7D Dynamic Range Testing

[Timber Wolf]

Hi people. Decided to repost in a new thread.

Methodology: Shoot a textured white wall with varying exposure settings at ISO 100. Check the histogram on the camera for blown-out highlights or blocked shadows.

Results:
JPEG Normal Contrast: -3 to +2.5 OK (5.5 stops DR)
JPEG -2 Contrast: -3.5 to +3 OK (6-6.5 stops DR)
RAW: -3.5 to +3.5 OK (7 stops DR)
If you really push your luck then RAW -4 to +4 (though i wouldn't count on the +4, the green channel blows easily)

One more observation is that my camera's meter runs cold, underexposing by -0.5EV normally, if the in-camera histogram is anything to go by. It's only when I set +0.5EV compensation does the spike which represents tt white paper centre on the middle line of the histogram. Design feature or bug?

Hope you guys can comment. Thankew =)
QX

[deslim27]

Hi Wolf,

Will you able to post your test shot, I interested to see them.

Cheers!

[kelster]

yup yup...curious for me as well...

[Timber Wolf]

Originally Posted by deslim27 Hi Wolf, Will you able to post your test shot, I interested to see them. Cheers!

Ok. Let me resize them first. I'm having abit of trouble with RAW processing though. There is info in the file but it just won't show. Will wait for Adobe Camera RAW.

[Timber Wolf]

Ok people. Here's your link: My Gallery

I've included Zone Matching in the tests. Apparently it gets you quite abit more dynamic range than otherwise possible. Look at the +3.5EV ZoneHigh as compared to +3EV Contrast-2 example. While the C-2 example struggles to hold onto the highlights the ZoneHigh does it very comfortably.

Enjoy. and Des (and others ), if you have the time, do test your cam k? Thanks. Hope to learn from you pple abt the "art" part behind photography.

Thx =]

[Timber Wolf]

Hi pple. Here to bore you again with descriptions of un-artistic wall shots

Firstly, Noise-Reduction at High-ISO with exposure time LESS than 1 second.
I shot a white wall, out-of-focus at ISO 3200, 1/6 sec f/38.

Observations:
The shots with NR and NR-Off are identical, in terms of noise pattern and grain. There is no discernible difference even when i turned my monitor contrast to max.

Using Neat Image noise-analyzation tools:
NR-Off Std-dev: 8.19
NR-On Std-dev: 8.31

So you can see that there's not much difference, if any, in images with NR On and exposure time Less than 1 second.

Secondly, Zone matching tests:
ZoneLow Dynamic range: -5.5 to +2.5 (+3 is completely blown out)
ZoneHigh Dynamic range: -4.5 to +3.5~4 (+4 quite hard to say, some parts blown)

There's clearly something KM is doing lol.

Lastly, some curiosities with KM's Zone Matching:
If I'm not wrong, ZoneLow operates at ISO 100 and ZoneHigh at ISO 250. This means that for a similar exposure, the shutter speed for ZH should be 2/5 that of ZL.

Here's the interesting part. When I shot ZoneLow, the exposure at +3EV was 1sec at F/11. This resulted in a largely blown picture. However, when I used ZoneHigh, at +3.5EV the exposure time was 1/2sec at F/11, more than the equivalent exposure for ZoneLow. But... the picture for ZoneHigh was full of detail! Nothing blown, and the histogram showed that nothing was even -close- to blowing.

What do you guys make of this? It's the same sensor, so by right signal saturation -should- occur at the same point, but it seems that processing matters alot in this case. If so, why not have a ZoneHigh/Low which gives us both sides of the DR?

Cheers people, and thanks for reading thru such a long post.
QX

[deslim27]

Too chim to digest..

BTW, NR will not be activated if shutter speed is less than a second according to the manual, so your finding quite correct.

As the zone matching, I have not tried it.. Don't really know what it will do.. KM 7D site has 2 example, with high and low key shot each, have you seen it?

Quote: Enjoy. and Des (and others ), if you have the time, do test your cam k? Thanks. Hope to learn from you pple abt the "art" part behind photography.

I'm a newbie lah..

[TME]

Manz.... this is so chim....

[Timber Wolf]

Hahah. Ok sorry for the cheem stuff lol. I'll try to put it into more simple terms. For Zone Matching, yes I've seen the pictures on the KM website. What it doesn't say is that you actually have to add +EV to make the whites actually look white. From my tests, the camera just attempts to expose for 18% gray (actually my meter underexposes 0.5EV). Likewise for dark subjects.

I think that the camera adjusts the tone curves alot, as what would be blown in one example is captured easily using ZoneHigh. That could mean that
1)The camera is performing magic with the captured data OR
2) In JPEG mode without Zone Matching we're getting shortchanged for about 1-1.5EV of dynamic range.

I haven't seen any improvement in RAW mode, but I suppose setting Zone Matching allows us to get the results we want with less post-processing.

Cheers =]

[forbytes]

I do not have 7D, following are my $0.02 on zone matching.

I believe that the camera is programed to under expose for high-key and over expose for low-key. the objective is to obtain a balance exposure overall picture. if you see your result, the positive compensation for exposure is 2.5 and 3.5 for low key and high key respectively. Difference is 1.5 stop for ISO 100 and 250 setting provided same aperture and shutterspeed.

If you shoot underexpose a high key, details will come out but dark, the brightness curve can easily increase for mid and shadow area. Remember the camera does have the capability to determine highlight area. overall is balanced. This can be easiy verified if you use the 14 honeycomb EV display found on D7 (I do not think 7D have this feature). High key always full of detail even you +EV since the camera already under expose.

for low key, camera goes into over exposure mode. then decrease the brightness for bright and mid. if you +EV, the chance of blow out is very high. That explains why Timber Wolf +2.5 EV and details already burnt and this phenomenon is not found in high key.

I believe when KM implemented this feature of zone, they are assuming that you have only mid and high for high zone and shadow and mid for low zone. The subject Timber wolf should be classified as a high key subject due to white, forcing the camera to go low zone will naturally cause burnt images. It is impossible to force detail out from this image using tone adjustment coz it is not recorded.

No offence to anyone.

[TME]

Forbytes, sounds like the 7D has a mini Adobe Photoshop onboard? What u describe is like tweaking the curves...

[forbytes]

maybe timber wolf measure the time taken from release shutter to end of write cycle (RAW) using the LED indication. If the time taken for zone matching is longer than normal, well, certainly some work was done by the camera "CPU" right?

[jonoq]

Whow..after reading the thread, i feel that i still studying in nursery ..Catch no ball