But sometimes you don’t have to go very far. My cat was sitting on a wall just outside my kitchen door, with the snow-covered garden behind. She didn’t have too much trouble contrasting with the snow and she threw in a mildly evil look, directly at me, as if to complain about having been left out in the cold.

I added one stop to avoid the meter being fooled by the predominantly white tone of the image, but white balance is always an issue with images of snow. There is always a blue cast. The question is how far to go in removing it. In the Lightroom histogram, the snow element of the picture came up as three distinct curves for the red, green and blue components. The curves were similar shape and peak height, red to the left then green and blue on the right. By warming up the colour balance the curves can be made to coincide reasonably well at which point the blue cast has gone.

This is what I did with the cat picture but wonder whether I may have gone too far. The overlap of the curves was not 100% perfect and we get most of the snow coming out pure white but some areas faintly pink and others faintly blue. Maybe I should have left a mild blue cast as I don’t really think there should be any pinky bits in the snow.

Here’s another crack at it. I’ve been less aggressive with the blue cast removal and taken the exposure down a bit to bring out more detail in the snow.

Last weekend I had an opportunity to put the newly released DxO Optics Pro 6 through its paces at my son’s graduation in Oxford. No use of flash allowed in the Sheldonian Theatre, but no need to worry because I could just ramp up the ISO (even on my trusty Sony Alpha A100) to say 800 and DxO would magically make it look like ISO 100 at noon on a summer’s day.

Except it did no such thing. It dealt admirably with the noise but threw in some ugly random artifacts and produced some unnatural “watercolour” effects, particularly on fine detail of faces and hair. I had previously experimented with DxO on pictures of things, not people. It may be alright with things, but definitely not with people.

Here is a 100% crop of an example produced with Lightroom:

And the same crop with DxO:

I wondered whether it might just be teething problems with the new version, so uninstalled the DxO 6 trial and reinstalled the older DxO 5.3.5. No better, I’m afraid.

In the end I went back to Lightroom 2.2. It isn’t quite as good as DxO at banishing noise but produces a far more natural and artifact-free result. Lightroom wins hands down.

Pass me a handkerchief, someone. Sob.

One such gem is DxO Optics Pro from DxO Image Science in France. Having only recently come across it (in photoclubalpha.com) I have given it a whirl. No longer do I believe that RAW conversion is a mature, commoditised technology. DxO simply does a better job of the basic RAW conversion process than the market leaders. It doesn’t have the polished interface or breadth of features, but it produces tangibly better pictures, particularly at higher ISO ratings. Now isn’t that what really counts?

Using DxO for image processing is easier, quicker and produces better results. It’s a different way of working compared with say ACR or Lightroom. With LR you have a lot of decisions to make and a lot of sliders to fiddle with. As against that, DxO is rather more “intelligent” and does most of the work for you. It still has all the sliders but has dedicated modules for particular combinations of camera bodies and lenses with oodles of data about how they perform at various combinations of focal length, aperture, etc and uses that to set the sliders in the optimum default positions for you. Most of the time it is best just to leave the sliders alone. Job done better and with a fraction of the effort. Maybe less attractive to compulsive pixel peepers and settings fiddlers – but just dandy for people with lives to get on with.

Version 6 promises to be a significant improvement, taking the high ISO performance even further. From DxO, I will accept nothing less than miracles.

It came out quite well given I had to resort to ISO 800 (on a Sony Alpha A100) with a shutter speed of 1/50th at a focal length of 120mm. The camera’s built-in image stabilisation just about kept it sharp, but the pony’s head was moving so it blurred slightly.

I still think the picture works quite well, helped by having a dark non-distracting background.

Well here is the best I could get out of Lightroom 2.2 with a particular ISO 1600 test image:

And this is the result with DxO Optics Pro 5:

Here are some 100% crops, first Lightroom:

And then DxO:

I find the DxO easier on the eye without sacrificing detail. The noise is not as prominent or obtrusive, and the overall result is a little more contrasty than I could achieve with Lightroom.

