind of hooks that attach to a Mocha-track and shape like they do in Nuke, FCP, AE, Shake, … etc. Inside/Outside blur/erode only goes so far, and you don’t have to look far afield, even in this forum, to find individuals who are not fans of the Resolve keyer anyway. If you have worked with Keylight or Primatte, or a variety of others enough, you begin to appreciate the difference. Yes, masking is more efficient and it is easier to build and track spline shapes in Version 10… parenting them or assigning track hierarchies… while sharing a correction… not so much. But then, you really are into an FX application, not a grading solution per se, and that is where Resolve is currently teetering.

jPo

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

[Joseph Owens]

Comment also posted on Fujiboard

Looks like a pretty standard MacBeth Greitag chart on the right, and what Fuji would advertise as equivalent to the Kodak 18% CinemaTools chart. “Chart” is the correct terminology.

All three manufacturers will have recommendations as to what values you should set these up for with what would become a sort of LUT or basemem for your session. Some google surfing should bring up the correct documentation. I don’t use them anymore… because…

We used to use these (or at least that was the theory) all the time to set up film transfers and to sort of calibrate a telecine to act as an extremely expensive sensitometer, because we had these charts with point/exposure correlations so we could relay the info back to the DP that film timers used to when rushing dailies.

For your purpose, I’d set up the brightest white square on the Macbeth to around 80IRE, the blackest area down to 0, of course, and then check the gamma linearity. Adjust seasonings to taste, as if they were, you know, salt and pepper and this was a recipe. When you have made the charts look nice, that’s the “secret sauce” for this session. Don’t expect it to apply to everything… or, really, anything because usually these charts are shot under circumstances that have nothing to do with the actual shooting environment. Every one of these values will change with exposure, lens choice (sometimes even from ‘full wide’ to ‘tele’ on a zoom), set lighting… you name it.

jPo

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

[Joseph Owens]

[Juan Salvo] “How do you create the look of shooting…”

Its a bit like how perfect everyone’s teeth are now… even when the movie is set in a time when oral care was non-existent and orthodontia wasn’t even for the rich. Article in this morning’s Globe&Mail had a picture of Raquel Welch in her fur bikini — with the caption, oh, yes, this is what women looked like one million BC.

What I mean by the example is that there is this fetish for recreating image finishes that are a product of a specific sequence of physical processes with their attendant optical physics and chemistry, dye layers, refractive indices, contact-printing… factors for which there really is no digital analog.

Look at the model (look) you are trying to emulate. What sets it apart from the digital source footage? What are you going to have to put in or take out? Is there a Sapphire plug-in for it? And why not?

jPo

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

[Joseph Owens]

Resolve is not clairvoyant, there are some specific requirements for the application to find and link to source media. If you were to review the manual regarding how to set up a round trip, you might find that Resolve, while it supports a vast number of media types, does not support all, and does need to see each file identified with a suffix — ie .mov. If you are able to navigate to a source directory and the application sees the media and you are able to import it, then the issue, and it sounds like this is what is happening — perhaps your source media is missing something else — a reel number and/or time code — or at least unambiguous or unique time code.

If connections are made, but they are not being linked to either the correct clip or the wrong trim points, then it is because Resolve is making a best guess where some or all of the correct references are not present. Another factor may be discrepancies in source frame rate. Although Resolve can be configured to take source frame rate into account, most offsets and frame mapping calculations are based on the project sequence frame rate, which will lead to errors if the source media frame rate does not match.

These are just some of the obvious first-up thoughts.

jPo

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

[Joseph Owens]

[Robert Houllahan] “I think Juan posted a workflow with a node tree “

ARe you maybe thinking about the node tree for synthesizing Technicolor 3-strip?

Might accomplish the same general look.

jPo

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

[Joseph Owens]

One thing to check out is the Quicktime Inspector, and make sure that what you think you are exporting is actually what format the file wound up being.

I often see surprise changes in files that are exported out of the FCP “Make Movie” dialog. The stuttering and looping you observe could be the result of a bad framerate selection. You might think you were exporting 23.98 or 29.97, but occasionally FCP/QT will create a bizarre framerate (ie 4.9fps) related to the space between your first trim-in point and an intra-frame in a Long-GOP format like mp4 or others.

Some people might classify this as a “bug”, but I’m sure Apple sees it as “your problem.”

Compressor doesn’t ever seem to exhibit this behaviour.

jPo

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

[Joseph Owens]

First off, happy Gregorian New Year to all…!

So, I’ve been having this conversation ever since the coffeecan MacPro started looking less like vapourware and started manifesting as a real thing. And as we all know it turns out to be a …. something else we have to work around. If you are starting fresh, no investment in peripherals or legacy third-party devices that run on busses that are no longer available… well, good, but otherwise, gee, too bad. “Shoulda waited….” Really? Seth and Amy could have some fun with that.

What about 4K and Thunderbolt? Well, to be blunt… what about it? Been working with higher resolution for some time now — yes, it took some investment to do it, but its working, and T-Bolt seems like a poor-man’s fibre, which… I’ve had in the shop ever since the beginning. The skype chat from this morning finished up along the lines of — can I charge any more for it if I’m running the same applications on a different platform? No. Is that a good business practice to splash out even more after-tax earnings on depreciates-right-off-the-cliff hardware?

Will Apple find a way to obsolete me? Yes. OS that kills legacy software and doesn’t run on the motherboard, that’s always a good one. But here’s a thing – in my market sector, one of the prime deliverables is a closed-captioned HD-tape master. Apple actually fenced themselves OUT of that sector by killing FCP7 in favour of a net-streaming edit app. Weird. Now I do realize that uploading air masters or striking XDCam disks is a trend… but so far: a) the “mezzanine” file hasn’t really been formalized (ProRes? DNxHD? H264?, how many audio tracks and channel assignment for full mix and stems? and so on) and b) XDCam? 8-bit, 1440 50MBits is whose idea of master quality?

jPo

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

[Joseph Owens]

[Margus Voll] “Maybe good Linux box will be also nice idea.”

Eventually that would help with DCI-workflow, but others have discovered one of the Murphy corollaries: “There are no perfect solutions, only ones that are more subtly wrong…”, in that LInux boxes have real trouble with the ProRes family, and that is one of the mainstay codecs of the present.

jPo

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