I've uploaded 12 clips to dvinfo.net site

[Paul Mogg]

Thanks, I wouldn't say sepia, but I think the beach scene is a bit dull, mainly because it was a bit of a hazy day, they could all do with some color correction here and there. The reds and blues definately have some chroma noise, but the main thing is that on a big HD screen they have so much more detail present than DV, which is why I like the format.

[Steve Mullen]

<<<-- Originally posted by Charles Papert :
Expose for the highlights, keep the blacks from crushing and stretch it back down in color correction. -->>>

Good plan but.

You can bias the exposure down. But in a high contrast situation, to avoid blown highlights--you've got to cut OVERALL exposure by at least 2 stops. Possiibly 3 stops.

Now, what happens to mid-tones with 2 to 3 stops under-exposure?

And, what happens to shadow detail?

In short, there is no way to keep the blacks from being crushed because that's the side effect of dropping exposure to save the bright areas.

That's because the chip has very little lattitude.

So Charles is absolutly right. The shots you want are damn near impossible to shoot with the HD10.

[Charles Papert]

It's a toughie.

I'm not experienced with the JVC camera specifically, I've just seen the stills and can get a sense of the exposure curve. It does seem steeper at the top than most cameras in this price range, resolution notwithstanding.

I'm always impressed with the underexposure latitude of digital video, from DV to HD however. It's arguable that at its best, digital can capture more in the shadows than film (each new generation of film stock is upping the ante on this characteristic, however). This is one of the reasons that Michael Mann elected to use the Cinealta to shoot the opening scenes of "Ali" where Will Smith is jogging through the streets at night; the HD camera was able to capture the clouds against the night sky, which would have been more prone to blocking up on film.

Also, it's well known that dropping the gamma will give video a more film-like appearance, whether in-camera or in post. It's a bit simplistic to put it this way, but in essence it's like underexposing just the midtones. I've always felt that the old standard of 70 IRE for skin tones is way too hot looking--very "evening news". I usually aim for no more than 50 for normal exposure.

So, in a situation where the overall exposure is reduced by a couple of stops to keep the highlights from blowing out, having the midtones drop with it is not such a bad thing for me--I think it results in richer skin tones and a more saturated look overall. The issue, as Steve points out, is that you will be giving up the bottom of the curve, in that details will crush in the shadows. This is why I elect to use an Ultracon in these high-contrast exterior situations to artifically raise the toe, which helps retain some of this information which would be lost. The effect is not linear, i.e. it's not the same as just raising the setup and then dropping it again in post, so that you do generally end up with a different degree of information in the shadows than if you had foregone the filter.

Drawbacks of this method? You will add some noise from having to boost levels in post, but I don't consider this objectionable--it adds some texture. One could even say it feels like film grain. As I have said before, using Ultracons takes a lot of care and judgement, they are not a set-and-forget filter. You can really screw up your photography if you use the wrong strength for a given shot, or allow light to hit the filter, and end up with a tragically mushy image.

I should also point out that I only go this route for exteriors. On interiors, I will light for the contrast ratio of the camera, and knock down the windows to the desired level. I don't doubt it would be possible to make some beautiful pix with this camera under completely controlled circumstances.

This is not a textbook nor a rocket-science approach; it's just what I do and what works for me. Taming the beast of high-contrast exteriors is the single greatest nightmare of digital shooting, in my mind and in a few years, this may all be moot, but in the meantime it takes a bit of voodoo and hocus pocus to "make the pretty".

[Steve Mullen]

<<<-- Originally posted by Charles Papert
So, in a situation where the overall exposure is reduced by a couple of stops to keep the highlights from blowing out, having the midtones drop with it is not such a bad thing for me--I think it results in richer skin tones and a more saturated look overall. -->>>

I too used to shoot 35mm stills 1.5 stops under cause I liked the more saturated color. But my tests with the JVC says to avoid blown whites you may have to go 3-5 stops under. Now white dog fur has details. Which is great! But human skin is now deep brown. And shadow detail is fully BLACK.


I'm also sketpical of ND filters preventing blown highlights. True they reduce light by 2 or 3 stops, which reduces AVERAGE exposure, the RANGE -- in stops -- remains just as great. Since latitude refers to Range, nothing has changed.

Comments?

[Eric Bilodeau]

Theoretically ND filters should not reduce the range (and it indeed doesn't) but in the case of blown out whites, the thing is that a lot of the bright areas are seen as complete white because of the overload from the CCD so all this range is lost by this wash out. If you reduce the incomming light in the lenses, the CCD will receive less light, the contrast in itself will not change, it is just an equivalent of under exposing a little. But a fact remains, by avoiding the overload of the CCD, you get the range in the highlights witch would have been lost (or washed as a white out area) because of the over sensibility of the CCD. So ND filters will overall avoid too much light to enter by the lenses to the CCD and burn the image. CCD's are vey sensitive to direct light. A good way of noticing that is by using a device like the mini35: as the camera shoots a depolish witch has received the image from other lenses (or if you prefer, the light does not enter directly by the camera's lenses) you lose all the edge artefacts that are normally occuring by the CCD receiving too much light. It is very revealing indeed.

