Search Results : label/DxO Review

Feb 012009
 

At the moment the prevailing weather pattern in Carinthia is high fog, covered by even higher clouds.

Yesterday I was unable to locate any sunny destination that could have been reached within less than two hours. I tried my luck in Gailtal, in the valley along the southern border between Carinthia and Italy. They had the most snow this winter, and I hoped to find a freshly fallen cover, alas to no avail. Everything had this typical, dirty look of old snow. Without much enthusiasm I tried some landscapes, and this B&W conversion is the only one that I even bother to show.

It’s not a great image, but at I guess it’s quite OK from a technical POV, and it illustrates the most important thing when shooting snowy winter landscapes: not to let the snow burn out. Exposure is very critical in snow. Why? Because you want to strongly accentuate local contrast, and this would make burnt-out highlights much more obvious.

The image is a sandwich of two different conversions made with DxO, one darker for snow and sky, and one lighter for trees and distant mountains. Then I have run PhotoLift on a “copy merged” layer and set the result to “Multiply” blending mode, applying it with reduced opacity and a mask to the snow areas and the sky. I frequently try a B&W conversion to my images, and here it looked very well, because it took the occasional dirt color out of the snow. I think the result looks not too unnatural, even though the local contrasts are borderline high, and in color I would have had to tone them back.

The snow image is much better than reality, and in reality I was so disgusted with the conditions, that I decided to turn around and drive back home. I could do so, because I already had an image, taken in the small gothic church of Saint Peter, to fall back.

The Image of the Day was converted in DxO like almost all my recent images. I took it with the Sigma 20/1.8 (like all my recent images, I’d like to say), and that lens is certainly not free of distortions. DxO does not support it yet (and likely never will, it’s simply too exotic), but of course PTLens does, and this is what I used to remove the distortions and to slightly correct perspective.

The Song of the Day is (Surprise!!) by David Byrne. It’s a somewhat bizarre version of “Au Fond du Temple Saint“, the duet from Georges Bizet’s “Les pêcheurs de perles”, with David Byrne interpreting it together with Rufus Wainwright (some might say it’s the other way round). An interesting idea, demanding more than a price for courageous folly 🙂

You find it on David’s outstanding 2004 album “Grown Backwards”. Deezer has the album, and YouTube has a more orthodox version sung by Robert Merrill and Jussi Björling, recorded in 1950.

Jan 012009
 

This is the blog entry for December 31, but when you read this, it has been New Year all over the world, thus let me begin with a heartfelt Happy New Year to you all!

This blog is now well in its third year and still running daily, with the images having been shot that very day. It’s a fantastic journey, exhilarating at times, exciting and very satisfying. Your feedback plays an important role in that. I remember times when I wrote practically into the void, and now I see a steady flow of comments. It makes all the difference. Thank you.

Of the things that I’ve learned this year, I’d like to point out only one, and that’s my new found appreciation of the square. This is completely due to the works of two great artists, Ted Byrne, who uses the square sparingly but with perfection (see only his retaliation on the bicycle front), and then of course the great master of the square himself, Mark Hobson, aka “The Landscapist“.

You have seen squares from me in the past, you will see some today, and the image in the following post will be a square as well. I know that, because it’s already post-processed and uploaded to SmugMug.

Speaking of SmugMug, along with the fact that this post has so many images, SmugMug is the second reason why I am late today. At the moment they have a hardly acceptable quality of service. There is one scheduled maintenance window per week, that means a short time, normally below half an hour in read-only mode, and I could easily live with that. The problem is, that over the last months they had lots of unscheduled outages.

I know, this is all for good reasons, SmugMug is growing into the video business (Vincent Laforet promoting them), has opened up to RAW storage, and these things have necessitated changes to the whole underlying infrastructure – but still: they are a premium service, they charge a premium, and as their customer I can expect that those things that I use, care about and pay for, work and are available.

Oh well. Sorry for the rant, but I simply have to say it: at the moment I would not recommend SmugMug to anyone doing business with their images. I’ll be the first to announce when the situation has stabilized.

Let’s forget the little annoyances of blogging life, let’s get to what this is all about: the images.

My plan for yesterday was to drive down to Italy again and take a mountain road from Chiusaforte (halfway between A23 exists “Pontebba” and “Carnia/Tolmezzo”) via Sella Nevea and back to Tarvisio. I have not driven the road for 20 years, but I remember some interesting sights.

