MFD Highlight DR…

Started Nov 25, 2016 | Discussions
Erik Kaffehr
Erik Kaffehr Senior Member • Posts: 2,772
MFD Highlight DR…
3

Hi,

One statement often heard is that MFD has a lot of highlight DR. That does not make a lot of sense as there is nothing like highlight DR. DR is the ratio of highest unclipped signal to the lowest acceptable signal.

There is an excellent tool called RawDigger for analysing raw files. Here is a pretty decently exposed raw file from my P45+.

P45+ image shown in RawDigger

The sun is almost included in the image, and there is a tiny part that is burned out. So, let us no open this image in Capture One and white balance on the grey marble on the building at the right.

Capture One with white balance set, otherwise defaults.

A significant part of the sky is overexposed, but we can of course recover it.

Now, let's open the same raw image in Lightroom:

Lightroom with white balance set, otherwise deafults.

So, Capture One processes the image much brighter at default setting. Lightroom automatically recovers the bright parts around the sun.

Now, let us go back to Capture One and switch to linear response from the default "auto".

Capture One, linear response that is no "film curve"

We can "recover" the sky with the "highlights" slider, but in reality we just show the part that Capture One has thrown away.

Capture One after highlight recovery…

My skills with Capture One are not so great, but this is about what I get:

With Lightroom I have more experience:

So, what I would say I have demonstrated is that default processing in Capture One makes the images to bright. The histogram may also be a bit conservative giving overexposure indications to early.

A similar shot from the A7rII looks like this:

A7rII image from Lightroom

This is a "photographic" rendering of a an image with excessive highlights shot on the A99

The disc of the sun is clearly visible trough the clouds

But we can still have good detail in the darks

Good detail in the darks

The dynamic range here seems to be around 9-10 EV (except the sun). It is limited by the large area of water reflecting skylight and probably also by lens flare.

Raw histogram

This image was shot at Lockenhous in Austria on the P45+:

P45+

And this one on the Sony A99 at the same place:

Sony Alpha 99

Here is the piano from the P45+ image:

P45+

And here is the piano from the Alpha 99

Alpha 99

I would say that the P45+ and the Sony exposures protect highlights similarly but the Sony Alpha 99 has cleaner signal on the near black areas.

This is what DxO-mark says about the two cameras.

Normalised DR on the P45+ and the Sony Alpha 99 according to DxO-mark.

MFD-s with modern CMOS sensors would have an advantage of size over smaller formats:

http://photonstophotos.net/Charts/PDR.htm#Olympus%20OM-D%20E-M5%20Mark%20II,Pentax%20645Z,Phase%20One%20IQ260,Sony%20ILCE-6000,Sony%20ILCE-7RII

This nice graph from Bill Claff's website shows how DR vs ISO improves with larger formats. Note the improvement of the Pentax 645Z over the full frame 645 size Phase One IQ-260. That is CMOS vs. CCD.

Pentax 645Z Sony Alpha a99
If you believe there are incorrect tags, please send us this post using our feedback form.
The Davinator
The Davinator Forum Pro • Posts: 22,041
Re: MFD Highlight DR…

There is no such thing as highlight DR. DR ks the whole range.  I too find the statement odd.  Latitude would be better...at least over mid gray

 The Davinator's gear list:The Davinator's gear list
Canon EOS D30 Canon EOS 10D Nikon D2X Fujifilm X-Pro1 Fujifilm X-T1 +17 more
marcio_napoli Senior Member • Posts: 1,454
Re: MFD Highlight DR…

Hi Erik,

I'm not trying to question your and David's opinions, just would like to present a different thought.

I'm far from a strict techie guy, as I mostly trust what my eyes see (regarding photography, I mean), so sorry in advance for my ignorance in case my presented method is super flawed.

What I usually do to test highlight headroom:

I choose a bright source of light that contains details, most of the time that's the sky with some clouds.

I over exposure by at least 2 EVs, and shoot with corresponding settings between cameras, like D800 , ISO 100 , SS 1/500 and whatever B camera, ISO 50, SS 1/250, same f stop.

Each sensor is saturated with the same exact amount of light.

As my sky (bright light) is already overexposed, when I dial down exposure in post, it's possible to see how many stops each sensor can recover while still keeping color information.

This simple test has shown many different performances, which I understand to be the highlight headroom of each sensor.

And contrary to popular belief, DBs are not great at that.

Excuse my super flawed eye method :D, but my experience went something close to the following (from what I can remember).

I could recover what appears to be a full blown sky (all at base ISO) by:

3.5 stops - Fuji S3

3.0 stops - Phase One P40+

2.3 stops - Nikon D700

1.5 stops - Leaf Aptus II 5

1.5 stops - Nikon D800

1.5 stops - H3D-31

1.3 stops - Nikon D200

0.5 stop - Phase One P45+ (not kidding! at least in my experience.)

0.5 stop - Nikon D1X

I tested these under this method, as soon as I got my hands on those cameras.

Don't take the numbers too seriously, they're not scientific.

But regardless of accurate or not, these are real world findings.

What one can take quite seriously is that the Fuji S3 really was a marvel, the P40+ is the only DB I've used that does justice to the "DMF super highlight headroom" fame, and that surprisingly the D800 is worse than the D700 on that criteria.

But how did DBs obtain such fame regarding highlights, I really don't know (and that is coming from a self admited DMF fanboy).

Anyway, what I think we cannot claim is that highlight headroom doesn't change between cameras. At least IMO, it does, but I could be saying loads of baloney

What do you guys think?

