Dynamic range and highlight recovery... thought I would lose it...

[FredBGG]

Dynamic range and highlight recovery is often touted as one of the big benefits of MFD often exaggerated by vendors.

I recently decided to ditch MFD in favor of a combination of medium and large format film along with a high end 35mm DSLR.

Having seen some nice files I bought a D800.

Anyway I live in southern California and often shoot on location in the desert or at the beaches... I favor the beaches as I can go kitesurfing as soon as I finish shooting

I often shoot with the model under flags ... sort of an arch. Sometimes I add a key light or reflector.

With this setup and shooting full length I often need to recover background detail as it tends to be over exposed.

So I did a quick highlight recovery test and it exceeded my expectations.



It is going to be a real pleasure to shoot on location with this camera, not to mention the particular look of the 85mm 1.4G.

I thought I would be losing highlight recovery by ditching MFD, but that simply is not the case.
It may even be better than what I have been able to do with Phase One P series backs.

it's a bit hard to see with the size I posted, but there is still tonality in the white of the numbers on the letter box....

[ondebanks]

Dynamic range and highlight recovery is often touted as one of the big benefits of MFD often exaggerated by vendors.

I thought I would be losing highlight recovery by ditching MFD, but that simply is not the case.
It may even be better than what I have been able to do with Phase One P series backs.

I'm not at all surprised, Fred.

Anyone can see the results for themselves in DxOmark...

Ray

[yaya]

Perhaps worth moving this thread to the 35mm forum?

[kers]

Hello Ray,

I am a bit surprise, yes.  I am used that overexposed images are usually quickly gone in the overexposed areas. Adobe raw still could do a good job sometimes finding the hidden information.
The D800 seems to have more headroom than the d3x in this respect. Looking at the small overexposed areas in this case it is fortunate that the image looks at first glance more overexposed than it really is.

[MrSmith]

Perhaps worth moving this thread to the 35mm forum?

why? each post has MFD info and might be handy for those wavering between a D800 and 40mpixel back and looking for user info.

[yaya]

Because the OP shows one image that was taken with a 35mm camera and discusses its qualities

But that's just MY opinion...

[FredBGG]

I think this is relevant to this forum, but let me elaborate.

First of all e really like the look of medium format, but find that unique look
more in larger medium format using film. While MFD has some of this look it is
actually quite close to the best 35mm system lenses. 6x8 film is 4 times the size of FF 35mm.

When I shoot film once I have the shot I shoot some backup with digital just for insurance purposes.
Due to the fact that my lighting setups are generally for film and mostly black and white
I need to have a digital camera with good dynamic range so I don't have to change scrims, fill or what not.
This makes the D800 an essential part of my medium format kit.

Secondly many people think of medium format for it's dynamic range and now it is just as good with
the D800, if not better. Also I think it would be safe to say that the D800 has some merit to be
discussed along side MFDB as it has a higher MP count than entry level MF thus making the d800 relevant to "large sensor photography".

Going back to the larger format look I find that with a combination of film and digital 35mm DSLR I am getting the best of both worlds.
While the D800 pretty much as good MFD with the exception of the 60 and 80MP the lens options for the D800
give me a vast range of looks. Nothing offered by Mamiya Phase gives me what the 85mm 1.4G or the 200 f2G give me.
Shallow depth of field is something I really like to use. The Fuji GX680 180mm 3.2 is IMO unmatched by anything from MFD manufacturers.
Shallow depth of field and tilt shift on a 6x8 capture area. I also find it to unreliable to shoot shallow depth of field with MFD.
Low viewfinder magnification and very limited focusing system as well as slow live view.

[bjanes]

The D800 seems to have more headroom than the d3x in this respect. Looking at the small overexposed areas in this case it is fortunate that the image looks at first glance more overexposed than it really is.

Before concluding that the image is overexposed, one should look at the tone curve applied by the camera or raw converter. Adobe Camera Raw and Lightroom use a hot tone curve for the Nikon D800. The actual exposure depends on the calibration of the light meter.

