Home of Professional Cinematography since 1996
Metering TV Screens
>Do spot meters have a problem metering TV screens? Is there an exposure compensation I should perform. I've been told that because a screen is scanned that it will read different that a continuous source.
>Any words of wisdom?
>In my experience the spotmeter reads 1/2 to 1 stop lower than the film actually responds. Don't know of a formula, but I bet someone here does.
>Spot meters ruined more film than any other photographic investment that I've ever made. They give very precise readings of the tiny dot they measure. I produced (read: ruined) lots of film with a precision exposed tiny portion of the scene this way. I know only use my spot for measuring the contrast of a given scene from brightest to darkest and not for any "overall" exposure information.
>You may be experiencing the same with your spot reading of a particular portion of the TV screen you measure. The "technique" I use for still photography is to take an "ambient" reading with the ball right up touching the screen on the spot with a middle tone.. With my Minolta Flashmeter III this seems to work fine depending on the content displayed on the screen at the time of the reading and the content on the screen when I shoot.
>Most spot meters (as will as other meters) have a weakness for excessive blue(or lack or red), such as a TV screen. "Silicon blue cells" were supposed to cure this problem, but haven't in my experience.
>Theo Van de Sande, ASC taught me this :
>Use an old spring-loaded (analog) Spectra Pro with the reflective light disc on the meter..place it right on the monitor, hopefully a close-up with skin tones..be sure and use the little dots and not the arrows for your calculations. It works.....it really does!
>You still must adjust the color for tungsten film, an 81A viewing glass helps.
>Wayne Kennan, ASC
>Many cameramen use the Minolta digital spot meter for most things, but have learned that it has a weakness when reading TV screens because of the nature of the moving bright spot target of the TV screen's scanning, combined with the Minolta's instantaneous measurement. It leads to misleading readings.
>Most cameramen on the forum agreed that the Pentax digital meter seems to "integrate" over a longer period of time and therefore averages the exposure better of TV screens.
>My recommendation is to use a manual exposure camera to shoot B&W Polaroid 667 film of the overall scene with the subject TV included in the test still picture. Use a shutter speed of 1/60th or slower to be sure to photograph the entire scanned image on the TV screen. You will be surprised how *dim* you need to set the TV so it won't look washed out on the film later. For filming, it turns out being set dimmer than most people would set it up to view it in a TV viewing situation in you home or office. You can't trust your eye, trust the Polaroid's, they do not lie, your eye can be fooled.
>Also another pitfall is you need to light your scene with daylight colored light (HMI) and put an 85 filter on the camera if you are using tungsten balanced film. Or light with tungsten and filter the TV screen with a piece of 85 plastic in front. Many larger TV's have a piece of clear glass in the front of the cabinet, you replace that.
>Or a third alternative is to have the video taped image itself shifted in the orange direction before you shoot film. This works if the video images are incidental, but I wouldn't use this technique if the video images are important, like a commercial for TV's!
>On a similar subject, it appears the Pentax does a better job of properly reading greenscreens and bluescreens than the Minolta does. I keep a Pentax around and only bring it out when I am doing TV screens or green/blue matting.
>I've found wildly differing readings from spot meters on TV screens, I've tried the Minolta, the Pentax Digital & the Sekonic L778, they are all out to some degree, the Pentax seems to be the most reliable in this situation.
>I always try to use a Polaroid to check this.
>If you're stuck then when I was a kid I found that 1/15 at f4 with 100 ISO was right for stills, I've used it as a basis for the last 30 odd years and it seems to work!
>I simply find the area that is closest to 18% grey (may be an actual playback of a grey card), then I take a reading with my spot meter and I underexpose it by an amount between a third and 0ne stop. If you're really in a hurry, just find the brightest spot of the actual image you're filming and expose for 18% with your spot meter; it works fine!
>Do a test before shooting and you can't go wrong. If you don't have time for a test, roll a 5 feet test (in stop increments) off the reel you're using, so you will have a reference next time.
>In my experience, yes and yes to all the responses. A lot of spot meters will be confused by the scanning process of the TV screen, and don't "integrate" well. Remember (here we go again) the vertical blanking interval actually turns off the display for the next retrace, much like an electronic version of the shutter in a projector blocking the light during pull-down.
>Result? The video screen is blank, or "black" some of the time. Faster than your eye can interpret. These dark periods confuse the meter, it tells you to open up to compensate. End result is overexposure... usually something like 2/3 of a stop.
>If you are running your own source material through the TV monitor and you have the luxury of setting up to bars, I usually read the green band (this correlates pretty well with 18%) with the Pentax and open 2/3-1 stop. Alternate method, read white and expose the picture Zoned between VI & VII.