[Kevin A. Sturges]

"the camera shoots a depolish witch "

Did we lose you there for a second, Eric, or are you using voice to text software? :)

[Eric Bilodeau]

Sorry 'bout that, the camera shoots a screen, the size of a 35mm slide, a depolished screen witch receives the image from the 35mm lenses up front. Sometimes (as I usually speak french) I tend to say things in a french way even in english witch is not always the most accurate way... :)

Anyways, I hope this clears up the mistake !

[Charles Papert]

"depolished screen" = "ground glass", oui?

Agreed that ND filters would only be useful if the aperture of the lens is essentially maxed out (which should be one stop open from the minimum aperture, for best results...i.e., if your lens stops down to f22, consider f16 your minimum aperture.)

Does the JVC camera have a "negative gain" function like the XL1?

Sad to hear that these cameras require as much stopping down as Steve reports to avoid blown highlights. That's a deal-killer to me.

[Eric Bilodeau]

Nope, no negative gain that I know of. I would be surprised to see a camera on witch you cannot control the iris and shutter at the same time having such an "evolved" function...

You dont have to close as much as f16. The automatic controls have a tendency to be a bit too opened so using a ND is useful once you locked the exposure
.

[Steve Mullen]

<<<-- Originally posted by Eric Bilodeau : Theoretically ND filters should not reduce the range (and it indeed doesn't) but in the case of blown out whites, the thing is that a lot of the bright areas are seen as complete white because of the overload from the CCD so all this range is lost by this wash out. If you reduce the incomming light in the lenses, the CCD will receive less light, the contrast in itself will not change, it is just an equivalent of under exposing a little. -->>>

That's why I asked for comments. There are three potential sources of problems:

1. Too much light on the CCD

2. Too much range on the CCD

3. Too much range for the DSP (only 10bits on the JVC).

I used a .6ND and still had blown highlights. Perhaps Paul used .9ND? Paul?

There are even stronger ND filters. But .9ND is 3 stops. What is 1/64?

[Charles Papert]

I hope I get the math right, but I've seen the designation of ND expressed on the filter wheels of broadcast cameras as fractions, representing the amount of light transmitted; thus an ND.3 would be a 1/2, ND.6 a 1/4, and so on...by this reasoning, a six stop ND filter (the rather unwieldy sounding ND1.8, more likely two ND.9's stacked) would be a 1/64.

[Jay Nemeth]

You've got it right, Charles. I don't know why television has to be different, I have also seen some ND filters labeled 2x or 8x (.3 and .9). The system of labeling ND filters is older than television itself, why reinvent the wheel?

Jay

[Steve Mullen]

Neutral Density factors:

BW101 (ND.3) (exposure adjustment = 1 stop)

BW102 (ND.6) (exposure adjustment = 2 stops)

BW103 (ND.9) (exposure adjustment = 3 stops)

BW106 (ND1.8) (exposure adjustment = 6 stops)

BW110 (ND3.0) (exposure adjustment = 10 stops)

BW113 (ND4.0) (exposure adjustment = 13 stops)

BW120 (ND6.0) (exposure adjustment = 20 stops)

I just tested a Sony 390L which used a 1/64th filter. As I remember, that's true of the other bradcast cameras I've reviewed.

Which means I/we have been using way too little filtering! (Paul???)

[Paul Mogg]

Again, I think we need to do some studio tests with various lighting setups and ND filters to see what is really required by this camera under various circumstances to get the best results. I'm still waiting to get a day free to be able to do this. I hope someone else has more free time than me and can do this soon.
Yesterday I shot another dance show and set up the JVC unattended alongside my Ike to see what I would get. I shot the first half with a .6 ND filter (2 stops) at 1/30th, and the second half with no ND filter at 1/60th. There was a lot of fast motion in the dancing. If I can get permission to do so, I'd like to post a couple of clips from this to show you. Overall the more pleasing results to me were at 1/60th with no ND filter. Less noise in the mid blues, nice filmic motion and not excessive blur, but it just totally depends on the lighting. This time the whites in the lighting were not screaming as much as in my previous shoots, and I got some very nice results on occasion. As I've said before, if you have control over the lighting you're fine. If not, you've got serious problems, and I wouldn't reccomend it.

[Alan Dunkel]

Paul, I have a Beaulieu 6008 Super8 film camera and some Kodak Vision Negative stock ( 320T ) if you'd like to run a side by side test I'll loan it to you and have the resulting film converted to miniDV or Digital Beta. It even has crystal sync if you want to get out the old clapper and record sound to DAT or miniDisc.
Regards, Alan

PS. I do think the latitude issues on the JVC can be controlled even if it requires gelling the entire car windshield/windows, lighting from the hood as it is towed etc. A shade is already generally needed over the hood area for some shots ( ie. into the car from the front ) to control reflection regardless of HDV or film is being used.