[ScottB]

Hi to all,

I'm currently calibrating my Sekonic L-758 meter with my new 645z. After doing the test shoots running it through DTS software I come up with an exposure range of 4.5 stops. This is truly not what I was expecting, My Canon 6D is even better at 7 stops range. After looking at the files in camera raw it shows clipping on both 4 stops over & under. I was hoping for at least a 10 stop range.

My question/s what kind of exposure range/s are others experiencing (camera firmware is 1.22)? Has anyone else tried calibrating the same combination and what results did you see?

thanks
Scott

[photoptimist]

That does not sound right.

How exactly are you doing this? RAW's or JPGs? Which lens? What scene?

[TonyW]

I agree the results are incorrect or at least by the sound of it your interpretation may be wrong. The 645z has a photographic dynamic range of around 12 stops. I say photographic because that would be classed as a usable range rather than the DR quoted elsewhere of 14+ EV (you cannot really have more DR than acquisition bit depth). My own testing found the DR to be around the 12 stops quoted and not far from Bill Claffs testing which showed 11.7 EV. The Canon 6D having a DR or 9.3.
Photographic Dynamic Range versus ISO Setting

First using the DTS software what target are you using and under what lighting conditions?
AFAIK you cannot use raw therefore you have to convert or acquire as JPEG for the DTS software - is it possible that the conversion has caused issues?

You may find that the DTS software is set to be fairly conservative with clipping point and you should consider what you need to set. The example here shows the DR as 12 and a clipping point range of around 9.3EV i.e. from -5.8 to +3.5 - clipping point is arbitrary to a degree and should be set by practical testing in the field

[leekil]

I agree the results are incorrect or at least by the sound of it your interpretation may be wrong. The 645z has a photographic dynamic range of around 12 stops. I say photographic because that would be classed as a usable range rather than the DR quoted elsewhere of 14+ EV (you cannot really have more DR than acquisition bit depth).

AFAIK you cannot use raw therefore you have to convert or acquire as JPEG for the DTS software - is it possible that the conversion has caused issues?

These two statements seem to be in conflict; if you can't get more than 14 EV with a bit depth of 14, how can you get more than 8 EV with a JPEG, which has a bit depth of 8?

[TonyW]

These two statements seem to be in conflict; if you can't get more than 14 EV with a bit depth of 14, how can you get more than 8 EV with a JPEG, which has a bit depth of 8?

You cannot - your camera acquires images as raw data and this is converted in camera (or in post) to JPEG and in the process data is discarded either by the camera designers visions if left in camera or by your interpretation in post

You are looking at different things here one being the acquisition device which is limited by the analogue to digital converter (A/D) and processing software and the other being the file type and its limits. In the case of a camera system and the linear A/D conversion DR measured in EV cannot exceed the A/D conversion bit depth. So in an 8 bit camera you would not be recording any more than 8 stops in fact you may be limited to around 5 stops - which is about the same as some analogue slide film

A very rough and quick attempt at an explanation:

Dynamic range is the ratio between the brightest and darkest intensities recorded by your sensor as a linear capture.

Bit depth on the other hand is the number of bits per channel used to display distinct colour shades between the minimum and maximum

Your acquisition device (Pentax, Nikon, Canon)has a linear A/D converter and depending on camera this may be 12 or 14 bit system.
If it is a 12 bit device the limit is 4,096 levels which increases 16,384 levels @ 14 bit.

A typical chart showing the relationship between sensor bit depth below. Please note this is an ideal scenario but does illustrate DR expressed in EVs cannot exceed the ADC resolution expressed in bits. No account taken of the unacceptable noise levels that would occur after the 9th or 10th stop with 12 or 14 bit system - noise being subjective with one person accepting more or less than another

Bit depth relating to colour or grey shades is different as an example:
A JPEG 8 bits per pixel maximum number of colours available 256 per channel for a total of 16,777,216 colour shades (=24 bits: 8 red, 8 green, 8 blue)
Deep Color of 30/36/48-bit (10/12/16 bits per channel) consists of billions of colours e.g 16 bits per channel (48 bit) 281,474,976,710,656 colour shades

[ScottB]

Hi thanks for all the replies. I was expecting about 10 to 12 stop range.