>Several years ago Ken Zunder told me his old Model M Minolta Spot read TV's and computer screens OK, but his newer Model F always seemed to fail him . . . there must be an integration differential between the two models . . . he was the monitor king while shooting the first season of Seaquest.
>I've always been able to meter TV screens fairly accurately with a spotmeter (Pentax Spot V; haven't tried my Spotmeter F on a TV screen yet.) I also measure color temperature by reading random static off the screen with a color meter. It works quite well.
>I just photographed some insert shots of a TV. screen playing back some videotape footage on 16mm at 29.97 using 320 tungsten stock with an 85. I simply grabbed a 35mm still camera, set the shutter speed to 1/60th, the film speed to 200, meter to center-weighted average mode and measured the on-screen image. I then worked the brightness of the on-screen image until I achieved the desired stop. All is well.
>I shot a commercial including lots of TV screens two years ago using 500 ASA Fuji stock. Before shooting I adjusted the TV's "picture" pots to obtain low contrast on the screen. I used my digital Pentax spotmeter and slightly (2/3 f-stop) underexposed the screens than the reading. The result was OK. This method was also mentioned in the manual of the Pentax spotmeter so I didn't make an invention.
>I always wondered why the monitors looked 'hot' first few times I tried to meter them. I gravitated to a system where I intuitively turned _down_ the brightness control to a point where I thought they would look right; that seemed to work. Later I ran into a situation where I had to balance an odd group of monitors and came up with my current system.
>First I choose a specific video image (whole screen) that I will meter on. I try to pick something that is mostly Zone 5. At the very least I try to find a shot with an _average_ brightness over the entire frame. Obviously, a c/u of a grey card properly transferred would be the best. Next is the most important part: I back up far enough that the set fits _entirely_ into the target circle of the meter. Now, I realise that this may not be possible. But, if it is, and you try metering from close and far you'll see what I'm talking about. The difference is usually about 1/2 or 2/3 of a stop, or even more.
>I always find the tricky part is choosing the frame to meter on in absence of a full field grey chart. If I have time, I usually get a grey card, lit to key, close to the set as seen from the camera so I can make last minute adjustments based on that. Usually, I feel that I should turn down the brightness on the monitor. I have never had a producer say, "Gee, that TV looks dark." They seem to command a screen presence that doesn't always require a full textbook (as I f) exposure.
>Also, Art Adams mentioned using video noise for a Kelvin reading and I think that works well; you can use the same noise as a Zone 6 (maybe 6.5) reading.
>We seem to be reaching a consensus here. Had a chance to absorb all these postings, and tried to put them into effect on my 35mm shoot yesterday (3/18/97). In this case, a 26" commercial grade "TV set" was being fed an electronic blue signal, so that in post-production footage could be inserted via an Ultimatte-type key on a Henry. Attended the transfer this morning (3/19) and the colorist said the TV screen seen in the footage was "as good as he could hope for" in terms of exposure and saturation. No Power Windows required! So I guess it all worked.
>Details: Arriflex 535A camera, shooting Kodak '79. I rated it at a 400 speed (personal preference), process normal. No filtration. Frame rate was 29.97 to sync to TV screen. Rest of the set lit to an F4 incident, as metered with both a Spectra (my gaffer's) and a CineMeter II (mine). Metered the TV screen with my Minolta Spotmeter F, ASA 400, set at 1/60th second, still rate (1/2 the frame rate, 180 degree shutter).
>Adjusted contrast & brightness on screen until spot meter reading off the blue from the TV was 2.8 1/3. Based on previous posts, I assumed the red and green cells in the meter were getting little or nothing, and therefore the meter was under-reporting. Allowed additional 1/3 stop for "no red" and another 1/3 stop for "no green" and we were at F4.0, the desired stop.
>THEN, had the video tech switch the signal feeding the set from blue to white (presence of all colors). Spot meter reported the exposure as F4.0... but remember, that's to achieve an 18% grey card value, and if we were reading skin tone, we'd be opening up 1/2 to 1 full stop, right? But, on the other hand, I remembered the suggestion posted here, to take the electronic blanking interval into effect (seems to confuse the Minolta meters, which don't "integrate" the chopped source from a TV set well) I decided the two factors would cancel each other out, and I accepted the F4.0 reading as desirable and shot the scene.
>No complaints, the colorist said it was fine, the Henry artist said the blue-screen matted out like a charm. Client was pleased, booked another job with me.
>Thanks to all here for their input, that's what's great about this forum!