Based on eye alone, what's your experience with the cameras you've used?

Best regards,

Marcio Napoli

Erik Kaffehr
OP Erik Kaffehr Senior Member • Posts: 2,772
Re: MFD Highlight DR…

marcio_napoli wrote:

Hi Erik,

I'm not trying to question your and David's opinions, just would like to present a different thought.

I'm far from a strict techie guy, as I mostly trust what my eyes see (regarding photography, I mean), so sorry in advance for my ignorance in case my presented method is super flawed.

What I usually do to test highlight headroom:

I choose a bright source of light that contains details, most of the time that's the sky with some clouds.

I over exposure by at least 2 EVs, and shoot with corresponding settings between cameras, like D800 , ISO 100 , SS 1/500 and whatever B camera, ISO 50, SS 1/250, same f stop.

Each sensor is saturated with the same exact amount of light.

As my sky (bright light) is already overexposed, when I dial down exposure in post, it's possible to see how many stops each sensor can recover while still keeping color information.

This simple test has shown many different performances, which I understand to be the highlight headroom of each sensor.

And contrary to popular belief, DBs are not great at that.

Excuse my super flawed eye method :D, but my experience went something close to the following (from what I can remember).

I could recover what appears to be a full blown sky (all at base ISO) by:

3.5 stops - Fuji S3

3.0 stops - Phase One P40+

2.3 stops - Nikon D700

1.5 stops - Leaf Aptus II 5

1.5 stops - Nikon D800

1.5 stops - H3D-31

1.3 stops - Nikon D200

0.5 stop - Phase One P45+ (not kidding! at least in my experience.)

0.5 stop - Nikon D1X

I tested these under this method, as soon as I got my hands on those cameras.

Don't take the numbers too seriously, they're not scientific.

But regardless of accurate or not, these are real world findings.

What one can take quite seriously is that the Fuji S3 really was a marvel, the P40+ is the only DB I've used that does justice to the "DMF super highlight headroom" fame, and that surprisingly the D800 is worse than the D700 on that criteria.

But how did DBs obtain such fame regarding highlights, I really don't know (and that is coming from a self admited DMF fanboy).

Anyway, what I think we cannot claim is that highlight headroom doesn't change between cameras. At least IMO, it does, but I could be saying loads of baloney

What do you guys think?

Based on eye alone, what's your experience with the cameras you've used?

Best regards,

Marcio Napoli

Hi Marcio,

I guess we mean different things. I don't care about ISO settings as they are pretty arbitrary. There are about three definitions of ISO for digital cameras, my understanding is that two of them relate to JPEGs. There is a third way of measuring ISO and that is based on saturation. DxO does measure that, on raw files.

The figure below shows measured ISO-s for three different Phase One backs.

ISO Sensivity for three different Phase One backs

If you check out the graphs, you will see that the IQ-180 has ISO = 29 at 100 ISO setting, he Phase One P40+ has 46 ISO at 100 ISO setting and the P45+ is has 95 ISO at 100 ISO setting.

So that means that at 100 ISO:

  • The P45+ exposes correct
  • The P40 exposes one stop under (thus yielding one stop of headroom)
  • The IQ 180 exposes 1.78 stops under thus yielding almost two stop of headroom.

I know that the above doesn't explain all your data.

My discussion is based on ETTR exposure that is based raw histograms. Most cameras don't offer raw histograms, unfortunately, so looking into that needs a turn over the computer. But looking into raw histograms is a great learning experience.

My experience is with the P45+, which you say has least of highlight headroom of all cameras you tested and the Sony systems, so it may be less relevant than if I tried a lot of other systems.

But, I am pretty sure that highlight headroom is dependent on exposure bias and also processing parameters. Adobe's DNG specification allows for an exposure bias. So zero-adjustment will be corresponding to BaselineExposureOffset.

  • 0.35 for Sony
  • -0.65 for Nikon D800
  • -1 For P30
  • 0.08 for Phase One IQ 180
  • -1 for P45+

I don't figure how I could use those figures.

Best regards

Erik

Iliah Borg Forum Pro • Posts: 25,756
Re: MFD Highlight DR…

Adobe's DNG specification allows for an exposure bias. So zero-adjustment will be corresponding to BaselineExposureOffset.

  • 0.35 for Sony
  • -0.65 for Nikon D800
  • -1 For P30
  • 0.08 for Phase One IQ 180
  • -1 for P45+

Adobe's exposure offset is a bias between how the camera exposure meter is calibrated and what Adobe consider to be standard exposure meter calibration (around 12.7% of saturation), sans Adobe's errors.

To quote DNG specification:

"Camera models vary in the trade-off they make between highlight headroom and shadow noise. Some leave a significant amount of highlight headroom during a normal exposure. This allows significant negative exposure compensation to be applied during raw conversion, but also means normal exposures will contain more shadow noise. Other models leave less headroom during normal exposures. This allows for less negative exposure compensation, but results in lower shadow noise for normal exposures.

Because of these differences, a raw converter needs to vary the zero point of its exposure compensation control from model to model. BaselineExposure specifies by how much (in EV units) to move the zero point. Positive values result in brighter default results, while negative values result in darker default results."

-- hide signature --
Erik Kaffehr
OP Erik Kaffehr Senior Member • Posts: 2,772
Re: MFD Highlight DR…

Hi Marcio,

Thanks for sharing your findings. Nice to have observations on many cameras! Unfortunately, I can only test the systems I own.

Best regards

Erik

Keyboard shortcuts:
FForum MMy threads