Blown highlights on Leaf back

[gsking]

The specular highlights on my Valeo 22 seem to blow out in a quite unappealing fashion that I never noticed on my DSLR. Is this to be expected? It's like shooting slide film, only worse.

I realize that digital in general doesn't provide the highlight latitude that print film does, but I didn't expect an MFDB to be more intolerant than a DSLR. It would seem the only answer would be to ETTL.

Or, the other possibility is that Leaf software handles the .mos files in a significantly better manner than LR3?

I am hesitant to use the MFDB outdoors, because unless I ensure I underexpose to retain the sky, I'll likely get ugly banding.

Thoughts? Known equipment limitation, or operator error?

 

[Graham Mitchell]

Does the highlight recovery tool in LR3 not help ?

 

[johnnygoesdigital]

Sometimes exposure meters need to be calibrated. Just for fun, use the zone exposure method to try to determine if your meter is off. Using the spot setting, find a scene and determine which area is zone 5 (18% gray), take a meter reading. If you don't have a gray card, usually, grass is close. Take a meter reading on a bright sky or other area that might be blown, and increase the exposure 2 stops from your first reading. This corresponds to + 2 stops from middle gray (18%) Highlighted areas usually aren't placed higher than zone 7, in digital cameras. Check your histogram, and look for any blown highlights.

 

[dougpeterson]

Several people here would love to help you but without
- an image illustrating what you mean we can only guess at what you mean
- details on where and how you are processing your image (specific version numbers and settings)

Uploading a raw file to something like yousendit for a few of us to download and play with would also be helpful.

 

[gsking]

Does the highlight recovery tool in LR3 not help ?

It doesn't help as much as it does "normally", but I suspect that's because when all three channels are blown, there's nothing to recover

 

[gsking]

Sometimes exposure meters need to be calibrated. Just for fun, use the zone exposure method to try to determine if your meter is off. Using the spot setting, find a scene and determine which area is zone 5 (18% gray), take a meter reading. If you don't have a gray card, usually, grass is close. Take a meter reading on a bright sky or other area that might be blown, and increase the exposure 2 stops from your first reading. This corresponds to + 2 stops from middle gray (18%) Highlighted areas usually aren't placed higher than zone 7, in digital cameras. Check your histogram, and look for any blown highlights.

Johnny,

It's not the auto metering per se, because I'm just talking about the nature of the blown highlights. Regardless of what the meter says, a "properly exposed" image seems to need to be exposed to the left to avoid blowing any highlights.

Subjectively, I don't have to try to avoid this as much with my DSLR, but it may be due to the JPG engine, or the RAW engine, or the Dynamic Range Optimization, or the converter in LR3.

So, generically, the question was if other MFDB users had seen a similar phenomenon. I suspect it's the nature of the beast, but wanted to make sure it wasn't just LR3 being lazy with the processing.

All this talk of increased dynamic range makes me think I can be sloppier with exposure, but if that means I have to be careful to underexpose to avoid ANY clipping, so be it.

Does that clarify?

 

[rga]

You mention specular highlights; are you using a polarizing filter to try to reduce them?

I have never been able to control true specular highlights: they are essentially direct sun reflections... About all you can do to reduce them is to use a pol filter.

As for dynamic range recording ability of film (positive) vs 35mm digital vs MFDB, I have found that most of the increase in dynamic range in MF comes from the recovery of shadow detail. There is a bit greater capture ability with MFDB on the right side, but most is on the left. Though I try to move my histogram data as far to the right as possible, I try never to clip the highlights. When in doubt I'll make several exposures starting as far right as possible and then decreasing the exposure for 2 or 3 stops.

I use Capture 1 for MFDB processing primarily because of LCC. I have been told its RAW processor is better at drawing out information than LR and I have no reason to doubt that it is at least as good (I continue to use LR for cataloging).

Hope this helps a bit,
Bob

 

[gsking]

Several people here would love to help you but without
- an image illustrating what you mean we can only guess at what you mean
- details on where and how you are processing your image (specific version numbers and settings)

Uploading a raw file to something like yousendit for a few of us to download and play with would also be helpful.

Doug,

I did mention LR3, but I probably could have spelled it out.

Here's an example specifically of the highlights, but my general comments about exposing sky is related.

In fact, in a related issue, anything with "pure" color seems to blow a channel much quicker than I would expect. Here's an example where the histogram doesn't show anything blown, but the detail in the bright yellow areas is gone.

It seems as if the CFA has extra dense coloring or something, so it enhances the saturation so much that it's hard to take pictures of anything vivid without it oversaturating.

Hope that helps you help me.

 

[gsking]

You mention specular highlights; are you using a polarizing filter to try to reduce them?

I have never been able to control true specular highlights: they are essentially direct sun reflections... About all you can do to reduce them is to use a pol filter.

As for dynamic range recording ability of film (positive) vs 35mm digital vs MFDB, I have found that most of the increase in dynamic range in MF comes from the recovery of shadow detail. There is a bit greater capture ability with MFDB on the right side, but most is on the left. Though I try to move my histogram data as far to the right as possible, I try never to clip the highlights. When in doubt I'll make several exposures starting as far right as possible and then decreasing the exposure for 2 or 3 stops.

I use Capture 1 for MFDB processing primarily because of LCC. I have been told its RAW processor is better at drawing out information than LR and I have no reason to doubt that it is at least as good (I continue to use LR for cataloging).

Hope this helps a bit,
Bob

Bob,

That helps. You've confirmed my suspicions at least. Do they end up looking jagged as in the examples above?

I never got used to C1 when I had a Phase back, and the last program I used from Leaf crashed my computer. Maybe I need to try it again.

