Making multiple cameras match

M

MRFilm

Active member
Hi,

I've been involved in several multi-cam DVX100 shoots. For the feature Film No Regrets (aired on Lifetime on 12-27-04) and several other shoots for the same production company.

The other day they wanted to use two cameras so "we didn't have to wait to rig one on a hi-hat on a ladder for a high angle".

The shoot was on a white limbo set (white cyc) with lots of props and an on camera talent. The high angle shot didn't have as much white cyc in frame as the eye levels angle. These didn't shoot simultaneously, but lighting was essentially the same.

When the editor was loading in the mornings footage while we broke for lunch he found that:

The lower angle footage had clipped whites and crushed blacks, whereas the high angle footage didn't have any crisp-blacks (although the talent wore black slacks) and the whites didn't even reach 100ire on the scope. So not truely "white" for video. The shoots looked the same general exposure on their respective LCD finders. However they didn't look anything alike in edit. The settings were this companies usual:

24P (normal) all scene file settings were at zero, cine-like Gamma & Matrix, Thick V_Detail and Camera setting - 7.5% setup which post specifies, evne tho' the DVCam deck they use SHOULD impose that IF they're using composite video out. . .

Now they ALWAYS shoot/edit and then go to the local transfer suite and do a tape to tape color correction. The killer here was that although things looked fine while shooting there was lost detail in the talent's hair (in front) and her brown eyes. We abandonned the "offending" camera for the rest of the day.

The other killer is that the other shooter and I didn't feel that what we were seeing in edit was anything like we thought the footage would look like. one crushed/clipped and the other milky & washed out. Grnated with milky & washed out they can make it look anyway they want thru the daVinci. I even tried the "hard reset" on the "contrasty" camera to see if there was a hardware issue that could be reset and there was no difference after.

Also the color bars we layed down had no relation (except maybe chroma phase) to the video.

I understand that this is "just a $3000+ camera" and that you have limited ability to make them match, but shouldn't the basic contrast that's captured & recorded to tape be similar? These looked entirely different.

It was almost like the Cine-like Gamma wasn't doing it's job.

I love my DVX but I have run into this on a couple of occassions where the color correction phase of post was so much more involved than we expected.

I hope Jan has some thoughts but anyone please.

Thanks,
MR
 
Re: Making multiple cameras match

A couple of questions which might help us help you:

Were you using zebras in each case?
I assume you didn't have a separate monitor?
Were the scene file settings precisely the same? Both the number on the camera body and the actual scene settings?
Did you have a color chart?
Did you do any sort of camera matching before shooting?
Are these both 100As or 100?
 
Re: Making multiple cameras match

These are both DVX100s (no A) the company bought 4 of them for the movie "No Regrets" and we used them for commercial work too.

I white balanced both cameras mainly so they would black balance and then as we always do we shot at 3200K preset in the studio.

Both cameras were on F5 with the previously mentioned settings confirmed on both cameras. see below:

>>>>"24P (normal)  all Scene File settings were at Zero, (detail, black level, Chroma. . .) Cine-Like Gamma & Matrix, Thick V_Detail and Camera Menu setting - 7.5% Setup which the post dept. specifies. . . "<<<<

We know we're going to do a tape to tape final color correction on the finished spots so outside of making sure they are dialed in to the same settings we don't try to make them match.

My point is both were set for cinelike Gamma and the whole contrast range is different on both cameras. We did confirm that the white was at 100ire (with Zebras) besides if the one that was clipped was just over exposed then the black areas wouldn't have been crushed also.

I'd like to have a scope but the other shooters feel like when we used one it caused them to second guess and come back with footage that they hated. I still wish we would have one for these shoots (white cyc).

Once again this isn't about the color not matching but the contrast looking entirely different.
For anyone who goes back as far as me this analogy might help. One looked like Kodachrome that had been printed (duped) and the other looked like ECO.

We did have seperate monitors which ALWAYS brings the question "is that what it's going to look like?" I find (on professional monitors) that the whole CineLike look appears somewhat washed out especially if you set up to DVX camera bars. So we end up having the client look at the LCD monitor and say "this is more like we're going to make it look after color correction." In fact we did change out one of the monitors back at video village, and I have to say I don't know why. I helped facilitate the change over but I have to admit I didn't ask why we doing it. . .

I have a lot of experience actually matching broadcast cameras as an engineer as well as being a DP/Operator. I believe there is an issue with this camera, some setting that you can't get to thru the menu and that it doesn't manifest itself in our other multi-cam shoots because they end up lit so flat, to accomidate 3 - 4 cameras.

Thanks,
MR
 
Re: Making multiple cameras match

Hey guys,

Any thoughts here?