When I started out at 11:45am, I was a tad late for the trip, because at 3pm I had to be back to Villach, but I started anyway. When I came to Dogna, the first village south of Pontebba, I saw that the road into the valley to the east was open, some 18km of dead-end street, that according to the map looked interesting and would lead me high up the mountains.

I reconsidered my plans and decided to take that road into territory so far unknown to me.

The first three images were made from the same place, already high above the mountain torrent, #1 looking back to Dogna, where I had left the main street, #2 looking into the valley, and #3 a tele shot across the valley onto the bizarre ice formations on those north cliffs that never get direct sunlight at all.

Interestingly enough I found a village after maybe six kilometers, perching in a nice, sunny place high above the valley, and there were even people on the street. Hmm … must be a strange way of living there. I suppose this village is cut off the main street during snowfall, and even if it’s not, the way to the next big supermarket, to the next media shop, to the next post office, to the next school, the next whatever, must be prohibitively long.

The road winds on and on into the valley, following the sunny side, mostly being free of snow, but still, where it winds into the shadow, the street is in parts blank ice.

It’s a funny feeling to drive such a road, when you don’t know how it goes on, when you always have to expect a place where you have to turn your car, and that on a road where places to turn are hardly available at all.

This is definitely waterfall territory, and I expect to return sometime in spring, when the snow is gone and there’s still lots of water. Two or three such places were particularly promising, one of them in the image with the small bridge.

One of the reasons why it’s difficult to takes photos in this valley, is the fact that you can’t choose your light. At least at this time of the year you are forced to do your images at or around noon, because that’s the only time when there is sunlight at all. Forget about golden light, forget about sunset, these are things that happen on the mountain peaks, 500 or 800 meters above you.

For this image I resorted to the D300’s built-in flash. This tunnel with its icicles is actually part of the road, I had to drive through. Gives you another funny feeling 🙂

All the images in this post were shot with the Nikon 18-200 VR and a Hoya Pro1 circular polarizer. You know my mixed feelings about polarizers, an
d this time was no exception. The blue of the sky gets slightly pale dark, with more magenta than is actually natural, and in the image of the Day you see well the effect of an unevenly polarized sky. Technically it is a bad image, I should have left off the polarizer, or better bracketed from the tripod, one time with, one time without polarizer, and then mixed the two exposures in Photoshop. Needless to say that I did not. That I have selected exactly this image for the Image of the Day is due to the wonderful effect that the polarizer had on the mountains.

I have converted all images using the DxO Optics Pro Photoshop plugin. Not only that it does a great job in equalizing lens defects (and any super zoom has plenty of them), it also makes opening up the shadows a very easy and controlled process. All images were finally taken to Photoshop, where I did any cloning, saturation adjustments and contrast fine-tuning. If you are new to this blog, you may be interested in my series of posts about experiences with DxO Optics Pro 5.3.

The last image was taken in almost the same place as the first. The sun was just vanishing behind the mountain, and for a moment the trees on the ridge, still bathed in sunlight, were a golden crest, not much different from the typical rim light used in studio portrait photography.

Most probably I should have switched to the Nikon 70-300 VR, but this was really a matter of seconds, thus I held on to the 18-200. That’s another of the problems that you encounter, photographing in such a place: you can’t stop wherever you want, and if you do, you have to hurry. There was not much traffic at all, but still, at one time I stopped for an image, just in the middle of the road, and sure at exactly that time another car came along and interrupted me.

Well, that’s it for this short and unexpected trip. As the weather permits, I will follow my original plan some time this weekend. We’ll see.

The Song of the Day is “Higher Ground“, not the Stevie Wonder composition, no, the song interpreted by Vanessa Williams on her 1994 album “The Sweetest Days”. Amazon has a sound sample, and there is also a video on YouTube, not from her version, but from one by Melissa Manchester.

Dec 142008
 

Well, I shouldn’t promise things that I can’t deliver. These are images of yesterday afternoon, and when I posted the lens index, I had all of them already taken but none post-processed. Doing so took me until mid-afternoon today, but I guess it was worth not rushing things, and there are even lessons to be learned.

Weather in Carinthia is crazy at the moment. Villach has about a foot of snow, Klagenfurt has none. They are 40km apart with a height difference of 50 meters. There is only a lake in between, no mountain range, no nothing. For all practical reasons they should have the same weather. It’s only they don’t. This was the third weekend in a row, that I arrived in Villach during snowfall. The only reason that the snow does not pile up higher, is that we are too low. Half of the snow does not make it down to us and ends up as rain.