Something to think about.

While there they collared the representative from Sony and asked the question that everyone has been asking.  How is it that, when the likes of Canon and Nikon are bringing out new video-capable DSLRs every other week, Sony for all its pedigree in consumer video can contrive to bring out stunning new cameras that pack in every conceivable feature bar video?

This is an excerpt from quesabesde’s article and my translation below.

“Sin ánimo de resultar cansinos con este tema, una vez repasada la apabullante lista de prestaciones y novedades de la nueva pareja Alpha no dejamos pasar la ocasión de preguntar a uno de los representantes de Sony por la ausencia de la función de grabación de vídeo en el cuadro de especificaciones de los nuevos modelos. La respuesta no nos sorprende demasiado. Sony es una compañía de vídeo -nos explican- y cuando dé este paso será para ofrecer una prestación sin limitaciones en el enfoque, los controles manuales y demás parámetros. Esperemos que la espera -valga la redundancia- no se prolongue demasiado. En todo caso, nos preguntamos si esa promesa de futuro no afectará al presente de estas réflex.”

“At the risk of becoming tiresome on the subject, after reviewing the mind-boggling list of features and innovations on the two new Alphas we couldn’t pass up the opportunity to ask one of the Sony reps about the absence of video recording functionality. The answer didn’t come as much of a surprise.  Sony is a video company, they told us, and when they take that step it will be to offer a feature set without limitations on focusing, manual control and other parameters.  We hope that the wait, and we think it’s worth restating this, will not be too long.  In any event, we have to ask ourselves whether that promise for the future might not have an impact on these new DSLRs right now.”

So there you are.  You can have it when it’s perfect.  They have standards to live up to and can’t quite get there yet with a DSLR. It’s a technology thing.

Well, it certainly doesn’t make sense as a marketing strategy thing.  Sony have a very long way to go to get anywhere close to the scale of dominance that Canon and Nikon now enjoy.  Getting established Canon/Nikon shooters to switch to Sony is going to be a very hard sell. Even where Sony models match or beat the spec for a lower price, those diehards will already have made a huge investment in proprietary glass, and in any case Sony’s lens range is still fairly sparse.

Sony have quite clearly realised this and are trying to build up DSLR market share by turning their attention to photographers just making the step up to DSLRs from digital compact (point and shoot) cameras.  They are hoping to attract anyone who loves their point and shoot but wants better image quality and interchangeable lenses without necessarily having to become a photo geek.  So they produce DSLRs which do not take former compact camera owners too far out of their comfort zone.

The problem for Sony is that these same people have got rather used to getting video out of their point and shoot cameras.  The lack of video on the Sony DSLRs will be a showstopper for a large proportion of them.

The other possibility is that internal politics at Sony may be holding things up.  But that is pure conjecture on my part.  At senior executive level, Sony should be able to work out that they would be better off incorporating video now, even if it’s not 100% as they would want it, because they would sell a damn sight more units now and that would pay dividends quite quickly in terms of establishing long-term market share.

Maybe that’s not quite true. I have come across the suggestion for example that Aperture beats Lightroom for RAW conversion of Sony A900 images.

Going much further with the notion that Lightroom (which shares its RAW conversion engine with Adobe Camera Raw) does a rotten job with Sony DSLR images is this recent post by Photoclub Alpha’s David Kilpatrick.

David is speculating that the loose tie up between Sony and French camera software company DxO labs may go further than first thought. In any event, he points out that the DxO Optics Pro RAW conversion program seems to do a far better job on high ISO Sony images than Lightroom/ACR or indeed anything else. Now this may not be just a Sony thing. Perhaps the science of RAW conversion has not been fully tamed after all, and DxO have come up with a superior conversion engine full stop.

I will in any event be trying out the DxO software. I can think of a number of noisy images which might benefit from better RAW conversion. It must be better to get the best possible conversion first, rather than rely entirely on Noise Ninja or similar in a subsequent step.

