Cinematography Mailing List - CML
    advanced

Varicam – Sync, Sound & Focus

I just got back from shooting a Varicam period drama where certain things unveiled themselves on the shoot. I'd heard that sync sound was an issue anywhere below 60fps and recommended the soundman used a DAT and that we sync to clap given that we had to end in 24p or 25p.

Has anyone out there experienced sound sync problems at normal acquisition speeds, and if so in what frame rates and what were you coveting to?

In the end I shot 25fps because as usual the end usage wasn't clear. "We want it both tv and cinema friendly". Given that, I chose 25fps because in the end there are more 25fps offline solutions available over here. The problem comes because in getting the footage into IQ for conform for instance, if the footage needs to go to a film burn it will slow down by 4% or gain 4% in length - had I chosen 24fps, in going to tv it would have sped up 4% or shortened 4% in length. We have the same issue in shooting film for both tv and cinema. Also, (and this may seem a weird question but I have to ask this) has anyone experienced no sync problems and if so, at what frame rates coveting to what?

Secondly...there's a recommendation in the manual to measure all settings on focus from the black screw on the handle above the focus mark on the viewfinder side of the camera. On checking the lens we found every measurement to be incorrect and imagined the lens markings to be incorrect until I realized that all calibrated markings on the Canon J 18 lens were set from the front lens element ! So: Anyone know what the focus mark on the side of the camera is for ? Sorry if this is a dumb question, but I need to know - I thought it was the place where the image was formed - but hey.

Thirdly...I shoot at -3db for safety at every point (given that the camera is rated by Panasonic at 800 ASA in tungsten position) and, always underexpose the image by at least half a stop - but I got perceivable noise in the lower lit areas - Anybody else come across this phenomenon? I find it hard to believe that the Varicam is so much more noisy than Sony kit.

Lastly...the so called "film menu" : Does anyone believe this stuff?

I remember when Digi-beta came out there was a healthy sales in "film look" chips but I took it that these were just a manipulation of internal settings and that if I were going to mess around inside the menus anyway then these chips were just a few extra pounds/dollars made by Sony from people's lack of confidence....and therefore It seems to me that the "film user mode" terminology in the Varicam is a sop to people who may be used to film language rather than video language, but my take was, if you use the EC3 remote unit then the knee, gamma, pedestal and colour paint dials etc would make all the film and tv gamma settings redundant. I understand that if you come solely from film you have a language that you're comfortable with - but does this just confuse the issue in the end ?

In the Varicam there was an added layer of "concern" in the shoot. As usual out front of the camera there's the lighting and how the camera receives that light (and how you have to build a tent for the monitor so that you can truly see what you're doing). Then there's the concern related to the manipulation of the image behind the lens - the mending system which I try to get beyond by using a remote so that I get hand and eye co-ordination together rather than prod a switch then look at an image, then prod a switch and again look at an image, etc....and - making sure at every conceivable point that it was going down ok and that of the many of settings I was playing with I wasn't missing some obvious simple foul up. it just seemed that there was more "mastering" to do than normal. Was it just my paranoia or have other people experienced this extra layer of concern ?

phew!

Terry Flaxton
UK
http://www.flaxton.btinternet.co.uk/


So : Anyone know what the focus mark on the

> side of the camera is for ?

You will find that certain manufacturers of HD lenses use the film style film plain markings and others use the front element, I won't mention names here.

Nick Paton
Film & Digital Cinematography
www.npdop.com


If this was for a feature where a color-correction session in post is planned anyway, I don't think it's necessary to spend a lot of time on set playing with the menu (I mean, you can if you like to...but many of these functions can be done in post anyway as long as you record enough information onto the tape so you can play around with it later.) There are only a few irreversible things like adding edge enhancement, plus clipping or crushing exposure information, that you have to watch out for in-camera.

For me, there is a limit to how much time I want to devote on the set to playing colorist when I'm just going to be doing all of that later anyway. My main goal is to record enough information and to get in the ballpark with my colors & contrast and then plan on making adjustments in post color-correction (because the average feature has hundreds of cuts, so some color-correction later is expected.)

Sometimes I think people worry too much over getting the "perfect" menu settings in their camera as if it were a life or death decision as to what gamma level to use, what percentage to use for black gamma, etc.

