Do you believe what I showed above is Moire? Please test. I did re-tests, but with less than 700 exposures on my original ZD and less than 700 on the loaner behind, and many of them trying narrow down above problem during my ownership, both on boring test shoots and in actual photography: I am frank tired and now want a product that does not give me problem.
I am now looking at Phase One and Leaf Aptus, even though still more $$ over my D200 or F100 was. Phase and Leaf can vaguely show magenta/purple in black/dark areas but it seems they significantly more control the artifacts, and they are also capable of higher ISOs and more stops below mid point neutral than ZD. I appreciate if someone with experience of such DBs would reply of those products in comparison.
Someone gave me a very interesting email reply on me sharing the same images sent by my agent to Mamiya above (confidential source):
“- The Image "All ISO": by looking at the different ISO settings one can easily notice noise at all ISO's, even with ISO 50. This noise can be seen in some green/magenta "patches". You can find exactly the same green/magenta patches in the 2 other images (girl & temple) in some shadow places. This shows simply a high noise level in coordination with an interpolation which seems not to be very good and efficient.
To be honest, all the ZD images I have seen so far have shown to me these very same green/magenta zones in the shadows and sometimes even in the 1/4tones. It seems here that the handling of the data out of the sensor is not really good (interpolation algorithm). Even if the sensor produces such noise, a good interpolation can get rid of it in most of the cases. In any case, "Neat Image" noise filter should be able to handle most if not all of this noise.
- What I can also notice in the 2 other images, besides the extreme high noise level, is a lack of details in the shadows: some (many) zones show RGB values of 0/0/0, this even by applying a "linear" curve in ACR. This with images shot at ISO 100 & 160: this is quite weird and unusual! Did you send the original data, or where those already manipulated by some toning curve applied?
- You are saying to have this noise problem under certain conditions: do you know when and under which conditions exactly?”My reply to questions in that reply are following: Neat Image and other noise ware sometimes can help (emphasis on sometimes), but they make image softer. It can also be cleaned similar to moiré by using paint with color (not pixels) in Photoshop (see Moiré Go Away @
http://www.russellbrown.com/tips_tech.html ). However, this makes color uniform which is not very acceptable for nature photos since color in nature is usually not uniform. Files that I emailed this person had not been altered. They were RAW files, same as posted above, and with no curves applied, and JPGs with default settings from SilkyPix (using AdobeCamera RAW result in same). The conditions that this can occur in are as I have described in above with dark and light subjects.
ETTR (exposing to the right) do reduce the issue. ETTR at ISO 50 is similar to shooting at ISO25 and is the safest way to reduce the problem, but it is of course also dependent on the DR of subject scene. ISO 25 is very slow for my landscape and nature photography. Per my testing the problem artifacts can show at ISO50 when exposed per camera meter neutral. If one instead underexpose it is similar to using higher ISO setting and can bring out noise and these artifacts in dark tones and mid tones (this is normal what happens to normal noise in dark and mid tones when underexpose). However, as shown by my test when scene exceeds 4-5 stops it seems artifacts may or may not become present in dark areas depending on the scene. This leaves that artifacts can occur at just 2-2.5 stops below mid point. In such conditions it seems slide film is more capable, or a DSLR. Medium format sensors are otherwise said to have about 12 stops of DR which is more than negative film and DSLRs.
Having now demoed and seen test samples from Phase One and Leaf Aptus (and correct me if I am wrong), it seems all MFDBs can show similar to above. However those other products much better control this so that at low ISO it seems as an insignificant problem, seemingly done both by more appropriate algorithms within the digital backs and in their designated RAW converters. Mamiya has not done so. In fact the Leaf Aptus 22 apparent have same Dalsa sensor as ZD but performs very significant better than the ZD.
Above are reasons why I have given up on ZD. Likewise I am frank not pleased that nothing was done to correct the problem when I reported it, and that I was not even given an appropriate fact reply from Mamiya.
The issue of very limited ISO range and indeed the above described problem made me give up. Else I see the ZD camera as having the best user interface of any DSLR I have used or seen. It is disappointing.
I believe Mamiya need to wake up!
Any comments from other ZD or MFDB users are most welcome.