Discussion of all digital SLR cameras under the Minolta and Konica Minolta brands
Forum rules
No more than three images or three external links allowed in any post or reply. Please trim quotations and do not include images in quotes unless essential.
I have just reread the DRO article on this site. Now that I have an A700 I wanted to test how DRO advanced underexposes. It's written somewhere that e.g. instead of shooting ISO800 w/o DRO I could shoot ISO400 with DRO since DRO will underexpose and give me the same (or similar) shutterspeed than ISO800.
Now I went into Aperture Prio mode, central focus point only, ISO400. I half pressed the shutter with and without DRO (set to level 5 for the test). I focused on the darkest point of the scene, the brightest point in the scene. I tried high contrast, i.e. half of the pic in the room (rather dark) and half of the pic out of the window, i.e. rather bright.
Result: Whatever I did, the shutter speed always was identical between DRO off and DRO level 5. Not a single difference.
bossel wrote:Did I misunderstand how DRO underexposes?
Umm ... yes, I think so. Why do you believe the purpose of DRO was to underexpose? Actually, it compresses the input dynamic range in a complex way so it will fit the camera's dynamic range and still look good. As a side effect, it's able to save underexposed frames if they happen for some reason (at the expense of increased shadow noise, of course---there is no free lunch, as the saying goes).
DRO does not alter the actual exposure. It does alter the effective ISO performance of the RAW file in certain conditions. This also applies to the earlier Minolta/Sony technology for 'Hi' key (Hi 200, Hi 250) which also will produce a 'dark' raw image. With Hi200/Hi250 this is consistent; the camera sets the metering to 200/250, but actually sets the ISO sensitivity to ISO 160/200, while the EXIF data shows 200/250. A 1/3rds stop* underexposed raw file is always produced using the Hi settings, and a special A to D conversion curve is used in the camera to lighten the midtones and shadows, while retaining the slightly darker highlights. This is how the 'hi' settings preserve detail. *I think the 7D is 2/3rds under.
Advanced or auto DRO+ can behave in a similar way - with DRO turned off you get 1/125 at f8 and ISO 100, and a normal raw file density; with DRO on you set get 1/125 at f8, it still says you are using ISO 100, but in fact the A to D is modified so you are really using ISO 80 and the raw file is 1/3rd stop darker. DRO needs this to be able to recover highlight detail as well as boost midtone and shadow detail.
DRO+ is linked to multi-segment metering and can't be used with spot metering, manual exposure (etc) because fairly often a setting like spot metering will result in an overexposed raw file. DRO can't process the data from an overexposed raw file properly, and must always work on a minimally exposed one. So when you are using it, there are certain conditions where DRO+ or DRO Manual will produce a darker raw image; in 'normal' conditions it will not affect ISO (gain) but in others it will.
Sony does not give any indication that raw files may be darker when DRO is used (for apparently identical exposures). Since the embedded preview and associated JPEG-type files up to HDTV resolution (part of the raw file) are processed by DRO, you 'see' a normal density raw and are unaware that it may have lower values. If you don't instruct Adobe Bridge/ACR to build high quality thumbnails, and opt for slow previews and high quality on preview only, you won't see any apparent difference in ACR either.
This is the main reason why DRO+/Advanced images can be noisier than images with DRO turned off, even when you use the raw file and ignore the DRO processed JPEG entirely.
Nikon does print a small warning that Active D-Lighting (a direct copy of Apical's DRO, but not using their chip in the DSLRs, only in their compacts like the P5000) may result in darker raw image data.
'May' is the operative word, because the link to matrix-type metering means that the effect is variable. My article shows circumstances where it is exaggerated, and where DRO+ Manual is also rescuing even more underexposure than required (the flash pictures). It would be easy to test DRO+ and think that absolutely no difference is made to the raw file; it's also easy to test it and conclude that DRO+ has no effect, if you don't choose the right conditions for the shot.
Thanks David this answers my question. So my test was doomed from the beginning. I was watching the shutter speed, which doesn't change. Because in fact DRO+ uses the ISO to alter the exposure, and without showing this anywhere!
I may be slightly wrong about 'changing the ISO' behind the scenes like the Hi setting - DRO+ may change the A to D output curve and make the raw file data more linear, which looks much the same (a darker file in the midtones).
I wanted to test how DRO advanced underexposes. It's written somewhere that e.g. instead of shooting ISO800 w/o DRO I could shoot ISO400 with DRO since DRO will underexpose and give me the same (or similar) shutterspeed than ISO800.