David Mullen
Cinematographer / L.A.


David Mullen wrote :

>Sometimes I think people worry too much over getting the "perfect" >menu settings in their camera as if it were a life or death decision as to >what gamma level to use, what percentage to use for black gamma, etc.

Amen.

I just had a feature screened at Tribeca film festival that was shot with a Varicam…

http://www.panasonic.com/PBDS/subcat/newsinfo/press_03/03_28.html

We had several basic set ups for the camera and within those set ups gammas were manipulated slightly according to situation. Post production schedule was extremely tight and we just managed to deliver for screening. We did not have a chance to color correct anything after the conform which had me very concerned. In the end it all went well, even though it needs some color polishing in my judgement.

Lets keep in mind that Both Sony and Panasonic give you control over every function and electronic parameter in the camera and in very fine increments, so fine in most cases that a few clicks one way or another are practically meaningless. You get infinite adjustability, but the most important thing is to figure out what scale the adjustments need to be done in. So if you have a "0" setting and the possibility of going + - 99, should you do adjustments in steps of 5, 10 ,20 or what before you see an effect.

Then there is the problem of judging the effect of what adjustments you make. If you have a 24" monitor what can you see on it compared to seeing your image on a 24 foot screen that it could be projected on as a final result?

I really don't have answers to these questions even after seeing my film on a 24 foot screen. If you are lucky, as a DP you have budget to test all the way through the projection process, but the reality is that if you are shooting a feature on HD, it is probably an indie feature without the budget to test all the way through projection, so you get to play the part of crash test dummy. This is just another feature of the digital landscape we moving into.

Mark Smith DP
Oh Seven Films Inc.


David Mullen wrote :
>
>Sometimes I think people worry too much over getting the "perfect" >menu settings in their camera as if it were a life or death decision as to >what gamma level to use, what percentage to use for black gamma, etc.

Yes, very true.

The ideal HD camera will, like film, record everything the CCD puts out, for later manipulation, with no compression or processing or (heaven forbid) enhancement or matrices.

Until that point, it always will be a compromise to some extent, because while the camera may giveth, the recording format (and the camera's processing and data reduction) taketh away.

Jeff "has preached this sermon before at the Church of the Unprocessed Pixel" Kreines


Terry Flaxton wrote :

>>Thirdly....I shoot at -3db for safety at every point (given that the camera >is rated by Panasonic at 800 ASA in tungsten position) and, always >underexpose the image by at least half a stop - but I got perceivable >noise in the lower lit areas

I gaffed on a few shoots recently where we had the same phenomena when watching on the LCD Panasonic HD-monitor. We checked the gain-settings, but they were set to 0dB. We even thought for a flash that it might have been a non-terminated BNC-cable. Apparently when the material was viewed back at the post-facility, it looked clean.

So we think it might be an issue with the Panasonic monitor.

Cheers

Martin Heffels
wannabe SFX-DP/editor,
Sydney, Australia


Mark Smith writes :

>>Then there is the problem of judging the effect of what adjustments you >make. If you have a 24" monitor what can you see on it compared to >seeing your image on a 24 foot screen that it could be projected on as a >final result?

This has to be the strongest argument in favour of an uncorrected picture when you shoot and a post grade, kinda the Viper approach.

Or shooting with a flat de-saturated image on the Varicam or F900.

Lets face it, if you've got the money to shoot the amount of tests that you indicate would be needed all the way out to film then you'll have the money to grade it, either as well or instead.

I still don't see the logic of having an entire crew and cast standing around as you try and tweak and guess what that small image is going to look like on a big screen when you can do it in a controlled environment under less pressure.

Cheers

Geoff Boyle FBKS

Director of Photography
EU based
www.cinematography.net


>>Thirdly....I shoot at -3db for safety

Forget -3db. Shoot at -6db unless you absolutely can't. Noise in the blacks is considerable at 0db. For safety you should be shooting at -9db, but alas, it doesn't go that low.

>>Then there's the concern related to the manipulation of the image >behind the lens

Then stop manipulating so much. Stay in Filmrec mode and set your camera up once to give you the most latitude in the areas where you want a lot of data to work with in CC - usually (but not necessarily) the blacks and highlights. This means a washed out image on tape, but a great canvas to work with in CC.

