Date
1 - 15 of 15
Vectorscopes of various cameras
|
Geoff Boyle
These are from matching frames from the July tests.
In ACES in Resolve with just the relevant IDT applied, contrast set to .85 and colour centred with offset. Alexa SXT, C300-2, C700, F55, F65, Varicam
|
|
|
|
Argyris_Theos_cml
Something does not make sense: Red is slightly off in all cameras? Cyan is not able to get to a saturation point in the Sonys? ? Argyris Theos, gsc DoP, Athens Greece, +306944725315 Skype Argyris.Theos via iPhone
|
|
|
|
Nick Shaw
On 11 Dec 2017, at 07:46, Geoff Boyle <geoff.cml@...> wrote: In ACES in Resolve with just the relevant IDT applied, contrast set to .85 and colour centred with offset. I have always thought that there was an issue with the S-Log3 IDT in Resolve (and reported it to the Resolve team) which results in the skew you see on the F55 and F65 vectorscope images. I have converted the official Sony S-Gamut3.CIne/S-Log3 IDT into DCTL (see here on my website) and using this does not show the same skew. Geoff, I will send you my DCTL off list so you can compare. Nick Shaw Workflow Consultant Antler Post Suite 87 30 Red Lion Street Richmond Surrey TW9 1RB UK +44 (0)20 8892 3377 +44 (0)7778 217 555
|
|
|
|
Art Adams <art.cml.only@...>
Without seeing parade RGB or what the chart looked like it's hard to know if the cameras are perfectly white balanced. The Alexa red is a little on the cool side, which is what I would expect, but the Canons and Sonys are as well. Canon reds often lean toward yellow/green, so that seems odd. Sony reds are neutral to warm. My guess is the light source was a little cooler than the cameras were expecting. Based on the linear smeariness around blue in all cameras I'm guessing this was a roughly tungsten light source.
Also, modern cameras don't put the colors near the boxes anyway as that's way too saturated for modern tastes. Sonys will in Rec 709 mode, but nobody uses that. (I hope nobody uses that.) The F55 and F65 look very wrong. There's no way that much smear in blue would result in acceptable images. I've never seen that before. It's got to be an error somewhere.
|
|
|
|
Argyris_Theos_cml
Did I speak of white balance? Colorists usually check white balance on the vectorscope, but they do this evaluating the center of the display. It appears that Geoff has properly balanced all images. To the best of my knowledge the polygon points are giving information on color saturation and on the vector angle. Best Argyris Theos, gsc DoP, Athens Greece, +306944725315 Skype Argyris.Theos via iPhone
|
|
|
|
Nick Shaw
On 11 Dec 2017, at 15:32, Art Adams <art.cml.only@...> wrote: The F55 and F65 look very wrong. There's no way that much smear in blue would result in acceptable images. I've never seen that before. It's got to be an error somewhere. That smearing is the artefact I was referring to from the issue with the S-Log3 IDT in Resolve. It isn't always obvious on images, but on a vectorscope of a Chroma du Monde it sticks out like a sore thumb. I am not running the latest version of Resolve, and I gather they have changed some things there. Geoff, what Resolve version are your grabs from? Nick Shaw Workflow Consultant Antler Post Suite 87 30 Red Lion Street Richmond Surrey TW9 1RB UK +44 (0)20 8892 3377 +44 (0)7778 217 555
|
|
|
|
Art Adams <art.cml.only@...>
No, I did.
That's a sloppy way to do it. As most of us should be able to see, the dot in the center of the image is not at all a clean, perfectly round dot. (This is not unusual.) Given how distorted each dot is, I'm not sure how you'd know each is centered. Most are ovals. A couple are clearly not centered at all. This doesn't mean that the image isn't white balanced, just that it's impossible to tell on a vector scope alone. Q: How can one determine whether a distorted soft-edged oval is centered on a target? A: One can't. Parade RGB is the most accurate way of checking white balance. Black balance can be verified by looking at a scope, but that's because black is a small, clean white dot on a vector scope, not a sloppy fuzzy distorted oval. I'm not saying Geoff didn't white balance. I am saying that I can't explain, from this alone, why these colors are not appearing the way I'd expect them to appear, and white balance could be a culprit, especially as cameras that I expect to make reds cool are making them too cool, and cameras that reproduce warm reds are rendering them as perfectly red or blueish. The super saturated hues are smeared like crazy, which is not right. They should be outside of the targets, as the Chroma Match chart contains some very saturated hues that are meant for use with P3 (I *think* they are, anyway—they may just be super saturated examples that don't align to anything, I don't remember) but not smeared like that.