So I will give it a try and post my findings. Should be fun.

Here’s the first:

The metering system has compensated for the bright disk of the sun by underexposing the rest of the image, so the sea and some of the cloud have come out close to black. The overall effect looks unnaturally dark, dramatic and maybe even a little bit spooky. The dark strand of cloud across the sun like the wing of a giant black bird … at this rate I’ll be imagining the spirit of Sauron rising up from Barad-dûr.

And here’s the second:

Taken from the same spot two minutes later, but with exposure compensation of +2 stops. It now looks much more like a “conventional” sunset, preventing the blacking out of the sea and lightening the areas of cloud, revealing some cute formations that are hidden in the other shot.

Personally I prefer the first picture. It’s a bit different and has more impact, even if technically underexposed.

Most people will have come across a striking sunset at some point in their lives and tried to capture it on camera. It is frankly impossible to capture the image as seen with the naked eye. The camera cannot cover the dynamic range. The best you can do is try to end up with a dramatic picture, whether it is faithful to what you saw on the day or not. Achieving that lesser aim is partly down to technique and partly down to luck. You can put yourself in the right place at the right time of day but it is really down to the weather conditions at the time. Either the setting sun is obscured or not, the cloud formations create interesting effects or not.

Beyond that, you really need a tripod so you can use slower shutter speeds and a longish lens helps to get the sun larger in the frame. Focusing should not be that difficult. The rest is experimenting with exposure and a modest amount of fiddling in post.

The geotag bonanza is not limited to iPhone users. Modern Windows Mobile phones and other platforms all have the same capability, if not the surrounding hype.  And we are starting to see GPS units incorporated into point and shoot cameras, for example the Nikon Coolpix P6000 which geotags pictures automatically before they ever leave the camera.

But what about DSLR users? When do we get in-camera GPS? And why would we want it?

I for one don’t see it as a gimmick. For people who take a lot of photos on their travels it can be hard to keep track of exactly where you took every photo. It’s not as if the Eiffel Tower will be in every shot. For example, I took the picture below on a Far East cruise two years ago.  I shot this from the deck of the ship as it was making its way from Nagasaki to Osaka, meandering around the Japanese islands.

I was fascinated and bemused by the unexpected sight of a mock-up of Venice. I’d love to know what it was all about and where this was.  But I don’t have the details of the route the ship took, so my chances of finding out are probably slim.

GPS-enabling DSLRs would be such an obviously useful thing, but as of today there are none.  The closest we have come to it is Nikon’s  GP-1.  This is a stand-alone GPS unit which attaches to the hot-shoe of the D90 D3, D300, and D700 cameras and adds automatic geotagging of photos.  Nothing at all yet for Canon, Sony et al.  So what to do?

Purpose-built geologging devices

Well there are dedicated devices for geologging, that is keeping a log of your geographical position by taking regular GPS readings and writing them to a file, along with a corresponding timestamp. The idea is you later use suitable software to match the photos to the location data, by reference to their respective timestamps, and the EXIF data updated to include the geotag.  There are also fancy geologging devices which accept an SD card with your photos on it and add the geotags automatically.  But such wonders do not come cheap.  This one from ATP is about $90.

Windows Mobile phone used as geologging device

I wondered whether I could do the whole thing for free by using the GPS unit built into my Windows Mobile phone, instead of buying a dedicated geologger.  My phone (HTC TyTN II)  certainly includes GPS in the spec, but did not come with any satellite navigation software and I had never bothered with any, so the GPS unit had probably never even had juice through it.  Still, maybe it had not totally atrophied from months of being ignored.

I hit upon a free application, GPSToday, which is essentially a clone of the Google map application for mobile phones. Crucially it has a geologging function.  You can set it going and, as you walk around, every 30 seconds it writes your location to a file.  You can change the frequency of readings or let it vary it “intelligently”. I guess that means it takes more readings when you are moving about more.