Unfortunately rain makes the snow quickly fall down from the trees, and somehow this looks bleak and sad. I wanted to have real snow, freshly fallen or at least looking like that, and so I took the car and drove the street up Mount Dobratsch. You know it by now, it’s that mountain that broke apart during an earthquake in 1348, the year when the Great Plague arrived in Europe. As if the plague wouldn’t have been fun enough.

From 700 meters on the road was solid snow, but I had no problem driving all the way up to 1750 meters. From there I could have gone up to the summit, but as I had entered dense fog at 1600 meters, I was not sure if it would be a good idea to go any higher. I took some images up there, made some wrong steps and was suddenly in deep snow up to my chest, in short: it did not look promising. Not winter wonderland, only winter, fog and no view. And here comes one of the lessons:

Every now and then someone posts about how RAW is inconvenient, how they don’t see a difference, how much less hassle JPEG is, while it provides superior or at least comparable image quality, and so on and so forth. Ken Rockwell is famous for it (and for all other sorts of interesting viewpoints, some remarkably lucid and some simply ridiculous), but even David Ziser did it for some length of time (Lightroom had him converted), and my friend Paul Lester does it still.

Actually there is nothing wrong with that, as long as it works for you. It certainly did for David as a highly successful wedding pro, and it obviously does for Paul. For me it does not.

JPEG is much too restrictive for me. I like to do extensive post-processing, enjoying the work in Photoshop almost as much as the actual shooting, but that is not my point today. The point is that without a RAW file, a good RAW converter and an arsenal of tricks in Photoshop, some images would not be possible at all. This is one of those images.

Take your time, click on the two versions, the second, the pale one being what the camera saw (and I as well, to be honest), and the first one what I did in DxO and Photoshop. Can you do that with a JPEG? No, you can’t. You would have to stretch the image so far beyond its capacity, that half of it would suffice to render it a mess.

Look at the JPEG from the camera: there is no contrast at all, no texture in the snow, and from that image alone, you’d get no idea of what it looks like up there. Sure, sure, I said I didn’t see much more than the camera, so why would I want to make an image that would not be true to reality? But that is the wrong question.

If I only had the JPEG, I wouldn’t need to bother showing it at all. This gray mess, what it is about? Well, not much more than bland grayness. You can’t really tell how the snow covers these barns, you can’t see how it smoothes out any jagged form, the picture is a complete waste of storage. As a JPEG shooter you’d have to throw it away.

So, obviously there is a reality of soft, ondulating forms up there on the mountain, and shooting JPEG, the true-to-reality plain JPEG, you can’t show it. Well, in RAW you can.

When I work on such an image, I always try to get the best RAW conversion possible. I know I’ll have to stretch the image contrast beyond believe, and I know that all flaws in the RAW conversion software, all flaws in my own technique will be amplified and thus revealed. Knowing that, I always begin as solid as possible, and since almost two months this means to convert the RAW in DxO Optics Pro. I have already written a lot about that program, I am still satisfied with it, and at the moment I use it for all high ISO and for all snow images.

Why for snow, you ask? Well, the reason is, that I need to stretch contrast, and often also to dramatically increase saturation. RAW conversion artifacts and lens flaws like chromatic aberrations would become dominant in such an image. I have experienced that once with a conversion done in Adobe Camera RAW (I believe it was this image), and when I then tried the same process based on a conversion by DxO, it was dramatically better.

DxO reaches that level of precision by doing much of its work before de-mosaicing. One good example is noise. In all these images I begin with a straight conversion in DxO, using one of the presets. I use the Photoshop import plugin to pull it as base layer into Photoshop. Then I duplicate that layer, push local contrast using Photo Lift, multiply this layer with reduced opacity and normally also a mask, often apply a photo filter adjustment layer to fine-tune colors, strongly saturate with “Hue/Saturation” layers in various blending modes, apply levels and a contrast curve globally, may add further local contrast with a masked curves layer, and finally apply a vignette.

All that of course increases noise, even if the original image was taken at base ISO (which is true for all these images), and in case of the barns covered in snow, I even had to apply a surface blur to the Photo Lift layer, but still, when you amplify so much, some noise is inevitable, and then it is of utmost importa
nce how this noise looks like. The noise that remains in images converted by DxO looks … crystalline. Not like digital noise at all. It does not harm any image, but it actually enhances snow images.