And thanks for reminding me about Cpols. That would definitely help with outdoor shots in general. Duh.

 

[Felipe]

I found that setting the curves in LR3 to linear helps a lot when working with the files from my digital back.

 

[Thierry]

This first image is showing "blooming" effect, an effect which was typical for older generations of sensors. It happens when the pixels are over-saturated with light photons and then filling up the neighbour pixels.
The manufacturers did overcome this issue with so-called "anti-blooming" built-in the electronics of the back to avoid or reduce blooming.

I am not sure if the Valeo did already have "anti-blooming" built-in.

Thierry

Doug,

 

[johnnygoesdigital]

gsking,

Yes, got it. As mentioned above I agree that's it's possible the photons are spilling over to other sites. I'm not familiar with your DB, but if you have options over the DRO, than I would dial that back a bit.

 

[rga]

Bob,

That helps. You've confirmed my suspicions at least. Do they end up looking jagged as in the examples above?

I never got used to C1 when I had a Phase back, and the last program I used from Leaf crashed my computer. Maybe I need to try it again.

And thanks for reminding me about Cpols. That would definitely help with outdoor shots in general. Duh.

Personally, I think this scene (from the black gloves to the direct reflection of the light) is beyond the capabilities of any single exposure using any recording medium available to consumers.

So we'll have to give on something; either lose the details of the black gloves and control the star shape of the reflected light, or get the blooming and details of the black gloves. My guess is that the range of this shot is way beyond 20 stops....

It would be fun to try an HDR image on this: not overly cooked, but naturally blended...

As for the flower, it definitely looks over exposed. Go out and play with a polarizing filter (it doesn't need to be circular unless your rig is autofocus) and then watch your histogram. Record it so that the highlights are just touching on the right side. Then try C1 and bring up the shadows (use the HDR sliders to protect the highlights and bring in the shadows). You may be really amazed at how much data can be recovered from the right side; I always am. You should also take a shot where the highlights are just barely clipped on the right and see if that works too.

Good luck and let us know your results!
Bob

 

[gsking]

Personally, I think this scene (from the black gloves to the direct reflection of the light) is beyond the capabilities of any single exposure using any recording medium available to consumers.

So we'll have to give on something; either lose the details of the black gloves and control the star shape of the reflected light, or get the blooming and details of the black gloves. My guess is that the range of this shot is way beyond 20 stops....

It would be fun to try an HDR image on this: not overly cooked, but naturally blended...

As for the flower, it definitely looks over exposed. Go out and play with a polarizing filter (it doesn't need to be circular unless your rig is autofocus) and then watch your histogram. Record it so that the highlights are just touching on the right side. Then try C1 and bring up the shadows (use the HDR sliders to protect the highlights and bring in the shadows). You may be really amazed at how much data can be recovered from the right side; I always am. You should also take a shot where the highlights are just barely clipped on the right and see if that works too.

Good luck and let us know your results!
Bob

Thanks for the input. I think I'm just spoiled by some of the "C-41-like" dynamic range compression from my A700. As you all noted, you can't overcome the physics of photosite saturation, and it just shows up more significantly with a minimally processed MFDB output.

As an aside, there was nothing magical about that shot. I was just testing the focusing accuracy of my RZ67.

Ironically, my next shoot is studio shots indoors. At ISO 25, I doubt my weak strobes (300ws and 160ws) will be even able to CREATE blown highlights. So I should be safe.

And I'll put my 58mm polarizer on top of my desk.

Much appreciated for everyone's comments.
Greg

 

[Geoff]

I've found some pretty good lattitude in the backs, even better perhaps than film.

Here are some shots from a Leaf AFI 7 II back. In bright light, the white cloth still has detail in the sunlight, and there is information still in the shadows. Grainy, but no noise reduction has been applied. Pretty much straight C1 on a shot with ISO at 200. Crops are supposedly 100% if I got this right....

 

[gsking]

Geoff,

Interesting. Those look very flat to me, very much like C-41, or more. The sky not only looks mid-tone, but it looks more grey than a "saturated" blue.

Maybe mine would look the same if I was more deliberate to avoid overexposure, but I usually still end up with over-amped colors like the one in the flower. I need to turn down the saturation somewhere.

Hmm...I guess I'll have to compare apples to apples and see how my camera compares to my DSLR in identical exposures.

 

[Geoff]

Leaf has a bunch of different profile settings for the backs. I don't know which ones you have, but there is a big difference in importing between portrait and product, with portrait being softer and less harsh than product. For delicate scenes, portrait is preferred. For more sharply defined and contrasty results, product.

These do look awfully flat, moreso than they did in C1. I brought them through Aperture and into jpgs, maybe something got lost along the way. There is a lot of flexibility to be found typically in the Leaf files, IMHO tho. They are quite elastic.

 

[Shelby Lewis]

I love starting with my leaf files in a very flat state... it always seems as though I can push them in so many directions with the amount of latitude offered by the subtle gradations of tone offered by a flat import.

It's totally a personal preference... but it makes dealing with highlights and shadow considerations a bit more easy for me, personally.

 

[gsking]

I tried all the settings. Yes, product was even worse. I use only portrait now. No HS settings here.

I WISH I could get my files to start that flat. I just can't figure out what the trick is. That's the way they come out...oversaturated.

Greg

 

[rga]

I tried all the settings. Yes, product was even worse.
I WISH I could get my files to start that flat. I just can't figure out what the trick is. That's the way they come out...oversaturated.
Greg

Buy a Phase One?
:ROTFL:
(sorry; couldn't resist...)