Light meters for DV camcorders

[Carlos E. Martinez]

Charles,

My intention was not to declare myself as a person who knows all and aware of all techniques. And my experience is quite likely a lot less vast than many people collaborating or reading this forum, you included. So let's go on.

Of course I understand your concern and agree with you on how something we declare as "universal truths" are only so within a context, particularly experience on how to use a tool. And there should be a certain responsability in not being light in our statements.

It happens that the danger can be the opposite too: not showing tools or routines that could be problem solving if used with certain care. We are also used to how expensive certain accessories, like filters, can become. Perhaps after a year we can look back and see how several suggestions here became real helpers in spite of looking "dangerous".


Carlos

Maybe it's time to open a separate thread somewhere to deal with grad NDs, their specific problems and how to best use them.

Your words there to be careful and mine to use them might work as they should: positively.

My concern with the word "budget" was mainly because I too wanted to take it out from seeing it as the equal of "cheap" and, worst of all, the opposite of "high quality". A strong example comes to my mind, which is how the late Nestor Almendros used mirrors to light the inside of a hut in "The Blue Lagoon", as he himself describes on his book. Coming from a budget oriented past, he used simple tools to get great results.

[Helen Bach]

Carlos foolishly asked: Helen: can you ellaborate more on that camera stop/NLE numerical values routine? That seems interesting. so blame him for what follows! Well no, blame me.

Best,
Helen

Here we go. Although I would like to make this thorough, I will compromise and make it brief. Apologies for any lack of clarity that creeps in, please feel free to add your own clarifications or corrections, or ask for more explanation. For the sake of completeness I'll cover some ground that I'm sure you already know. Of course, I make no claim that the methods I'm giving are correct - they are just one approach, and much variation is possible. For brevity I have not explained all the implications of varying from the given method.

Preface: the most important conversion is from emotion to image. There are no formulae for that.

Method 1: for use with an incident meter (assuming no real-time digital WFM available).
Point the white-balanced camera at an evenly illuminated 18% grey card and defocus enough to lose any surface detail. Either let the auto exposure settle on an aperture setting (shutter speed remaining constant throughout), or set the aperture manually to give mid grey. Adjust the lighting so you get an approximately correct exposure at you favourite aperture (or exactly mid-range, say 5.6), then fix that aperture. Now set your meter to that aperture and adjust the exposure index ('film speed', EI) so that the meter is balanced when placed on the grey card (use a flat receptor if you have one, rather than a dome). This tells you the approximate speed of your camera, and this will be sufficient for most purposes. To get a more accurate answer, record a series of tests for later analysis with an NLE or whatever. Suppose that the approximate speed comes out to EI 160. Set your meter to EI 100 and adjust the lighting until you get the correct reading at your chosen aperture (throughout these particular tests the exposure should be varied by varying the lighting, not the aperture). Record a few seconds. Do the same at 125, 160, 200 and 250 EI, varying the light each time. Now import the footage into an NLE or After Effects (or import a still into an image editing program such as Photoshop) and use whatever tool is available to read the RGB values for each EI setting. Whichever EI setting leads to RGB values closest to 127 is the camera speed.

The characteristic curve of the camera can be determined by using the aperture and shutter speed settings to over and under-expose over a range of stops from the 'correct' reading obtained by using your meter at the EI you just determined, then reading off the RGB values as before. This method is dependent upon the accuracy of your aperture and shutter speeds (the shutter speeds should be accurate!).

Method 2: for use with a reflected light meter (including a spot meter).
This is much the same as Method 1, except that you can use any neutral-coloured card, it doesn't have to be an 18% grey card. With a spot meter you can also use a grey scale when finding the characteristic curve. The method should be obvious.

