Thanks Alexander. I put a video on vimeo with the material.

http://www.vimeo.com/986544

Sami

I know this is not the answer you want, but with a new camera that we all are getting to know, my suggestion is to try it all ways and the YOU tell us what you found best. For me I try to do as little post as possible but that is not necessarily right. Does the camera do post better than the NLE? Is the info there? I think we all are still getting to discover that and that is why there are few definitive answers in this whole long thread.

God luck and please do your test and tells us what you find.

...now let's work on a picture profile for the "Bourne Ultimatum" look! :)


...no really!

The difficulty besetting this subject is that one needs a sound understanding of the controls involved and their quantitative effect on the image. There are too many variables to just fiddle. Bill Raven really got things going with valuable contributions in determining (quantitatively) desirable settings, and I was pleased to find Simon Wyndham's SAW results that actually did define the nature of the gamma curves. I've seen only the SAW results published by Sebastien Thomas, which I understand were recorded at 3dB gain which resulted in flats in the region of 80 to 90. Bill has looked at this and mentioned that he has determined that the full curve (without flat) is seen only at 3200K 0db (white balanced). Has anyone output those curves for our reference? I unsuccessfully searched for the methodology for doing this myself, so maybe I just need a little primer or a link. I've been into the service menu but that's one area where I get nervous when I'm not absolutely certain what I'm doing.

Here are the steps to follow when evaluating SAW curves with the EX1:
1-Select any SD mode with the EX1.
2-connect the firewire link to your computer
3-Fire up HDRack
4-enter the maintenance menu and pick SAW
5-the WFM window in HDRack will display the gamma curve. note the WFM scales.

I assume it would be possible to record(capture) with Vegas and look at the WFM after the fact.

I, also, assume, it's possible to look at the SAW curve via the SDI or component out ports of the EX1 feeding a conventional WFM and vectorscope, altho' I've never done this.

Thanks Bill.

afaik.....

sony has always made cameras for the masses....like the metaphor for a Chevy Nova. Sony finally stepped out of their box and made a Cinealta available for the masses...albeit a wee bit expensive. OK, so, their marketing gurus told them most people would never deal well with the myriad of controls. And, I suppose, in a way the marketting folks were right. Nevertheless, I'm very happy with what i bought. OK, vignetting, backfocus, labels that wear off, etc., etc., I got a pretty decent camera for the $$$ I spent. I'm not exactly ready to believe Sony the next time around, but, who knows?

...and how does this matter in the grander scheme of things? well, not worth a shyte.

Vegas does produce the plots of gamma curves. So no need to go through OnLOcation (DVRack).

Except that with HDRack you get real time viewing of the curve, thereby seeing the immediate effect of changes to the camera settings. It's inconvenient, if not totally unworkable(at least, for myself), to make changes, then export the recorded result to Vegas to see the result of those changes. Using vegas is a good workaround, I suppose, if you can trust those somewhat untrustworthy Vegas scopes.

Agree. I was just wanting to investigate gamma curves, so that lends itself to going offline. Hopeless for the variables you were looking at.

I just finished all 20 pages of this thread. Many thanks to all the contributors.

If you wanted to use Std3 gamma, is there a reason why this would not work?

1.) Choose a high contrast outdoor scene.
2.) Expose on a neutral gray card, centering the graph on the x-axis of the histogram.
3.) Remove the card and compose the scene.
4.) Turn on Zebra 2 (100%) , adjust knee downward until zebra stripes go away.
5.) Adjust black so that it's not crushed on the histogram.

It needs to be an 18% reflectance gray card. Kodak makes them, others may too. Available at any pro photo store. An 18% card is made to fall in the center of the scale.

At a Nikon photo school I was taught that the palm of a human hand, regardless of race, is 36%. You can hold your palm up, meter and open up one stop. If your own palm is in front of the lens, your thumb will remind you to open up 1. Make sure the light level and angle matches you subject. In direct sunlight, no problem. In artificial light, the palm has to be in subject position.

In my tests this has always been correct, as improbably as it seems.

a grey card will work. S0, in fact, will any card whose reflectance in perfectly neutral....RGB255 to RGB0. I believe the convention is a pure white 255:255:255

Not if you are setting exposure. A 255:255:255/90% card set to the middle of the histogram will give a major exposure error. Great for white balance though.

An 18% card will, in essence, tell the camera the amount of light falling on the subject. It's an alternative to directly measuring the light with an incident meter. If everything is calibrated correctly, both will give the same exposure value.

I have done that for years, the palm trick + 1 f-stop with 35mm SLR, but I wouldn't have dared admit it here.

lots of people seem to use zebra set at 70% for skin tones. I've found 60% to give me a much better exposure. Anyone else notice this?

