Understanding DRO+ / D-Range on a55 and RAW images

[DrScottNicol]

OK, I've been poking around on the internet (including the article on this site by David), trying to understand the DRO / D-Range function, specifically whether, as I'm predominantly a RAW shooter, I should leave it off, on auto or some other setting. Given Sony gave it its own button on the a55 (that I've not touched since I have the camera but I'm now a few 1000 shots in), I thought understanding its effect might be a good idea .

So, I think I understand the basics of what its supposed to do (selectively bring out detail in shadows without blowing out the rest of the image?) and that according to one poster referring to an a700 its hardware accelerated 'in camera' because it has to work on each pixel separately and optimises the image on that basis (rather than a global adjustment of colour curves?) and it only really works on jpegs (but that includes the jpegs used for live view viewing). Someone also mentioned that different cameras in the alpha range have different implementations of DRO from basic to hardware accelerated licensed from a third part (apical?) - is the implementation in the a55 the more sophisticated hardware accelerated version?

I've just performed a very quick and dirty (and if I'm honest, slightly shaky / blurry) test of the function at different DRO levels, (off, auto, +3 and +5) of a well lit room, medium lit corridor and dark room at the end, all in the same image. Now on the back of the screen, with DRO off or on auto, the dark room is very dark and I can see little difference between the 2 images on the liveview screen or when I look at the images in the sony PMB viewing software. As I crank up the DRO setting, the well lit room looks fairly similar (I can see the shadows brighten), the orange wall colour of the corridor gets a little brighter and a few objects start to appear out of the shadows in the dark room. When I look at the RAWs in the PMB viewer, I can still see these differences.. but when I take them into my raw converter (in my case, DXO optics 6.5), I can see no difference between DRO off and DRO +5. From this, I'm deducing / making the assumption that the despite what the live view looks like, DRO does not influence the RAW files dynamic range, just the thumbnail jpeg thats generate for liveview, that the PMB software either emulates the DRO effect in software or just grabs the JPEG thumbnail for quick viewing purposes and that DXO Optics renders a true 'RAW' image (or at least its conversion of the basic data before you apply its corrections). On a side note, I can quickly recreate a reasonable facsimile of the DRO effect using DXO's auto HDR functions but the DRO effect looks a little better without any further tweaking.

So, based upon the above.. am I along he right lines with regards DRO and RAW files? I'm a little confused as the the Auto setting - does it see I'm shooting RAW and switch DRO off? Or is it a little less predictable than that? If so, what would people recommend when shooting RAW...

1) Turn it off - its resulting in misleading liveview images that might result in disappointing RAWs if your not careful
or
2) Turn it to AUTO - in RAW mode it knows to switch it off and shouldn't interfere, and if you occasionally use a function that demands JPEGS, it will kick in at an appropriate level

Let me know your thoughts / point out any misconceptions in the above
regards
Scott

[David Kilpatrick]

DRO does not affect the raw file except in one way, which is that in the A700 and A900 it can favour underexposure in order to preserve highlight values, and then do its processing to create a brighter JPEG with zonal adjustments. I don't think it has the same effect on exposure with later cameras (at least, not so visible). It appears to do this by reducing the effective ISO, the actual exposure is what you would expect.

DRO only has its visible effect in the JPEGs produced in-camera. DRO Basic is simply a gamma curve (brightening midtones) and is replicated perfectly by PMB/IDC. DRO Auto can, in theory, move between this and Advanced - which is the Apical local contrast adjustment process, using the Iridix chip in the A700/A900/A850, but I think this has been incorporated directly into the BIONZ (Fujitsu) in later or lesser models.

The Manual strengths simple instruct the DRO process (Apical's firmware/software) as to what degree of adjustment it should apply.

IDC, DxO etc can not replicate this without using the patented Apical algorithm. Canon's DPP and Nikon NX2 appear to include processes which are closer to Apical's method, but after the Coolpix 5000 (I think?) Nikon abandoned Apical and developed their own alternative. Canon has never used it.

The reason Sony has remained as a user of Apical's chips and patent algorithms is simple enough - the real value of this process is found in TV and movie cameras (as well as in military and surveillance). The DRO Advanced Auto process is effectively found in higher end video cameras, using one or more processors, even if not labelled as DRO.

Because Sony continues to use the DRO technology in-camera but chooses not to pay a stack more to encrypt and embed it in software raw processors, you end up with a very weak imitation of DRO in IDC.

People think Photoshop fill light etc can do the same, but they can't, they have a global effect and are not zonally adaptive.

I do not use DRO routinely but would know when to do so, if it could rescue a situation in a big way. The best example has to be any kind of bounce or direct flash work indoors. DRO can really fix flash illumination problems pretty well.

David