Using a waveform monitor
If you have a waveform monitor (WFM), just set the camera aperture to give 50% IRE, then find out from the meter what EI will give the set aperture. If you were using an analogue WFM, it would be worth checking the recorded digital signal with an NLE.
Waveform monitors that read the analogue output of a digital video camera are not necessarily the best way of measuring the signal. The best way is to read the values of the digital signal. Some WFMs do this. This applies to using a monitor from an analogue signal as well.

T-stops and f-stops. If your camera has T-stops, they are the 'correct' ones to use. If you only have f-stops, use them. You are measuring what is important to you: the effective speed with your camera and your meter. As long as you are consistent in the use of T-stops or f-stops and don't mix them you will be OK. The effective speed using f-stops will probably be a third of a stop slower than using T-stops (no big deal really).


end of part 1

[Helen Bach]

part 2

Background:
Grey cards. The standard grey card has 18% reflectance. This is taken as mid grey. 18% is about 2.5 stops down from 100%.
In video this is usually set at between 45 and 55% IRE (assuming a range of 0-100% IRE for the sake of this discussion, using a % value instead of absolute IRE to avoid set-up complications) depending upon how you want your video to look (the mood?).
Kodak say 45% IRE (they use % IRE for the same reasons that I'm using % IRE) for an 18% grey card. I'll use 50% for this.
How does that transfer to RGB values? 50% IRE transfers as 127, which is half way between 0 and 255 (ie 256 values). The relationship between IRE and RGB values is normally linear, sometimes with an offset to the IRE values.

Hold on, video is supposed to be linear. Shouldn't that mean that half the light leads to half the IRE, so one stop down (rather than 2.5 stops down, for example) should be 50% (and 127 RGB)? No. When video is said to be linear, the linearity is between the response (IRE or RGB) and the log of the amount of light reaching the chip. This is easy to explain with diagrams, but I'll try to put it into words. Imagine the familiar response curve of video: the line is straight and rising left to right. The left-hand (vertical) scale is IRE, and is linear - ie the distance on the graph between 20 and 30 IRE is the same as between 80 and 90 IRE. The bottom (horizontal) scale may be 'stops'. The physical distance between stops is the same. This scale is not linear in terms of the amount of light falling on the chip. Suppose that the stop scale runs from 0 to +8. The half-way mark (+4 stops) is not the half-way mark in terms of light, it is 1/2 x 1/2 x 1/2 x 1/2 = 1/16 of the way in terms of light. So it isn't really linear. To put it more technically, gamma is a log-linear ratio, not a linear ratio.

There is further non-linearity because video cameras do not have a straight line relationship between the IRE response and the log of the amount of light reaching the chip. They will show some roll-off at the shoulder (top right of the curve) - to compress highlights a little and give a bit more dynamic range before those white holes appear. There will also be some rounding at the toe (bottom left of the curve) which can lead to the muddy shadows characteristic of bad video.

What does calibrating to 18% mean? It means that an object with 100% reflectance (there is nothing that achieves true 100% diffuse reflectance, by the way) will not be over-exposed if it is illuminated at the same level as the grey card, and if your camera has 2.5 stops of latitude above the middle value - which it probably has.

Is an 18% grey card necessary? Only if you are using an incident meter to calibrate. Any neutral coloured card will do - grey or white - if you are using a reflected light meter, including a spot meter. It's best to avoid white paper with optical brighteners. The value of the comparatively expensive grey and white cards made for photo, film and video use is that the manufacturers should have put some effort into achieving a neutral colour. The reflection density (greyness) of the card does not matter, so long as the card fills the view of the camera and of the light meter cell. If you are using an incident meter, you must use an 18% grey card.