Same here, at least if I use STD4 or Cine4 I will get too bright picture when zebra is set at 70%...


Dennis

Zebra level
 

Hi Bill,
first of all - thanks to all your investigations concerning the EX1 profiles. Great work and even more important - great results. I am very frequently using your profile with Cine 4 ...

Concerning Zebra level:
I am actually using 90% zebra for the low level. The reason is, that I use Zebra only as a white clipping indicator. So seeing zebra marks would warn me, that I am almost over exposing. The setting of Zebra very much depends for what indication exactly you are using it. As "the last warning before clipping white", I prefer to set zebra low just very little below actual clipping and Zebra high just above clipping level. So when I see Zebra low, I can react - if I would ever see Zebra high... it is too late ...:-(
But Zebra high setting makes sense, as sometimes I even want pictures to clip in some very selected highlight peaks...

My goal is, to make maximum use of the camera's dynamic range seeing as much detail in white and black parts of the picture. Of coarse there are scenes, which have much higher dynamic than any camera can handle - that's why we compress parts of the shades not to loose details by crashing blacks or whites (or worst case both...)
It is up to the DoP to decide the best compromise and target the film/video to the right expression using all technical possibilities.

That's why I appreciate your work Bill - as it targets to squeeze maximum usable dynamic range out of the EX1 with natural colours...

Recently I found, that very often auto Iris works very well to keep the image well exposed. Of coarse this is only useful in shots where depth of field is not critical. At least my tests confirmed, that in auto IRIS mode using Bill's Cine4 profile comes very close to optimum exposure (no clipping of white).
As long as I minimize clipping of white without crushing blacks, I can do everything else in post production...

regards,

Ulli

Ulli...

Indeed, using zebra for highlight exposure control is what zebra is designed for. I think the conventional wisdom is to use zebra for monitoring highlights, either on white objects, as you describe, or by monitoring highlights on skin. On interior shots, such as a stage production, where there are no true white objects to meter, what is important is to not blow out the specular reflections on human faces. This is where a 60% (or 70%)zebra setting is invaluable.

Can you actually trust HDrack with the EX1? I mean the only way to see the scopes is to go in with firewire right? Will it give an accurate representation of the signal? Or HDrack works now with HD-SDI too?

You might try Veescope Live. http://www.dvdxdv.com/ I have tested the free demo and it seems to work as advertised.

If you have a Kona or similar capture device it will display whatever your device will capture from HD-SDI to composite plus Firewire and matches my hardware scopes.

Monitor and scope display has about a one second delay.

Did you test it to be accurate with the EX1 also when using firewire or just SDI?

HD-SDI, component and Firewire. All matched within readability of the screens. The screen is less sharp on Veescope.

I'm only interested in the scopes function as I can't always have hardware scopes out on the field. But if using firewire to connect to the laptop, I would think you would have to switch the EX1 to SQ ("HDV") mode in order to work with the firewire? In that case, if what you are shooting is HQ, will checking the scopes in SQ still give an accurate representation of your signal? The question goes for both Veescope Live and Onlocation by the way.
The problem with Veescope Live is that it is Mac only and my Laptop is a windows machine, so Onlocation may be the only option. I may have to buy Premiere just to get Onlocation. Or is there another option for the PC?
Just making sure before I buy anything. Thanks.

Michael, I also would love to hear feedback on your question as well...

FWIW...

HDRACK(which I assume is the same for Adobe's version), distinguishes REC601 from Rec709. Unlike Vegas, which never tells you WHAT it's reading, HDRACK will annunciate the color space it's reading. I would assume there's no difference in color space between SQ and HQ modes with the EX1. Simply a matter of higher bitrate doesn't mean a change to the color mapping.

Whether you like the results, or not, the color profile I posted here was done with HDRACK and the firewire connection in SQ mode. Applied to HQ modes yields the same results. Fairly convincing, altho' perhaps not conclusive, evidence that it's accurate enough.

This is somewhat good news then. So switch to SQ, check on DV Rack, switch back to HQ and record. Sounds like a worthy workaround to me.

What version of DV Rack did you use Bill, DV RACK HD 2.0? I hear CS3 has some improvements.

If only DV Rack was still available as a stand alone option.
Is there another option for the PC?
Thanks.

Bill,
Allow me to expand the question out of the EX1 slightly...

if using scopes with a downconverted SD signal for color checks, is there enough of a color space difference to warrant this as not a good idea if recording in an HD mode whether HDV or HQ, or would it be fairly accurate? (applying this question to the broader range of HD cams - like the XHA1 for example in addition to the EX1...)

Thanks,
Lonnie

I have only tested with the free demo of Veescope.

