Histogram and highlight warnings

[SrMi]

I wonder what approach X1D/907x owners are using to expose correctly. You may want to use ETTR or not, but everybody wants to avoid blowing highlights.

I typically use histogram and highlight warnings in image review. Since histograms use JPEG instead of raw data, they can be misleading. The histogram accuracy improves by specifying a neutral JPEG mode, but X1D does not offer that option.

I have compared the in-camera information with the information provided by the RawDigger program (raw data analyzer).
In typical scenes, when an in-camera highlight warning appears, there is still about 1/2 stop of leeway. E.g., X1D shows highlight warnings at 1/1000 sec, but 1/750 sec still has no blown highlights. However, when bright red color dominates, the highlight warnings are very misleading. E.g., in-camera highlight warnings appear at 1/750 sec, but 1/180 sec still has no blown highlights in the raw data. Using histogram/warnings leads to heavily underexposed images.

[JCM-Photos]

The easiest and fastest way is to spot meter the brightest image zone and apply to it à + correction (rear wheel) depending how bright you want to render this highlight and then block it with permanent AF lock.
After some trials you will exactly now how much you have to correct.

[SrMi]

Thank you for your response. I tried it before, but I could not use spot metering on the brightest part in a reliable manner.

One disadvantage is that it is easy to overlook the brightest part of an image, which leads to an underexposed image.

Also, "spot metering can be accurate if three readings are made, each one with a filter which makes the meter's response the same as the response of one of the camera's raw planes" (quote from Jim Kasson). In my tests, depending on the colors in the scene, the correction for spot metering with X1D is anywhere between +2.5 (balanced colors) and 0 (bright red).

I have found an article (link below) that suggests a different approach for bright light: set the camera to daylight WB and use the green channel to determine if clipping occurred. Fortunately, X1D provides a large RGB histogram. I will turn highlight warnings off and use the green channel to evaluate proper exposure. Wouldn't it be nice if Hasselblad would include a raw histogram ?

https://discuss.pixls.us/t/the-myth-of-digital-reds-or-how-to-properly-expose-to-the-right/14390

[SrMi]

Wouldn't it be nice if Hasselblad would include a raw histogram ?

You mean like in my $200 Panasonic ZX-40?

I really doubt that ZX-40 has a raw histogram. Are you confusing a live view histogram (based on JPEG) with a raw histogram?
AFAIK, only Phase One and some Canons with custom firmware offer raw histograms, and Phase One apparently has issues with it in the latest firmware.

[SrMi]

Wouldn't it be nice if Hasselblad would include a raw histogram ?

You mean like in my $200 Panasonic ZS-40?

I really doubt that ZS-40 has a raw histogram. Are you confusing a live view histogram (based on JPEG) with a raw histogram?
AFAIK, only Phase One and some Canons with custom firmware offer raw histograms, and Phase One apparently has issues with it in the latest firmware.

You are probably right, though the ZS-40's 500 page user manual is never explicit about it. (Truly I gave up reading it on page 5, but still...) However there is indeed the option to overlay a histogram on top of  LV , and it does change dynamically depending on the illumination of the spot meter.

Most cameras have an option to overlay a histogram on top of LV. All of them are based on JPG, and therefore of limited usability. X1D owners have been asking for a histogram in LV but to no avail.

I doubt the usefulness of a histogram unless it is an RGB histogram. Fuji cameras have a nice implementation where you can press a button to quickly turn on/off a large RGB histogram in LV.

[OleBe]

Good day dears,

following up on my last thread (http://www.hasselbladdigitalforum.com/index.php?topic=6088.0) I have addressed this to Hasselblad.

I have executed a lot of tests and send all that data to them. Camera thinks something is overexposed as from +2,4 stops in RAW it will start clipping as from + 4,0 stops. Quite a bit of difference.

They told me that this is on their list for improvement. The current standard JPG conversation in camera plus 8 bit limit cannot do it better.

Hopefully this will lead to a real RAW histogram instead of the JPG based one we have now...possibly with live functionality. Guess that depends if the cameras processor is quick enough and the current sensors' readout times allow this.

[disposable@tx.rr.com]

I tried it before, but I could not use spot metering on the brightest part in a reliable manner.

One disadvantage is that it is easy to overlook the brightest part of an image, which leads to an underexposed image.

Even with a completely accurate raw histogram to tell you something is clipping, you would still need to evaluate the scene in order to determine WHAT is clipping.  Some elements SHOULD clip, and this implies judgement/experience in determining exactly what is the brightest part of the image.

I'm going to suggest that JCM-Photo's advice in this matter is a sound way to proceed.  If it didn't work for you in your tests I would add that, like most everything else, mastering a new process requires practice, practice, practice.  For talentless hacks like me, it requires LOTS of practice, practice, practice

One of the programs allowing examining raw histograms for existing images, such as the excellent FastRawViewer, is an outstanding way to evaluate the results of this practice, practice, practice and master the response of your particular equipment.

And as you are already employing a fairly leisurely workflow, bracket.