ISO 100 both meter and camera, shooting in Manual mode
Color checker passport is the target, in ambient light out of direct sun, metered the 18% grey card (newer ones have a dedicated sample now)
Set Camera Default profiles in the meter for all three camera profiles before starting process.
Latest firmware in the meter, and 5.1 on the DTS software.
Took 5 shots +/- 4 and 8 stops and metered correct exposure.
FA 80-160 set at F16 on the lens.
Shot raw then used Camera Raw in PS to convert them to jpg, choose highest resolution.
In the software choose extended, entered both ambient and reflected meter values, adjusted green frames, choose ISO.

Tony your attached graph is what I normally see when I calibrate for my 6D a nice curve but the 645z there was no curve just junk. I've followed the above steps many a time and always gotten the results I normally expected to see.

Glad to hear others are getting the results I was expecting at least I know something is a skew other than my imagination. I will run the entire test again tomorrow.

Thanks
Scott

[leekil]

You cannot - your camera acquires images as raw data and this is converted in camera (or in post) to JPEG and in the process data is discarded either by the camera designers visions if left in camera or by your interpretation in post

Right -- so that means that the OP should not use a JPEG as the input to the dynamic range testing software. No matter if there is a problem in the conversion or not, it will only give 8 stops at the most. I guess I was confused because you didn't mention this initially, but seemed to imply that it would be OK to use a JPEG.

[Kolor-Pikker]

Right -- so that means that the OP should not use a JPEG as the input to the dynamic range testing software. No matter if there is a problem in the conversion or not, it will only give 8 stops at the most. I guess I was confused because you didn't mention this initially, but seemed to imply that it would be OK to use a JPEG.

It's a bit complicated, an 8-bit jpeg can show high dynamic range without actually being able to store all of it, by way of compression, whether it be highlight/shadow recovery or another method. So each physical stop on-screen is represented by less bits than are required for them to be considered "lossless", however if you try to edit such an image tonally, it will immediately break apart and cause posterization. JPEG is a delivery format, simple as that.

The issue with the 645Z is that preview images or minimally processed jpegs do a terrible job of showing what the sensor is capable of, because most of its dynamic range is biased into the shadows, the histogram you see in-camera maybe represents only the top 2/3rds of your available latitude.
Converting straight into a jpeg without performing any shadow recovery will just throw away most of the dynamic range, as data near the extremes becomes regularized to "fit" the new data constraints.

[TonyW]

Right -- so that means that the OP should not use a JPEG as the input to the dynamic range testing software. No matter if there is a problem in the conversion or not, it will only give 8 stops at the most. I guess I was confused because you didn't mention this initially, but seemed to imply that it would be OK to use a JPEG.

No JPEG is OK for this purpose and is what the software expects.
Briefly you make a series of exposures of a colour chart either side of the indicated exposure 1EV apart - none are more than +/- 8EV. These images are fed into the Sekonic DTS software which analyses the results of these images and produces the results shown in the posted graph, showing DR in terms of plus and minus EV either side of the correct exposure to produce mid grey (RGB values 118 - sRGB or Adobe RGB)

[TonyW]

Scott,

Just an afterthought:
Something in the back of my mind seems to recall that the 51MB files of the 645z may be troublesome for the DTS software.

Not done it for a while but you may want to consider trying this.

1. Acquire images as normal in raw DNG
You may also want to try small JPEG as well/ instead

2. Convert to TIFF while also reducing image size by at least half

With luck you should now be good to go

[ScottB]

Hi to everyone.

I sent my files to Macgroup and apparently the problems were that my images were rotated vertically and contained to much of the color checker. Once rotated to landscape and cropped to just have the color test squares and not the entire color checker it created an acceptable calibration. See attached. Guess I've never had those issue with the earlier version of DTS but maybe I was just lucky. Guess I thought that's what the green alignment corners were for

I think I'm going to try using the extended mode and see what results I get. At one time the product manager also suggested instead of 3 stops to do 5 stops for the test.

Scott