Joseph Owens
16
COW Coins
16
DaVinci Coins
Forum Replies Created
-
Trying to cross colorspaces (especially with a relatively consumer-level setup) is always going to be a simulation, as Tero alludes. Disregard the DCI-P3 colorspace for your grade. WHAAAAAAAAT!?!??!??!
Choose a Universal colorspace (*A.C.E.S*) and choose your Input/Output Transforms for your system. Grade in cct. You will be closer with your final DCI-P3 output (agree there may be tweaks if you do finally get to see it in a true P3 Display Reference) – but a monitored 709-mediated session will get you far closer with a very accurately calibrated judgement screen than a poor approximation with an economy setup/emulation mode. This is what ACES is designed to do.
Because the brochure says a display “covers” a certain gamut doesn’t mean it is doing it accurately. Remember a “fair ball” can be anywhere between first and third.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
Video Levels vs. RGB. Read all about it. Range issue.
Gamma is not involved.
Look at your signal measurements. If all that is reported is “gee it looks darker”, then you haven’t looked at your levels.
ProRes 4:4:4 can contain either RGB or Y’CbCr coded media, and it will make a sometimes arbitrary decision (wrong almost 100% of the time) about your video range. If you are converting from a Y’CbCr origination, it is pretty much certain you will get a range error when attempting to render out to a 4:4:4 container.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
[Marc Wielage] “what is frustrating is that you also can’t render them. You have to either turn it into a compound clip (which is a real file) or place one “real” file format on the timeline before you can render a mostly-generator timeline.”
Of course, the truly resourceful find a way around that, but it was annoying to have to work it out. I like sending bars and a sawtooth through the various re-encoders (DCP, etc.) to see what the process is doing to my previously-beautiful work.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
The filetype “Uncompressed” can open a can of worms.
It may be advertised as an Apple Quicktime .mov but at least in my own experience there can be misinterpretations in how the media is unpacked. How was the media sourced?
Can you open it in a third-party player that exhibits the correct behaviour… that is, if it is indeed 25.000 fps and 9 seconds long, that the frame count is around 225?
Your Start/End TC is telling us 11 frames 10:18 to 11:04 — and 32 MB seems pretty low for around 10 seconds of 2K. It should be more like 1.8 GB.
If VLC or Switch or something like it can open the file (if it can be opened), curious about re-coding it.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
[Michael Gissing] “The DCP spec calls for ext3 formatted drives (Linux format) but in practice most cinema management systems run on Windows.”
That is correct, and it is all the more frustrating that so many exhibitions are running with ad hoc installations that are incompatible on a variety of levels with the DCI specification. Even when you do comply, the chances are that the submission will be rejected for one reason or another, the vast majority of the time because the exhibitor is out of compliance.
NTFS will work… if you are a macOS-based service, it may well be the best option, even if you run one of the LInux overlays (ie. Paragon). I just sent out a package a few days ago, and was almost immediately requested to do an upload. Sure, 64GB, here you go…
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
[Jeremy Hansen] ” use the color match (an X-rite color checker passport”
I’d be curious to see your X-Rite before-and-after. It sort of looks like the highlights in the background are correct, and the *clip* foreground has been pumped up over the top, forcing a cyan overshoot.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
I don’t really have anything against Resolve Color Management, other than I am suspicious of attempts to deploy proprietary approaches.
Also before I was ready to embrace ACES, when it was in the Release 0.n versions, I took the trouble to reverse- engineer linear transforms of my own, with the help of various cinematographers, charts, many cameras and a lot of patience and scope reading.
For log projects, ACES is an automatic, pretty much default choice for me now. If something isn’t working, something is wrong — and I will choose the shortest, simplest strategy every time. Choose the I/O transforms, apply them, go to work. I do not support LUT displays for judgement or conversion. On one hand, too crude, on the other, no control.
In the past, I have developed tables for the User Interface — so that clients looking over my shoulder do not question the pass-through values that may appear differently on the Resolve edit/grade preview pane. It is absolutely why it is not the best, first choice for making grade judgements when they have to count. I also only use Client-3 Calman profiles developed with a sensitive probe for my iMac and do not use Apple characterizations. For so many reasons. The main system graphic display for the User Interface was almost close-enough after some time spent with CalMan just creating an Advanced User RGB balance in the monitor’s own adjustment software. Actually close enough to not bother with a display LUT. It’s better when there are fewer interventions.
jPO, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
Sounds like another tower if you need all of those distribution methods, at least from one device.
Be aware that not all fanout devices can support any and all peripherals — it’s a power supply thing. For example, afull complement of Element control surfaces might overwhelm your average USB hub, and definitely if you also added a mouse/keyboard, external drive, maybe your Resolve dongle… expect brown-outs.
A couple of years ago, there was an “industrial” USB box offered by Startech — no power supply. The user was tasked to calculate the overall power draw and wire in an adequate wall adapter. the usual suspects can supply about 2.5 watts at nominal 5.5 volts. If you are not aware of the relationship V*A=W, you will have to trust the brochures.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
Lower the bit rate.
Advise that “dial-up” is dead.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.
-
[Peter Robertson] “I have 40 Gigs of memory.”
Is far more motherboard memory than Resolve15 will ever use — it is GPU/VRAM that is critical. That your 8K “slows down” is a red flag. Not knowing much more about your platform, you should look into the recommended hardware config that Black Magic publishes — its on their website in the Support column.
jPo, CSI
\”I always pass on free advice — its never of any use to me\” Oscar Wilde.