I plan on checking out all 4 of their cameras (and mine along with them) to see how inconsistant the look is. . .

MR
 
Re: Making multiple cameras match

Wow,

I'm surprised, no real replies to this thread.

Yesterday I set up a test with all 4 of the production companies' DVX100s and mine AND an XL2, on the white cyc.

all of the Panasonics looked like I expected them to look. I did introduce a Techtronics WFM90 waveform/vectorscope/monitor into the mix, so no overexposed white, etc. . .

BTW I tried grossly overexposng the cyc to see if that could have caused the problem. Perhaps the camera couldn't handle the contrast and clamped down the blacks as well. . .

Nope, the darker areas got milky and basically the blown out whites contaminated them, like you would expect. . .

plus the shots looked well exposed when we were shooting them. Just the one went to extremes, and I couldn't recreate the "look".

I mainly wanted to definitively narrow it down to the one camera and be able to say "this one needs to be sent in for some TLC. Now I have a test tape with 5-DVX100s all looking basically the same, and an XL2 with Pro35 adapter that looks great too. Just different.

There is a whole camp that would have loved for me to find a "problem" with the DVXs so they could point a finger and say throw that consumer crap away! We need to go back to our 16mm camera OR get a "real" camera in here for these jobs.

These 4 cameras have been used a lot, most have had some degree of service work performed on them.

Has anyone else had times where the camera was having a "bad day" and looked entirely different than usual?

Throw me a bone here, eh?

MR
 
Re: Making multiple cameras match

You've got a real challenge not having a true multicamera engineering set-up, and I'm curious about a couple of things...

With no CCUs, you ostensibly ran the composite video output from each individual camera to the WFM90 for scoping?

Did you say that you had a MacBeth chart / scale on-hand?

Did you run the cameras through a composite switcher (bypass all gamma, chroma controls) prior to the WFM90? I think this will help somewhat - giving you a quick reference as you switch from one head to another.

Effectively, you'll need to "CCU at the head" - which is what you're already trying to do. I'd suggest that you should get the baseline CCU engineering work done (e.g., master ped., iris, phase) prior to getting into the 100a's myriad cine-processing modes.

I'll point one of my sportstruck buddies, Gerry Platt, to this thread and maybe he can help further!

Cheers,

e
 
Re: Making multiple cameras match

Hey Thanks Overland.

First off these are DVX100s (no "A") yes I made sure that all settings (scene file & camera setup menus especially) were exactly the same. We've done this on several shoots.

Rather than try to actually match all of them we're just getting them to a "baseline" very "vanilla" look so as to have maximum latitude for post color correction on a daVinci.

During our test, I didn't use a MacBeth I used my Kodak Grey Card Plus. The white, middle grey & black patches on it for all 5 available DVX100s were identical levels on the WFM90, and white balance was (as close as I could tell) a match on all of the cameras with a 3200K preset in the studio.

The test was exactly what we'd hoped for.

The previous shoot however. . .

This is what I'm really fishing for:
Has anyone else had this problem?

Identical camera "setups" BUT one camera has a completely different look and nowhere near the expected look!

I mean in the contrast "range" specifically, chroma levels and phase were actually quite close.

One of our cameras had clipped whites and crushed blacks. The other had a quite low contrast image giving tons of latitude. . .

I really wish the one camera had looked the same way twice - maybe it's just NTSC hosing us! (-; - then I could have had it sent in for servicing. Was it the phase of the moon?

Once again we had the scope and chart on the test and not the shoot a week & a half ago. The other shooter felt that having a scope around was making him second guess his lighting and exposure and he was happier with his past results setting up to the LCD finder. ie: he ended up with a look after color correction that looked similar to (but better than) what he'd seen on the LCD while shooting. . .

I really feel that this is a transient problem, because I could not recreate it. I could clip the whites or crush blacks BUT not both on a white cyc setup. If I clipped the whites then the blacks were contaminated and milky.

If it would just break all of the time I could get it fixed. . .

Thanks,
MR
 
Re: Making multiple cameras match

just to be very specific, i know you have them all set to f5, but did you try working you way through the menu's and picking the command to reinitialize it back to factory settings??

it probably sounds redundant and you have like 1000x the experience i have but i would give that a shot.

matthew
 
Re: Making multiple cameras match

Hey,

Thanks Matthew.

I didn't go to the scene file and restore to initial settings. I DID perform the hardware reset (the recessed button on the side of the camera near the shutter controls) when we discovered the problem and it didn't seem to change the look of the camera.

A week later all of the cameras look the same!
It makes me wonder about a cracked circuit board or cold solder joint that doesn't always mess things up, but when we're counting on it. . .