The Radigger web site has a good article on the calibration of the camera light meter. I performed the suggested test with my D800e and determined that the headroom was 2.75 stops from mid gray. This was for ISO 100 and a color temp of 4500 (late afternoon sunlight). Saturation of the sensor was 14.8%, whereas mid gray is 18%. The pixel values for a 2.2 gamma space would be 107 and 117 respectively [%saturation^(1/2.2)*255)]. ACR 7.1 rendered the 14.8% saturation at 173 in Adobe RGB.

Camera makers adjust headroom by changing the ISO to a lower value than would be indicated by the ISO standard. DXO measures the actual ISO. For the D800 at an ISO setting of 100, the measured ISO is 73. This means that the camera will give 0.45 stops less exposure than would be given by the ISO standard. This would place the highlights 1/2 stop below clipping. For the Phase One IQ180 at an ISO setting of 100, the measured ISO is 29, which would place the highlights 1.8 stops below clipping.

Regards,

Bill

[ErikKaffehr]

Hi,

The original poster said he moved from MF digital to 24x36 digital and large format film, so I would say that his experience absolutely relevant in this context.

Best regards
Erik

Perhaps worth moving this thread to the 35mm forum?

[LKaven]

I'm not sure I understand why the word "recovery" is in this discussion.  Either one (or more) of the channels was blown, or it wasn't.  Knowing when your camera manufacturer is playing coy with your perceptions is key.

Bill is correct that "standard" tone curves will exaggerate the highlights.  You should be using UniWB to get as accurate a histogram as possible, without applying any tone curves whatsoever, in order to truly judge the exposure.

[FredBGG]

I'm not sure I understand why the word "recovery" is in this discussion.  Either one (or more) of the channels was blown, or it wasn't.  Knowing when your camera manufacturer is playing coy with your perceptions is key.

Bill is correct that "standard" tone curves will exaggerate the highlights.  You should be using UniWB to get as accurate a histogram as possible, without applying any tone curves whatsoever, in order to truly judge the exposure.

I am talking about recovery is the sense of using recovery slider in raw or shadow highlights or a combination of other tools.
IF a channel is blown out beyond the very top of recordable dynamic range it's not recovery it's inventing it.

The point I am making is that recovering extremely bright highlights with this camera returns excellent results due to the linearity between the three channels.
With mode cameras you get color shifts and strange looking tones as if you had thrown your curves correction around too much or to many times.

After making these corrections I did the detail and colors were remarkably close to those of a correctly exposed image, even the detail and texture within the white numbers.

[LKaven]

...recovering extremely bright highlights with this camera returns excellent results due to the linearity between the three channels.
With mode cameras you get color shifts and strange looking tones as if you had thrown your curves correction around too much or to many times.

I would think that using "camera neutral" (in LR) or "linear" (in C1) would return linear data for all the Nikon cameras.  It seems there's more to what you're saying, but I'm not sure what it is.

One thing is for sure.  The highlight adjustments in LR4 are much more "photographic" than on previous versions, which were all but unusable.

[kers]

...... For the D800 at an ISO setting of 100, the measured ISO is 73. This means that the camera will give 0.45 stops less exposure than would be given by the ISO standard. .....
Regards,
Bill

Bill,   thank you for pointing this out..in accordance i would think

1  the optimum exposure for the d800e is perhaps about 60asa..  optimum in the sense of getting the tonality at its best and noise the lowest...

2 D3x did not have this headroom for it performed a lot weaker at high ISO and Nikon did not want it to be a 100-800 asa camera but an 100-1600 asa camera....

please correct me where you think i go wrong...

[Gel]

I wouldn't exactly call that a great test image. Nothing was really blown, just overexposed.

The 5D3 has better recovery than that. The D800 has great range in the shadows though.

[jduncan]

Dynamic range and highlight recovery is often touted as one of the big benefits of MFD often exaggerated by vendors.

I recently decided to ditch MFD in favor of a combination of medium and large format film along with a high end 35mm DSLR.