On the cheap, you can use still frames and Photoshop histograms to really dig into this set up process - props to Scott Billups who mentioned this technique in one of his articles. Then light properly, resorting to "menu madness" only to solve the hairiest of problems that you can't solve by physical means. If you keep monkeying with the settings you'll never get your groove on in the other departments because you have no baseline response from the camera.

Some say signal processing control is an advantage of HD over film. I think behind the lens manipulation is an artefact from the broadcast origins of these cameras - the need to tweak in the field so you can fit vastly varying light conditions into a very limited dynamic range. But the dynamic range isn't that bad on the Varicam - especially if you set it up well, and in a production environment light can be manipulated, so that old way of using a camcorder needs to be retired.

Why build skills manipulating data when you could be building skills manipulating light?

If there's no CC in your chain, then you are forced to play the painful WYSIWIG game. But the same "ideal" applies - get a good solid setup in a controlled environment where you can really think it out, then manipulate in front of the lens, not behind.

Greg Gillam
Producer / Director
Red Sands Production Co.



>Forget -3db. Shoot at -6db unless you absolutely can't. Noise in the >blacks is considerable at 0db. For safety you should be shooting at >9db, but alas, it doesn't go that low.

Greg,

Please correct me if I'm wrong, but I understood that with the filmrec on, your gain was limited to -3db. You'd need to take that off to achieve -6db, which, while cleaner, will limit your dynamic range.

Take care,
Chris Cooke-Johnson
Creative Junction
Barbados


Greg Gillam wrote :

>>Forget -3db. Shoot at -6db unless you absolutely can't. Noise in the >blacks is considerable at 0db. For safety you should be shooting at >9db, but alas, it doesn't go that low.

I have to say that having seen " The Look", a film I shot with Varicam 27v ( not even and F) projected from Hd on a 30 foot screen in a 500 seat theatre, that noise was not a factor. The film includes a number of very low key scenes where we were at my limit T2 or T2.8 for lens aperture and underexposing in general, shooting at -3db and 0 db when helpful and I never saw what I would call objectionable noise on the screening.

I recall from when shooting the film that there was apparent noise on the monitor in these low key scenes but in the projected version the apparent noise was really was not as great as I expected it could be.

Mark Smith


You are correct, Chris. On the Varicam -6 db is not available in the FilmRec mode. I do all my shooting in Film Rec now and see no problematic noise in the black at -3 db.

Leo Ticheli
Director/Cinematographer
Birmingham/Atlanta


>>Thirdly....I shoot at -3db for safety at every point...but I got perceivable >noise in the lower lit areas...I find it hard to believe that the Varicam is so >much more noisy than Sony kit.

With full disclosure that we own the F900 and not the Varicam, one reason that we did not purchase the Panasonic camera is exactly because of how noisy the image was, shooting side by side with the Sony. I've observed this in our own tests and on the footage I've seen as well. While I understand that there are people who prefer the Panasonic look, the additional noise is undeniably there. If the economy were better up here in the Bay Area, we would probably purchase a Panasonic as well, to better serve the market but I would alway disclose to clients the additional noise and also the "coarser" picture created by fewer imaging elements. I know that Walter says that since you are throwing away a lot of that info from the Sony when you record to HDCam, the recorded images are the same, but I totally disagree based on what I've seen with my own eyes.

Oversampling works!

>>Lastly - the so called "film menu": Does anyone believe this stuff?

With an engineer who knows what they are doing, you don't need those pre-programmed settings. But I believe that since many shoots don't hire an engineer, or prefer to hire a camera assistant who has been trained on where the menus are but who doesn't really understand what they all do or how they interact, these presets give them some comfort. And they are a great marketing technique - must have the "cine gamma curve" for my shoot. I think Panasonic just got a leg up on the Sony marketing machine. We'll have to see what Sony's answer will be.

By the way, thanks to Art Adams for the nice words!

Leigh Blicher
Videofax
San Francisco


You can't call HDCam oversampled, since the CCD resolution is 1920 x 1080, the same as the format it ostensibly records and outputs. Oversampling would require a CCD with greater res, ideally 2x in each direction, or 3840 x 2160, to make Mr. Nyquist happy.

The fact that Sony decimates this to 1440 x 1080 still doesn't mean that you can refer to it as oversampled. But I think you could call it "under recorded."