Most of these images were made using the Sigma 10-20, a fine ultra-wide lens for DX format sensors, the one exception is the barn image, for that I used the Nikon 18-200 VR. Have you ever changed lenses standing up to your chest in snow and not knowing how to best get out of it? Funny, I can tell you 🙂

The Image of the Day was taken on the way up, at a height of about 1100 meters, at the same place where I shot the last of the series, on my way down, like all other images.

There is another lesson though. Just look one more time at these images. They all look slightly different, don’t they? Just rightly so, I’d say, given that they were taken in different places and over the time of about two hours. And still, they make no proper series. Remember my SoFoBoMo book? There I had shot all images in an afternoon, but it has taken me weeks to post-process them in a way that they had the same look.

All the images on this page look fine individually, but were I to make them a true series, for example a book, I would have to invest much more in visual coherence, probably giving up individual “truth” for the flow of the series.

The Song of the Day is “Let It Snow, Let It Snow, Let It Snow” from the 1992 Manhattan Transfer “Christmas Album”. Hear it on YouTube.

Nov 302008
 

When I posted “771 – And Winter Came” a week ago, I was met with mild mockery. I admit, there was not much snow to be seen 🙂

These two images are from Saturday, the first one from the morning. It had snowed from Friday afternoon all through the night, and when I went fetching breakfast in the morning, I made this shot, using the Nikon 70-300 VR at 300mm.

The image out of the camera had almost no contrast and definition at all. It was shot at ISO 200, the Nikon D300’s normal ISO, and it was exposed more or less “to the right”, thus I saw no reason to favor the precision of DxO Optics Pro over the convenience of Adobe Camera RAW, especially as I wanted to take the image to Photoshop anyway, but this turned out to be an error.

Mind you, this is a color image, but of course there was no color at all. Trying to push local and global contrast, I quickly found out that Adobe Camera RAW had introduced all sorts of tiny color artifacts, and pushing contrast greatly amplified them. Of course simply working with layers in “Luminance” mode would have hidden that, but wanting to fix it at the root, I re-developed the image in DxO Optics Pro, and, boy, did that make a difference. This is one of those areas where DxO shines: precision. There was a lot of fuzziness, that is to be expected when one looks through 500 meters of dense snowfall with an effective focal length of 450mm, but there were no color artifacts at all. Impressive. Read more about DxO in my ongoing series of posts that document my experiences with it.

The second image, the Image of the Day, was shot in the afternoon when something almost like the sun began to shine through the clouds.

This image is a rule-breaker. One of the things that Craig Tanner constantly preaches, is that in photography, two is an odd number, and that so often when one has the choice to use one, two or three of a kind of subjects, either one or three are preferable to two. The reason is, that two subjects of the same kind tend to look boring and symmetric, redundant in a way.

Well, here it works. We have two trees, symmetrically arranged, but everything else is asymmetric on multiple layers.

The other thing normally not to do is to have a fence or something like that in the foreground, running across the frame. It makes the image inaccessible, keeps the viewer out. Here we have something similar, in fact the edge of a pond, and although it works like a wall in front of us, the two trees provide a mighty and powerful door frame, leading the eye through and into the winter wonderland behind.


I took this image to DxO as well, trying different presets until I found one that would work as a good start. Then, in Photoshop, I copied the layer, applied Photolift (more about this useful plugin here) for a strong push in local contrast and put the result in “Multiply” mode with reduced opacity, added some curves layers with masks for local contrast adjustments, and strongly pushed saturation. A levels adjustment, a curves layer for global contrast, some vignetting, some sharpening, and that’s it. With the exception of saturation adjustments, this was a similar procedure to the one used on the snowfall image, only more subtle in the contrasts.

Update – Monday: I don’t do this normally, but last night I have processed another image of this series. I would have normally displayed it together with today’s post, but it really belongs here, to the other images of one marvelous winter day in Carinthia. Here it is.

The Song of the Day, “My! My! Time Flies!“, is another one from Enya’s new winter album “And Winter Came”. See a video on metacafe.

Nov 112008
 

A bicycle? I must be back to Vienna 🙂

Over the last two weeks I have spent quite some time with DxO Lab’s new version 5.3 of DxO Optics Pro. When I began processing today’s image, taken at f1.4, 1/60s and ISO 3200, I found that the demo had expired and I was back to Adobe Camera RAW and probably Noise Ninja. I reined in my first impulse to finally buy DxO and began working on the image with my usual tricks. I even uploaded the result to SmugMug, but the more I looked at it, the more I felt the need to go back and try it once again with DxO.