It's not clear to me at all. I'm not saying that he didn't, only that it's impossible to tell from scopes alone. The JPEGs posted do appear to match very well, and appear neutral, but there's something going on with red hue angles that's not right, and the cool colors seem distorted as well—not in hue, but saturation. It's what happens if you take matrixing to an extreme, although I've not seen it that extreme before.
|
|
|
|
Argyris_Theos_cml
In that case and with all due respect I believe there was no reason to quote my text. Argyris Theos, gsc DoP, Athens Greece, +306944725315 Skype Argyris.Theos via iPhone
|
|
|
|
Luis Gomes <gomes.luis@...>
we don't! check white balance on a vector-scope!
Luis @finland having the audacity of calling himself a colourist more like a QC expert
|
|
|
|
Tom Tcimpidis
I refuse to get drawn in to yet another senseless battle of semantics and procedures but will simply say that long-time, experienced VCs have been using the vectorscope to white balance for eons… They know exactly how to interpret the fuzziness and lack of roundness… It is obviously not the only tool for the task but can work very well by itself or in conjunction with others in the right hands…
Tom Very old L.A. VC and tech maven
_,_
|
|
|
|
Geoff Boyle
The cameras were all in preset as it says on the web pages for the tests, it also lists the lights, BB&S area $* so daylight and some of the most accurate lights I’ve used!
toggle quoted messageShow quoted text
I then balanced in Resolve using saturation only and entering up the vector scope. Contrast not saturation was at .85 Exposure was corrected to give mid grey at 50%, some of the cameras came in at tis without alteration. The Resolve Vectorscopes were X2, I couldn’t make this adjustment in the Prelight ones, I wish I could. Otherwise everything was bog standard.
|
|
|
|
Geoff Boyle
Correction
toggle quoted messageShow quoted text
I then balanced in Resolve using Offset only entering up the vectorscope. It’s in Prelight that I used Saturation. Whilst the centring of the vectorscope may not be perfect it’s as close as I could get.
|
|
|
|
Geoff Boyle
Now it gets really weird.
I’m trying to test the CTL that Nick Shaw sent me but I can’t because whatever I do I can’t get a non-IDT result from Resolve 14.1 F55 with Resolve Sony RAW IDT  Sony F55 with no IDT  I shouldn’t get anything like this without an IDT, I’ve checked that the system is working fine by removing IDT’s for Alexa, Canon & Varicam and they all work as expected.
|
|
|
|
Luis Gomes <gomes.luis@...>
Woa. I am talking about now versus 1982 when I did use the vector scope because I did not had access to a parade waveform monitor.
toggle quoted messageShow quoted text
Of course I keep an eye on all five calibration scopes. But to judge white balance on a vector scope is spot on only if you have a reference. Again with this group one has to be crystal clear on things. Where Mike Most? Luís. Senseless assignments since 1982 Finland.
On Tue, 12 Dec 2017 at 7.13 Geoff Boyle <geoff.cml@...> wrote:
-- Gomes.luis@... http://fi.linkedin.com/pub/luis-gomes/20/11b/335/ Freelancer video Professional. Finland.
|
|
|
|
Nick Shaw
On 12 Dec 2017, at 05:12, Geoff Boyle <geoff.cml@...> wrote: I’m trying to test the CTL that Nick Shaw sent me but I can’t because whatever I do I can’t get a non-IDT result from Resolve 14.1 I'm not running Resolve 14, so I can't test. I guess they've changed how things work. A raw file needs to be decoded to some colour space. In Resolve 12.5 with an F65 RAW clip it would default to S-Gamut3.Cine and S-Log3 (greyed out, so you couldn't change it) so I could apply my IDT. Given that the "no IDT" version in your new screen grabs looks far better than the one with the Sony RAW IDT, I suspect that Resolve is requesting ACES data directly from the Sony RAW SDK, so no IDT is required. I don't know exactly what the "Sony RAW" IDT does, but it certainly doesn't seem to doing anything good in this situation. What it "should" do is not actually apply a transform, but simply set a flag which says "ask for ACES from the SDK". But this seems like it may be happening anyway. Nick Shaw Workflow Consultant Antler Post Suite 87 30 Red Lion Street Richmond Surrey TW9 1RB UK +44 (0)20 8892 3377 +44 (0)7778 217 555
|
|
|
