Olympus E-PL1 and Photoshop - "Correct" Exposure

Olympus E-PL1 14-150mm

While I've been very impressed with what Capture One Pro from Phase One can do, I am unsure as to whether I'm going to buy it when my trial period runs out. Firstly its expensive, though they are offering a 30% discount until 28th August. Secondly, its more complicated and takes longer to process images than Adobe Camera Raw and Photoshop. Thirdly, its not really how I like to work. I prefer a simpler raw conversion that lets me do the majority of the work I want to on a file in Photoshop, rather than in the raw conversion software.

Olympus E-PL1 14-150mm

Olympus E-PL1 14-150mm

The trial I have downloaded also lets me try out the Express version of the software, though the nice thing about the Pro version is that it lets me open the files immediately in Photoshop.

I have tried most of the raw converters on the market and as I wrote yesterday this version of Capture One is the best I've ever used. The question is whether I want to go for the extra "complication" of stand alone software, even though it does integrate with Photoshop.  It is very tempting, since some of the files I've produced are very impressive and it does do a very good job with m4/3 files.

Olympus E-PL1 14-150mm

However, using Photoshop and Adobe Camera Raw works very well too. All the files in this post were processed using that combination. I wrote yesterday about the need to expose and "develop" carefully with m4/3, in order to get the best out of it. While I think the criticisms regarding the more limited dynamic range of m4/3 are often overstated, there is no doubt that it is more prone to blown highlights than other larger sensor formats. Particularly when it is also the case that the default metering settings of all the Panasonic and Olympus cameras that I have used have a tendency to over expose for my tastes.

I generally have all my m4/3 cameras set to - 0.7 in terms of exposure compensation, and this, more often than not gives me the exposure that I find most accurate. In fact, I do this with most of my cameras. It happens with both my Nikons and there has been a lot written in forums about how the Nikon D7000 has a tendency to over-expose slightly. While I can understand doing this in point-and-shoot compacts when the users probably won't make any adjustments to the file after taking the shot, it does seem a little strange in cameras that are aimed at a different market.

I'm not advocating deliberate under exposure here, I'm actually talking about getting the exposure correct. There is this theory that with digital cameras, you should expose to avoid having to lift the shadow areas, because thats where noise can be generated. However that does assume that you get the exposure right in the first place. Also using the new Sony 16MP sensor in the NEX-C3, Nikon D7000, Pentax K-5 etc. there are far less problems in exposing for the highlights, as that sensor allows shadow areas to be lightened quite dramatically with no or very little increase in noise.

I wrote a piece on it here:- http://soundimageplus.blogspot.com/2011/01/pentax-k-5-three-lenses-low-iso.html and this image, taken with a Pentax K-5, shows what you can do.

Pentax K-5 15mm F4 lens

Micro Four-Thirds is not quite as forgiving as this, though here's an example where I've lifted the shadows with no great increase in noise.

The other great advantage of continuing to work with Photoshop and Adobe Camera Raw, is familiarity. I feel that I've only begun to scratch the surface with Capture One Pro, and the question is do I want to allocate the time to learning all its possibilities. I have no doubt that this would be a worthwhile exercise but whether I decide to do that remains to be seen.

Olympus E-PL1 14-150mm

Olympus E-PL1 14-150mm

Panasonic GH2 - Part 5 Dynamic Range

One thing I noticed in my testing yesterday was the different "look" of the GH2 files when imported into Photoshop ACR for raw conversion. They seemed to have a "flatter" lower contrast look. Dynamic range has always been a problem up to now for m4/3, but I think Panasonic have gone some way to improving it. 


All the images in this post are converted from raw with all the parameters in ACR in Photoshop CS5 "zeroed out" I've added the histogram to the picture. You can clearly see the headroom at the highlight end. There's a lot of shadow detail too. 





As ever exposure is critical here. Its easy to moan about a cameras dynamic range when you either over or under expose, and its really important to get the exposure right to retain the maximum amount of information available in order to work on the file later, where increased contrast and overall "punch" can be added.