Not to mention over-compressed....

Jeff "semantics cop" Kreines


>>Oversampling would require a CCD with greater res, ideally 2x in each >direction, or 3840 x 2160, to make Mr. Nyquist happy.

Oops, a correction is required.

Martin Snashall (not on CML) adds this :

Even then Mr Nyquist would not be really happy. You can't really apply Nyquist theories to a CCD, which averages over an area. Nyquist sampling only works on a continuous waveform, and then with a infinitesimally short sample point. Point a CCD camera with 1920 pixels at a 800 line pair chart, and the frequency response will be all over the place!.

Of course, that doesn't mean that oversampling doesn't have its advantages even if Mr. N isn't involved.

Jeff "not an engineer" Kreines


Whoa hold it right there yourself Mr "not an engineer/semantics cop"

Leigh who is also not an engineer said that the picture looked better with less noise. This was her point. Whether it is over sampled under sampled Nyquist limited or whatever. The picture looked better to her and her client on the F900. I agree. I saw a very very good example of this with the DigiPrimes on a 27FP this last week in Brazil. I had never shot anything other than a lot of "tests" with the 27FP and had a chance to shoot real world.

This was a brand new in the sealed box camera with -3db gain. The emperor had no clothes. To be fair the image looked very nice and I really liked the camera. But quiet and sharp and crisp it was not compared to the F900. 720 X 960 is not 1080 X 1440 regardless of the knife used to slice it.

Michael "likes very sharp and quiet pictures" Bravin
Chief Technology Officer
Band Pro Film/Video, Inc.


Jeff Kreines wrote :

>>The fact that Sony decimates this to 1440 x 1080 still doesn't mean that >you can refer to it as oversampled. But I think you could call it "under >recorded."

Throwing gasoline on the fire, you could then start a discussion about "chroma".

Stepping away from the fire now...

Mark Smith DP
Oh Seven Films Inc.


Michael Bravin wrote :

>>720 X 960 is not 1080 X 1440 regardless of the knife used to slice it.

I agree completely.

My point was simply that there's no way you can call HDCam oversampled, as the CCDs' resolution is the same as the output format, not greater. Leigh appeared to be implying that since HDCam decimates the image to 1440 pixels, the extra res of the CCD could be considered oversampling.

Then again perhaps Leigh's point was simply that more resolution is better than less resolution, which I agree with. But that's not the same thing as oversampling.

>> Michael "likes very sharp and quiet pictures" Bravin

Should I put that on your tab, Michael?

Jeff "prefers very sharp and quiet and uncompressed pictures" Kreines


I agree the emperor has no clothes. Only in my tests, it was the CineAlta standing naked.

For television production, the Varicam has many advantages :

1/. A prettier picture, especially in the chroma department.

2/. A much friendlier form-factor with built-in SDI out and intervalometer instead of the added-on expensive boxes for the Sony that turn it into some kind of Franken-camera.

3/. Under and over cranking from 1 to 60 fps. This is huge; not just a subjective difference. If you shoot commercials, it's essential. There is one particular Sony salesman (not you, Michael) who keeps trying to convince customers that the CineAlta is a variable frame-rate camera. It's not.

4/. The Varicam is much quieter than the CineAlta with it's fans.

5/. The Varicam has a very affordable system price.

Now I'll give one to the CineAlta. It's much more widely available as is it's post gear. I'd also like to mention that I've got a good friend who owns a CineAlta and doubtless he makes really beautiful pictures with it. It suits his business perfectly, just as the Varicam suits mine.

I'm not sure these Sony vs. Panasonic debates are useful when based on subjective opinion of the look of the pictures, but I felt the need to offer my own view and experience. That emperor's clothes thing...

Best regards to all, whichever camera you choose,

Leo Ticheli
Director/Cinematographer
Birmingham/Atlanta


Mark Smith wrote :

>>Throwing gasoline on the fire, you could then start a discussion about >"chroma".

What, 3:1:1 isn't good enough for you?

>> Stepping away from the fire now...

But you're already a bit charred...

Jeff Kreines


I certainly believe that tweaking the camera on the set is not the best situation, but has it happen to me right now with a F900 in 24P, and so often on a documentary shooting, where, has usual they wont have money left for a good color correction, it mean get the best of what you can get right now. Not pleasant, stressful, but on the other hand it's help you to judge and respond fast if you want to keep a good continuity. At the end, it's a solid learning process.