The result is, I did it, I am now owner of a license of the standard version of DxO Optics Pro, and I am glad that I did it. The result is clearly better than my first effort with Photoshop alone. I spare you another comparison on pixel level, fact is that I see a difference and that the difference is big enough to make me shell out the money.

DxO is not perfect. We have seen that Adobe Camera RAW plus Noise Ninja can outperform it under pathological conditions, and it also lacks in terms of workflow. Rather than acting as an import plugin, I’d like to see DxO being integrated into Photoshop as an alternative RAW converter, but all that is nit-picking. In practice DxO gives me the best RAW conversion that I’ve ever seen, at least for the Nikon D200 and D300. I probably won’t use it all of the time, but when I need it, it will be available.

The Song of the Day is “Try Some, Buy Some“, David Bowie’s George Harrison cover from the 2003 album “Reality”. Hear the album version and a live version on YouTube, and when you’re already there, why not compare it with the Harrison original?

Oct 262008
 

In Saturday’s post we saw that DxO 5.3 performs fine on the noise front, but it won’t boost your D300 to ISO 25600. Well, at least it did not boost mine, I suppose it won’t do yours either 🙂

In all that pixel peeping we have completely ignored what other things DxO can do. Today’s image is old as well. I have not shot a single image today, not even test shots.

This one was shot in the cloister of the dome of Magdeburg, Germany, in early June 2006. Here is the original JPEG out of the D200. It does not exactly show my abilities at their best. It is tilted by accident, the subject, the cross, is hardly visible at all against the strong back light. This is clearly a case when I would have had to use a flash. I can’t remember if I had been allowed to do so, maybe not, fact is that I didn’t.

The Nikon 18-200 VR, the only lens that I had at that time, is nice, but it shows strong barrel distortions (well, actually something more complex than simple barrel distortions) on the wide end, and I even had to point the lens slightly upward to get everything into the frame. Not exactly ideal for an architectural shot.

DxO handled all that nicely. The 18-200 is a supported lens, thus all distortions were removed automatically, along with chromatic aberrations. Using Photoshop’s “Shadow/Highlight” to lighten up the foreground would have produced halos, thus I would normally have used a curves layer with a luminance mask. DxO does this with one slider under “DxO Lighting”. Basically they analyze the image, automatically isolate regions according to tonal value, and then apply contrast and exposure to these regions independently. This works really well and is extremely easy.

For the correction of tilt and perspective distortions I have used a simple tool where you paint a rectangle into the image and then drag the corners of the rectangle to indicate the desired correction. You can immediately see the outcome in a second window. I have first seen this in Paint Shop Pro and always missed it in Photoshop. You can choose to automatically crop the result.

Well, that’s it for Sunday. This series of posts about DxO will go on as I discover things or find ways to demonstrate features that I like.

The Song of the Day is “The Cross” from the 2002 Blind Boys of Alabama album “Higher Ground”. Sorry, no video.

Oct 242008
 

These are images of Thursday. I took some in the morning and then late in the evening. All images were converted with DxO 5.3, but with the exception of the last, none was shot at particularly high ISO values.

Speaking of DxO, I’m just speeding through the documentation to get an impression of what a workflow with this tool could look like.

My first impression, just from working with the program without any consultation of the docs, was one of overall simplicity. There are two versions, a standalone program and a Photoshop/Lightroom plugin. I don’t have Lightroom, thus I can’t say how it’s activated there, but in Photoshop you get into the plugin via “File / Import”. Makes sense after all.

Once in the program, regardless which version, you are first presented with a file browser where you can select the images that you want to process. DxO is a batch/pipeline oriented program and employs a “Project” metaphor. A project can contain any number of images. Those get processed in a batch and the result is either that all images are opened in Photoshop (plugin) or get stored in a format of your choice, in the same or a different directory as the original (normally RAW) files.

As I said, I’m just now looking into the documentation. All images so far were processed in “experimental mode”, and I can say that the user interface is intuitive and simple. You begin on a “Select” screen where you include images into your project, walk through a “Prepare” screen where you can specify how each image is to be converted and then start the conversion on a “Process” screen. On my quad-core processor, processing was two images at a time, we’ll see what it does on a dual-core processor when I have installed it in Carinthia.

Speaking of multiple installations, the program has to be activated and activations can be transferred from one computer to another, but as far as I have seen, activation on at least two computers is permitted by one license. That’s quite OK. The idea is to have it on one Desktop and a laptop. We’ll see how that finally turns out for me, because I regularly use two desktops and a laptop.