Certainly Panasonic claimed that they had increased the dynamic range when they announced the camera and though only using the camera for a day it looks to me like they have. 


DxO Mark claim the GH2 has a "worse" dynamic range than the GH1. However I'm extremely skeptical about DxO "results" since they don't offer any examples or say what they are measuring. 


As ever I use my eyes for my own assessment. Right from the first picture I took with the GH2 I did notice this lower contrast. 


This is of course incredibly useful and deals with one of the problem areas of m4/3. I will be looking at this again as I use the camera more to see if it is as good as I think it is.





David

Dynamic Range

In a recent forum discussion on the Panasonic GH2 samples, which look really good, someone rightly pointed out the concerns about dynamic range. This has always been a problem with m4/3 in particular and digital sensors in general. 


From time to time we do some property photography, for estate agent websites and brochures. Interiors are very hard and test a cameras dynamic range to the limit. 


Below are two examples from a bracketed sequence, taken with a Panasonic GH1 and a 7-14mm wide angle zoom.





The bottom shot is getting close to what is required for the interior, but the window is completely blown out and unrecoverable. So to get a satisfactory result we had to use the top version and lighten the dark areas both in camera raw and photoshop. We prefer this to using HDR as, for obvious reasons, a natural look is required. By working with the file we eventually came up with this.





The "look" is right. Nice light bright interior and some detail in the window. Some of the exterior is burnt out but it works quite nicely. However there is a large amount of noise generated by the lightening and for the final clients version we would do more work, getting rid of that. Probably by replacing areas within the room with sections cut and pasted from other lighter shots. The section on the left with the shelf is particularly nasty and would have to be substituted.


With a suitable budget and the time to do it, a better result could be obtained by using some flash heads, but the circumstances dictated that this had to be done quite quickly and cheaply!!


So some problems with m4/3, but shots like this are very difficult with any camera. They were difficult with film and they are difficult with larger sensors also.


Heres another from the same shoot. This is the camera jpg taken with a Canon 5D MK II and 20mm lens.





The range is slightly less than the first shot, as there is more light and less dark areas.


The final result after Photoshop work looked like this.





Again an OK result with a light interior and detail outdoors. However there is still a burnt out sky, and a considerable amount of noise which would have to be eliminated.


Interior shots like this are incredibly difficult. You only have to look at the problems movie makers have with situations like this. Unless they use banks of interior lights, they will suffer from the same problems as above. In many cases, it just has to be accepted. Currently we have no film nor camera sensor that comes close to what our eyes can cope with. Someone in a recent forum expressed the concern that manufacturers obsession with higher resolutions has led to the problems of dynamic range being largely ignored. He rightly made the point that by now we should have something better than we have. Even the very best film or sensor camera/lens combinations are a long way off recording what we see with our eyes. We can come up with solutions because we have to, but it would make things a lot easier if we didn't. I'm afraid in-camera or software HDR isn't the answer. Can be useful for static subjects, though we are not convinced by what we have seen, but completely useless when one shot is all you get. 


In terms of m4/3, yes dynamic range is more of a problem than with larger sensors, though we've not seen any great improvement from the APS-C sensors we've used, and in the really difficult situations other solutions have to be found. They have to found for full-frame sensors as well as m4/3, so the notion that somehow these larger sensors solve all your DR problems is somewhat inaccurate. They don't. They may be marginally better, but it is marginal and we need somebody to come up with a satisfactory solution. However it still seems a low priority for camera manufacturers.


Finally, going back to the GH2 samples, they did seem to be no worse in terms of DR than the GH1. The high ISO performance looked excellent, with samples at ISO 3200 and 6400 looking very similar to the Canon 550D, which is high praise indeed. The upsizing of the sensor seems to have worked very well, as would have been expected. Panasonic wouldn't have done it if they knew it was going to be inferior to what they have already. If the samples translate to real world performance then they really have come up with something special, and deserve to be regarded as at the forefront of current sensor technology. A little more concern about DR wouldn't go unappreciated however.


D & A