Best.

Marc Gadoury, csc
Montreal


All this -3, -6 sounds quite conservative to me, but I'm not familiar with this camera yet.

Is this noise level (in the case of 0 or -3) a penalty you are paying for more highlight latitude in "Film Rec" or "Cinegamma" mode ?

And how much is this noise visible in a filmout ?

Sam Wells


Mea culpa. I should have put "oversampling" between quotation marks. What I meant is that the image that you get when you use a denser chip to take in information, even when you then go and "under record" (I don't argue deny that), is smoother or finer than the image you get using a chip with half the pixel count.

Sheepishly,
Leigh Blicher
Videofax
San Francisco


My bad - yes, -3db is the lowest setting in Filmrec mode. I shot for so long at -6db when my camera was a "V" that I still have that -6db thing in my head - demonstrating the "forget" part of my "set and forget" philosophy.

Also - I've seen this sync thing come up many times but I have never had any sync problems at 24fps. I'm usually not critical of sync on off speed material, so I could have missed problems at other framerates, but now I'm concerned as I need to shoot something at 25 this summer. Is there anything to this topic that seems to pop up every now and then? I've always assumed people were just seeing a processing delay in the image as it passes through an HD to SD converter. Whazzup?

Additionally, I've had problems off and on with my auto gain distorting signals that are only -10 db on the meters. I'm generally running manual gain with a sound mixer pre-camera or going to DAT, but when I need to get something off the camera mic and I switch into AutoGain I get problems. Anybody with a Varicam experiencing these? My first thought was the mic, but I've gone through 5 different mics and always the same problem. I even opened up the camera (with Panasonic's blessing) and set the channel jumper to a lower gain, but that didn't solve the problem.

Greg Gillam
Producer / Director
Red Sands Production Co.


Greg Gillam wrote :

>>My bad - yes, -3db is the lowest setting in Filmrec mode. I shot for so >long at -6db when my camera was a "V" that I still have that -6db thing >in my head - demonstrating the "forget" part of my "set and forget" >philosophy.

I've been silently watching this train go by for sometime, but now wish to now make just one short comment : Remember that every db that you reduce the gain is a db that is no longer available for headroom. Thus, for sake of argument, if you reduce the gain by 6 db to -6, you have also reduced the headroom (or overexposure latitude, if you prefer) by that same 6 db (one stop) but you have improved the signal to noise ratio (or grain) by that same amount, 6 db...As in everything else, it is a compromise, and you decide which side of the equation is the best compromise for the task at hand. YMMV.

Tom Tcimpidis


Tom - reducing gain shifts the exposure latitude rather than diminishing it, correct? No? In tests it appears that I have the same latitude at -3 as I do at 0. Am I missing something?

Greg Gillam
Producer / Director
Red Sands Production Co.


Greg Gillam wrote :

>>Tom - reducing gain shifts the exposure latitude rather than diminishing >it, correct? No? In tests it appears that I have the same latitude at -3 as I >do at 0. Am I missing something?

This is a gross over-simplification but think of it this way : You have, for sake of argument since the actual numbers are different, 1,000 possible steps of information with which to work, from sensor black to sensor maximum white. This total number is fixed and can not be changed by user control. At zero db gain, we save 500 of those steps for the exposure area above what would be nominally 100% (or .7mv) video level.

This is our over exposure latitude or headroom (which in this example is one stop). If we go to -6db gain, we still have the same 1,000 steps but we have expanded the entire range up by 100 percent. We now, in this example, have no over exposure latitude or headroom whatsoever since step 1,000 now corresponds to 100% nominal level. Any information above this point will now be lost and can not be recaptured by knee compression or the like. We still have the same total dynamic range and actually have (2X) more available steps in the nominal exposure region (and less noise), but have gained this increase at the expense of the ability to capture information above the nominal exposure region. So, yes, the absolute total latitude is nominally fixed by camera design; we can only move the window within that latitude to best leverage that window for a given purpose. Also keep in mind that one camera's zero parameters may very well be different than another camera's zero parameters - rarely are the two directly comparable.

As always, there is no free lunch.

Tom Tcimpidis



Copyright © CML. All rights reserved.