Effective speed, the film analogy:
There are at least two distinct ways of determining film speed: toe speed and mid-tone speed. The former is exemplified by various standards and by the zone system used in B&W still photography (expose for the shadows, develop for the highlights). Toe speed is primarily a feature of the film: the development process has comparatively little influence. The second method is exemplified by the LAD (Laboratory Aim Density) method. The LAD method emphasises the importance of mid tone/skin tone exposure. In a nutshell: the film manufacturer tells you what the R, G and B densities should be for an optimum grey card exposure. For current Kodak motion picture camera negative films they are 0.80, 1.20 and 1.60 using Status M filters in the densitometer. To determine the speed you do a series of grey card exposures under different lighting levels using the same camera/lens settings. You set the lighting level by fixing the aperture setting on your meter (assuming the shutter speed is also fixed) and adjusting the film speed index in third-stop decrements and increments around the manufacturer's rating. So for 250 speed film you would set 160, 200, 250, 320 and 400 for example. You would then adjust the lighting level to give the correct exposure at each speed index.

[Carlos E. Martinez]

Thanks a lot, Helen!


Carlos

[John Jackman]

Originally posted by Carlos E. Martinez : No doubt a CRT monitor and a waveform are excellent tools to work in video, probably some of the best.

But metering a frame using a combination of incident and spot meter may go a bit farther.

Not as far as a field waveform monitor, which is really the ultimate way to go.

As long as you can use the right ASA reference (where do you get that for different camera models?) and know how to handle a spot meter (not many really do) what is the real problem?

No problem if you compensate for hot spots more aggressively than film --

Two techniques for finding the Exposure Index are in an appendix my book (ASA technically refers only to film).

[Helen Bach]

John wrote Two techniques for finding the Exposure Index are in an appendix my book (ASA technically refers only to film).

John,

Just out of interest, are they anything like any of the methods I mentioned?

Thanks,
Helen

[Carlos E. Martinez]

<<<-- Originally posted by John Jackman :
Not as far as a field waveform monitor, which is really the ultimate way to go. -->>>

It's quite likely I may be alone in saying this, but I still believe that in practical terms (like portability, ease to use and quickness) an incident light meter comes first, a spotmeter second and fwm third. All three can be extremely accurate and consistent if properly used.

The fourth most important element is a well regulated field monitor. In fact, using just one of those three with a good monitor might be the right combination.

Perhaps a mix of incident meter and fwm could be the best, as I think the incident ball provides a consistency that is fast and easy to follow. But this is probably nitpicking and a matter of opinion. All three can be used properly and provide superb results.

Of course we are talking of situations where we can adjust the lighting. If the lighting is set, all three should provide the right stop. Though in the end a good monitor and the DP experience should decide which stop that is.

My intention is to suggest which tool might be better for every situation: documentary, fiction, location, studio, etc.

<<<--No problem if you compensate for hot spots more aggressively than film -->>>

No doubt. The about eight stops difference between film and video latitude demand to do that. Perhaps some day we will be able to "paint-down" hotter spots during recording.

<<<--Two techniques for finding the Exposure Index are in an appendix my book (ASA technically refers only to film). -->>>

Can you comment about them here?

Of course ASA refers to film index. But we can cross reference film sensitivity with video sensitivity and get to some numbers. In fact camera manufacturers could provide this number.



Carlos

[Karl Heiner]

i will shoot a out door theatre 2 part performance. first part is during day light, and the second part is at night with stage lighting.

my thought is to sit in, during a performance prior, and take meassurements with a spotmeter to get some readings.

could anybody comment please..

thanks

[Charles Papert]

Karl:

OK in theory, as long as you know how your particular camcorder relates to the resulting range of f-stops. Unless the performs happen to be wearing nice flat costumes made of 18% gray material, you'll obviously be taking numerous readings of highlights and shadows and interpolating the results, then making an evaluation of what exposures seem best, then setting those on your camcorder for the taping--but whether or not those are the optimal exposures that deliver the best image may be suspect for the reasons detailed in this thread.

If you have the camera in your posession for the rehearsal, it may be more efficient to simply bring it along and shoot some tests; you can try bracketing exposures (make note of the timecode and f-stop, or whisper the f-stop into the onboard mike) and then you can evaluate the footage on a good monitor (or use a waveform, histogram etc. within your NLE) to give you a foolproof set of stops that will serve you for the actual taping.