DxO has one fully automatic mode, a lot of image presets, e.g. for slightly low or high key processing, one tuned for high ISO images, presets for different saturation/contrast combinations, etc, and of course you can set everything manually and save that as a preset.

Whatever you do on the “Prepare” screen, manually setting details or applying presets, it is always immediately displayed in a big preview that can be zoomed in up to 200%, thus you always exactly see what you do.

If you don’t set anything at all, the image gets converted in fully automatic mode, and what that does is usually quite OK. It may not be your desired look, but so far it has always produced a usable result. Automatic processing includes geometry correction for supported lens/body combinations. Some of my combinations are supported, some not.

All of the images in this post were converted by the plugin version using presets. Then I have added further processing in Photoshop, but usually not much.

What Adobe Camera RAW does and DxO seemingly not, is the automatic elimination of hot pixels. This is a bit of a bummer, but no worse than Capture NX.

EDIT: Sorry for the false information, by now I have found out that DxO in fact can automatically eliminate hot pixels, it’s only a tad hidden and not on by default.

One final thing that may be interesting from a workflow perspective: The standalone version of DxO can produce linear DNG files, i.e. DNG files that already contain a demosaiced image. These files can be processed by Adobe products and retain the flexibility of RAW files. I’m not really sure about the consequences, but this could mean that it would be possible to batch-convert all your files and e.g. apply only demosaicing and noise reduction. Anyway. We’ll see soon.

That’s it for today. This series of posts about DxO will continue as long as I learn useful things. Stay tuned.

The Song of the Day is “Just Another Day” from Brian Eno’s 2005 album “Another Day on Earth”. Hear it on YouTube.

Oct 242008
 

OK, finally there’s one image of Wednesday, almost plainly out of DxO 5.3. It’s ISO 1600 from a Nikon D200 with Sigma 30/1.4, a combination that is supported by DxO for automatic lens correction. I think you won’t be able to tell from this resolution (even if you click on the image for the higher resolution) how much better the RAW conversion is, especially compared to Adobe Camera RAW. Nikon’s own conversions are slightly better but still inferior. I shall post some examples as we go along with a series of blog posts that will end up as a sort of review of DxO 5.3, but upfront I can already say that the examples on their website don’t lie. The quality improvement is dramatic.

They achieve this by removing noise not after demosaicing but before, directly on the RAW data. This makes sense, because a conventional Bayer sensor uses a pattern of interwoven pixels, half of them green (where the human eye is most sensitive), and a quarter each red and blue. That means that the actual resolution of the sensor for green is half of what you’d expect from the pixel count, and for red and blue it’s only a quarter each. Demosaicing is a process of interpolation, where the full resolution red, green and blue channels are reconstructed from the actual data in the reduced channels and from luminance information from the neighboring pixels of another color. Scary, huh? That’s for precision.

The only sensor type on the market that does not use such patterns at all (some by Kodak and Fuji use different patters or variations) is the Foveon sensor used in Sigma’s cameras, although at the price of much reduced actual pixel resolution. Those pixels upscale well, but an image resolution of 2640×1760 (4.6 megapixels) does not look so sexy today. That impression is further marred by a light sensitivity that’s slightly below today’s standards.

But let’s get back to Bayer sensors and the process of demosaicing. We’ve seen that the actual image is reconstructed from greatly incomplete data. I’d even say much of it is invented. Of course this technology is not new and the algorithms are fairly mature, but it is also clear that at high ISO noise starts to spread out. When neighboring pixels are used to reconstruct the exact color channels per pixel, then noise in those neighboring pixels comes into the equation. This sums up, and the result are ugly blotches of color or strong de-coloration as a consequence of noise reduction algorithms.

This is where DxO sets in. They reduce noise on the original pixels before they get combined in demosaicing. Therefore the noise in one pixel does not spread out to neighboring pixels. What exactly they do and how they do it is not known to me, is most probably patented or secret, but the results are clearly better than everything else that I’ve seen so far, producing very fine grain without much loss of detail and without much loss of color at high ISOs.

It’s really like a camera upgrade, i.e. the improvement is about the amount that you’d expect by going from one camera generation to the next. The Imaging Resource has quite a lengthy interview where these things are explained and where a D700 image is shown, taken at ISO 6400, underexposed by 3 EV and pushed to 51,200 equivalent. Crazy? Yes, but surprisingly there is still usable output.

The Song of the Day is “Tell Me You’ll Wait For Me” from Ray Charles’ 1959 album “The Genius of Ray Charles”. Sorry no video, but Amazon has samples.