MR
 
I haven't responded 'cause I can't imagine what would have caused it -- especially if it's working fine now!

All the main questions have been asked, and you've answered 'em right... can't think of what else to speculate on...

... only thing I can think to ask is -- what f-stop were you working at? The DVX looks very consistent over most of the f-stop range, but there are a couple of ticks at the very bottom of the scale (f/16+1 and f/16+2) that look pretty bad, diffraction gets the worst of the camera... perhaps you had the "bad" camera in that range? Although I can't fathom how that would lead to crushed blacks and blown-out whites... it's a mystery. The only other thing I can think of is that during the actual shoot someone changed the scene file dial. I mean, seriously, the only things that can explain what you're describing are a) a broken camera, or b) different settings. You've determined the camera isn't broken, so that leaves different settings... maybe someone switched it over to F6 or F4 when that one shot was done?
 
Thanks Barry.

Yes I've considered that as a possibility. It could have been switched to F6 accidentally, but neither of us working with the cameras caught it, if that is what happened. . .

F6 was still set at factory settings, so the black level would have been dialed down. We usually run everything @ "zero" because they always take the edit in for a tape-to-tape color correction on the Rank. Basically we are trying to leave the most latitude.

I love the DVX. No it isn't a $50K+ Pro camera. For the $3K+ that was spent on them they are fantastic.

Mine is very low mileage compared to the 4 that this company owns.

There are those who'd like to find fault with these and replace them. Preferably with a Pro level tool or at least just new 100A's or XL2s. . . There may be some logic to that. 3 have been in for service at this point.
Two for being dropped, one for a big speck on the front element of the zoom group, the one we had the "problem" with, BTW.

Today we used one and for an insert to go into an older spot shot on film and everyone praised the footage, tomorrow we'll be using 3 simultaneously again. . .

The "broken" (bent?) ;-) one is our back-up. (I'll have mine handy too!)

- Mistakenly switching the settings, OK. That's why they put that lip around the Scene File dial on the 100A, after all. Just wish we caught it, if that's it. . .

- Broken, it can be sent in and repaired.

Bizarre problem, cracked circuit board perhaps, that is hard to duplicate/diagnose. . . stinks!

Thanks again,
MR
 
I have some experience with still photography, so I'll take a stab at this from a different angle.

It sounds like your eye level camera was working correctly. In a scene with pure white and pure black, the CCD may not have enough latitude to show details in both. So the camera will set the exposure to the middle, which will give you blown out whites and crushed blacks.

The low contrast in your high angle camera may have been caused by lens flare. I've had this problem in still photography when the light is set in front of the lens. Even though the light is pointing away from the lens and you have a lens hood, some of the light still shines on the lens and causes the low contrast problem. This is problem is worse if you use a wide angle lens.

So you may not have any problems with your cameras at all. Just a thought.
 
Someone else on the boards owned 3 DVX's or something and one had a very different image from the others. It's an old, old, post but I'll try to dig it up and see if it's relevant.

I've only done one multi-camera shoot, and I found both cameras produced similar, though not identical, results.
 
The "look" of the high angle camera is what all of the cameras looked like when I tried to recreate the "problem". Also there was less chance of lens flare from the high angle than the lower angle being closer to the same height as the instruments and looking away from them.
The lower camera would have actually had an over-all ambient "flaree" because it actually saw more of the white cyc. plus we do check for lens-flare before we shoot. thanks.

I tried over and under exposure to extremes during testing if I crushed the blacks then the white cyc was grey, if I clipped the whites the blacks were washed out.

We are all concerned that this is an intermittent problem with this camera.

The thing I'm going to test next (per Barry's suggestion) is if it's possible that the one could have been switched to F6 and no one noticed. These are 100s not 100As. I just don't think the factory F6 settings will do this either. . .

Like a posted before their "normal" setting is everything zeroed out in the scene file. So as "middle of the road" as possible, since they always do a tape to tape color correction a the local lab/film transfer house.

These cameras will pass their 2 year (purchase) anniversary in May or June. They have been used used used. . . I think they should sell them as used and replace with new 100As before the next big project. This lets someone get a great deal on a DVX100 and gets them brand new cameras which with the Cine Like_V setting will reduce some of the "why doesn't it look like it did when we were shooting. . .?"

MR
 
Well, reading through this, no one mentioned that maybe it was a bad tape since it only has happened that time. Maybe record something at the end of it and see if it comes back? Or maybe by accident a magnetic source was near one of the cables or cam.
 
I suppose that's possible. I haven't seen the tape since it went into the post production environment.

Stranger things have happened. . .

Thanks,
MR
 
You must log in or register to reply here.
Back
Top