That said, the day exposure may be variable depending on weather and how much ambient light penetrates the proscenium. You might have to improvise on the day (and on a partly cloudy day, be ready to ride the exposure).

[Karl Heiner]

Originally Posted by Charles Papert

Karl:

OK in theory, as long as you know how your particular camcorder relates to the resulting range of f-stops. Unless the performs happen to be wearing nice flat costumes made of 18% gray material, you'll obviously be taking numerous readings of highlights and shadows and interpolating the results, then making an evaluation of what exposures seem best, then setting those on your camcorder for the taping--but whether or not those are the optimal exposures that deliver the best image may be suspect for the reasons detailed in this thread.

If you have the camera in your posession for the rehearsal, it may be more efficient to simply bring it along and shoot some tests; you can try bracketing exposures (make note of the timecode and f-stop, or whisper the f-stop into the onboard mike) and then you can evaluate the footage on a good monitor (or use a waveform, histogram etc. within your NLE) to give you a foolproof set of stops that will serve you for the actual taping.

That said, the day exposure may be variable depending on weather and how much ambient light penetrates the proscenium. You might have to improvise on the day (and on a partly cloudy day, be ready to ride the exposure).

hello charles,
thanks for your advice, and you're right. i am better of to sit in with my camera and do some shooting/ measurements/ readings in regards to lighting and sound. i will also tape on vhs for a later evaluation.
my biggest problem will be the audio, since they do not use any sounds system. "shakespeare". by the way, they have the best costumes, flat, props etc i have ever seen. most of my ethnic dance troupes have white satin costumes!!

thanks again. will keep you posted.

greetings

karl

[Stefan Day]

are you thinking of doing anything for that sound situation? or jsut gettingthe best you can with the camera mic?

[Karl Heiner]

Originally Posted by Stefan Day

are you thinking of doing anything for that sound situation? or jsut gettingthe best you can with the camera mic?

hello stefan,

thats a very good question, and i don't have an answer at this point.
this is an open air theatre, which plays music during the performance, but the actors do not use any mic's whats so ever. a lot of spoken words and singing. there are 2 marathon shows, at the first one i will take charles advice, run the camera to get some readings in regards to lighting, but also to see about the sound. the acustic was excellent.
there is no way to set any mic's anywere.
i have seen both shows, and think that i can pick up enough quality sound with my on board mic.
since that is a big professional company, it is interessting to see how all those different level work, just by agreeing to have a videographer.

will keep you posted.

greetings

[Jos Svendsen]

What an interesting idea. Do a play with all actors in 18% grey square grey costumes, and using a giant Kodak Q-13 color card as backdrop. All lightning done with 5500K softies of course. All actors miked with lavs in advance, with all frequencies adverticed in programme.

Now - who could goof that? ;^)

[Karl Heiner]

hello charles,

well, it's over. what a long weekend. from 1:00 pm till mid night, both days.
did set up my camera on saturday switched between av and tv. (day light show) that did not rtealy help, since on sunday there was a total different light conditions, around 5 pm, both days, there was such a strong contrast.....
the two nights shows went very well, since they had excellent lighting. no refletive light from the costuems/ faces etc (over exsposure)
i like to think that the canon tents to have very warm colors, leaning into the color red. don't know yet how to control that. should i use filters?
the audio was terrible, no mics, no sound board, only my on board mic. lucky me since they do not care about the audio quality.

here are the things i learned:
warm cloth, incl gloves!
write down every setting on the camera during rehearsal
write down what the actors do
watch out for low flying bats

greetings

[Karl Heiner]

help....

just got some feedback from my weekend client, the afternoon shoot came out nice, but the night show still has to much red in it.
setting:
nd filter off,
3200k
-3 gain
1/60
2.4

where is my mistake?????

thanks

karl