Within the readability of the scope image on my Mac/Dell 30 - HD-SDI, firewire, component HD and composite SD all seemed to match the hardware scopes. Veescope seemed to adapt to the changing color space. All of the computer screen scopes I have seen or used make an image that is less sharp than hardware scopes. Comparing requires estimating whether the "center" of a small fuzzy blob is in the same spot as a sharp dot. If you really need to know within 1 IRE, buy hardware scopes.

After a brief scope test my 3 week old EX1 failed. The IS system went into uncontrolled vertical cycling. Sony tried a repair that did not work and has it back to try again. No word on when it might come back. I had the camera 3 weeks, Sony has had it for 6 weeks. So I can't recheck anything.

Answering Lonnie's question, the EX1 appears to correctly translate from 709 to 601 for an SD monitor feed. If the SD feed was correct on SD scopes, the HD signal was correct on HD scopes, until the camera blew up.

Sorry, I only got a couple of hours of scope testing and color calibration before the camera fell apart.

Real Color Picture Profile
 

Here is a new PP matrix settings for you to try.

I have tried as much as possible for max saturation while centering the vector angles.

This was produced with the same technique as Bill used for his TC PPs - DSC Labs chart, color correct lighting, HD-SDI to hard and soft scopes, but I got a different set of numbers. These profiles look more balanced on my camera, they may look less so on yours.

I have called these Real Color 1 and Real Color 2. RC1 is very close to total saturation, with just a tiny bit of head room. RC2 is right on the mark for 50% bars, no head room at all. If you find you are sometimes clipping colors even with RC1, try setting Matrix Level to -10.

If you give these a try, let me know how they work for you.

Matrix ...............on
Select................hisat
Level..................RC1 0 / RC2 +10
Phase.................3
R-G...................72
R-B...................-5
G-R...................-1
G-B...................23
B-R...................13
B-G...................-29

I have attached a jpg of RC2 on FCP scopes. FCP live scopes show the HD signal as SD colors pace, but this also looks correct on other hard and soft scopes.

The yellow circle marks 63%, the correct max for Red and Green on 75% scope and 50% bars, when viewed as NTSC color space.

I would have posted these a month and a half ago, but my camera failed while I was creating these profiles and I just got it back from Sony. I needed at least a brief chance to test them in the real world before posting.

Hey George - glad to hear your camera's back! Any idea what it was?

George.
Which gamma did you use with these settings ?
Thanks.
Paul.

And Black and Black Gamma?

Carroll Lam

Real Color Matrix Profile - This is only a Matrix profile for balanced saturated color. Set everything else where you wish for the scene you are shooting.

Gamma: Std 1-4 and Cine 1-4 all maintain the vector axis and gain is barely changed after the exposure is corrected for the gamma change.

Black gamma also has minimum or no change, after setting exposure. Same with knee.

Low Key Sat WILL change the vectors (it IS a saturation control) but mostly above or below + or - 10. Using + or - 99 will make some very weird colors.

I have attached a jpg of the Sony standard matrix for comparison. Notice that it is very low saturation and the yellow, green and blue vectors do not align with their vector targets. Just like pictures taken with the standard profile look on a calibrated production monitor.

Ronn - This time they put on an actual NEW lens. I've only tested it for a couple of hours, but everything seems to be OK.

btw - let me know how the RC matrix settings look on your camera.

George,

Here is a sample shot with that RC1 profile you posted.

It was shot through the not-so-clean window of my office which overlooks the lounge area outside the building.

George's RC
 

Hey George - here's a 30 second clip I shot yesterday of my fountain, using a variant of your RC settings. I actually split the difference between the two (Level was 5 vs 0 or 10) and detail off. 720P24 with 1/48 shutter and polarizer.

http://www.box.net/shared/dlka9t48os

Daylight
 

So what is the best recipe for shooting in heavy sunlight?

I found the TC2 profile excellent yesterday in mid-afternoon sun, but fairly unusable indoors without a lot of light. The RC1 profile does better in low light and really looks nice.

Perrone - Thanks for the still. Too bad there wasn't more color range in you subject. Let us see what some full color range subjects look like with RC1 on your camera.

Ronn - Great clip. Nice range of colors in the subject and moving in to frame to show skin tone is a plus. Very saturated without any strange colors that I saw. Should look great from 0 to -15. I had not thought about it before, but I will try a Tru-Pol on the charts to see if that upsets the balance or levels. Polarizers might want a profile of their own.

RC1 is about the max saturation I would be brave enough to use on a real shoot, and then for clients that insist on no post grading. It provides a bit of head space on saturation, but exposure or white balance errors could still give chroma clipping.

RC2 is to show where the top is, DSC 50% color chart taken all of the way to 50% on scopes. Pretty much no chroma head space left.