Having seen some nice files I bought a D800.

Anyway I live in southern California and often shoot on location in the desert or at the beaches... I favor the beaches as I can go kitesurfing as soon as I finish shooting

I often shoot with the model under flags ... sort of an arch. Sometimes I add a key light or reflector.

With this setup and shooting full length I often need to recover background detail as it tends to be over exposed.

So I did a quick highlight recovery test and it exceeded my expectations.



It is going to be a real pleasure to shoot on location with this camera, not to mention the particular look of the 85mm 1.4G.

I thought I would be losing highlight recovery by ditching MFD, but that simply is not the case.
It may even be better than what I have been able to do with Phase One P series backs.

it's a bit hard to see with the size I posted, but there is still tonality in the white of the numbers on the letter box....

Thanks for the example.

Is my understanding that, as general rule,  the last generation H4Ds still have an edge  in recovering highlights against the D800, but the D800 owns the shadows.

I mention that in part because this cameras are so good that, some times, at the crop level, the only way to know is side by side comparison.  I


http://www.youtube.com/watch?v=9UBTE4xpvpk

Best regards,

James

[EricWHiss]

It's a ncie example, Fred, but I think the same can be done with most of the newer MFDB.  Certainly that's my thought anyhow and having mostly used MFDB's I wasn't surprised by how far you could pull down the highlights because this seems pretty normal.

[torger]

All this talk about "highlight recovery" seems to be because Lightroom etc hides how a digital camera actually works. A sensor registers light linearly and either it clips or not. Typically the R G B channels do not clip simultaneously and then a raw converter can guess the clipped channels values, more or less successfully, I'd call that "highlight reconstruction" though, since the raw converter invents data that has not been recorded. It differs greatly between raw converters how good they are at guessing clipped data.

I often hear people say that MFDB has great "highlight recovery". Sure, it is just how the camera reports the sensitivity and histogram shows the data, if it underexposes the images then there's more "highlight recovery", but it only means that you get noisier shadows since the image is underexposed. Do a couple of experiments to learn how your camera's histogram and ISO works. If you think it has too little "highlight recovery", just underexpose. I know those that have their exposure correction dial constantly at -1 to get more highlight recovery when shooting with auto-exposure.

[FredBGG]

It's a ncie example, Fred, but I think the same can be done with most of the newer MFDB.  Certainly that's my thought anyhow and having mostly used MFDB's I wasn't surprised by how far you could pull down the highlights because this seems pretty normal.

My point is that in ditching MFD I would lose this kind of control or would at least have to do color correction to the recovered areas
What I am saying is that it is as good if not better. If I consider the better shadows and overall better dynamic range I'm better of now
from an IQ point of view. If on top of that I add all the ergonomics, focusing, range of lenses, better bokeh, speedlight integration, image stabilization,
reliability, speed, proper live view, video, NPS... well I think it's a pretty clear picture.
While I was pretty confident of my decision and this highlight recovery/control is the icing on the cake.

[Guillermo Luijk]

the last generation H4Ds still have an edge  in recovering highlights against the D800, but the D800 owns the shadows.

Talking about 'recovering' highlights vs 'recovering' shadows is arbitrary, the only thing that matters is total dynamic range which includes both.

A camera that is good at 'recovering' highlights (the H4Ds according to you) is a camera with a low effective ISO, and becomes bad at recovering highlights as soon as we expose more. For the same reason a camera that is bad at recovering highlights is a camera with a high effective ISO, and becomes a highlight champion just by systematically underexposing every shot.

If you are a RAW shooter, the only thing that matters is total DR of your sensor, i.e. how many stops below clipped highlights you can still have information with a decent low visible noise.

Regards

[Schewe]

If you are a RAW shooter, the only thing that matters is total DR of your sensor, i.e. how many stops below clipped highlights you can still have information with a decent low visible noise.

And how well the raw converter can extract usable image data from the raw capture...LR4 and ACR 7 raised the bar...