This sounded desperately promising, so I installed it and got it working.  It works perfectly, happily latches onto legions of satellites when you are out and about, and on request records your travels in a text file (I elected to call it geolog.txt) in the following format:

time,latitude,longitude,altitude (feet),speed (mph),heading
06-16-2009 17:14:48,34.361445,21.317156,440,0,0
06-16-2009 17:14:59,34.361486,21.317144,407,0,0
06-16-2009 17:15:10,34.361502,21.317144,404,0,0
…

That’s the first snag.  It’s just a flat file, not in any recognised format for location data such as GPX or NMEA.  How could I expect it to be recognised by any suitable photo geotagging software?  GPSToday assure us that a future version will output in GPX format, but for now we’re stuck with the existing non-standard format.

Getting the location data in a recognised format

Enter GPSBabel, a free software application that translates location information between a myriad of formats (including all the proprietary SATNAV manufacturer systems) and even “custom” formats such as the one used by GPSToday.  Mind you, it is hardly straightforward.  You have to use the “xcsv” mode and create a “style file” to tell GPSBabel how to interpret the non-standard location data.

This is how I did it.  I created a folder on my C: drive called GPSBabel and in there put all the files downloaded from the GPSBabel site.  I also copied the location data file geolog.txt from my phone to the GPSBabel folder. I then created a text file called GPSToday.txt, in the same folder, into which I inserted the folowing text:

# Format: GPStoday

DESCRIPTION  GPStoday
EXTENSION               txt

#
# FILE LAYOUT DEFINITIIONS:
#
FIELD_DELIMITER COMMA
RECORD_DELIMITER NEWLINE

PROLOGUE time,latitude,longitude,altitude (feet),speed (mph),heading
IFIELD ISO_TIME,”",”%s”      # time
IFIELD LAT_DECIMAL, “”, “%f” # Latitude
IFIELD LON_DECIMAL, “”, “%f” # Longitude
IFIELD ALT_FEET,”",”%.0f”    # altitude
PATH_SPEED_MPH,”",”%.1f”     # speed
PATH_COURSE,”",”%f”          # heading

I saved it and changed the extension to .style so the file became GPSToday.style.

Now I should explain that I did not arrive at that style file at the first attempt.  Most of it wasn’t too bad but the time field setting was a nightmare.  I tried just about all the options in the GPSBabel help system and none worked.  GPSBabel could not understand the date/time data in the format output by GPSToday, no matter what I did.

In the end I decided to use the ISO_TIME option and run a script to rearrange the date/time information into a format that would be understood by GPSBabel in conjunction with the ISO_TIME setting.  In practice, this meant writing a bit of Visual Basic code.  For convenience I did this from Excel.

This is the code:

Sub FixTime()

Dim TextLine As String
Dim LineBuffer As String

Open “C:\GPSBabel\geolog.txt” For Input As #1
Open “C:\GPSBabel\geologISOTIME.txt” For Output As #2

Line Input #1, TextLine
Print #2, TextLine

Do While Not EOF(1)
Line Input #1, TextLine

LineBuffer = “”
LineBuffer = LineBuffer & Mid$(TextLine, 7, 4) & “-”
LineBuffer = LineBuffer & Mid$(TextLine, 1, 2) & “-”
LineBuffer = LineBuffer & Mid$(TextLine, 4, 2) & “T”
LineBuffer = LineBuffer & Mid$(TextLine, 12, 8) & “Z”
LineBuffer = LineBuffer & Mid$(TextLine, 20)

Print #2, LineBuffer

Loop

Close #1
Close #2

End Sub

On running this VB macro, a new file, geologISOTIME.txt is created in the same folder, with the same data as geolog.txt but with the date/time format changed so that it would be acceptable to GPSBabel.

You can then run GPSBabel, by double clicking the file GPSBabelGUI.exe.  You have to point it at the input file, geologISOTIME.txt, and specify that it is in xcsv format and also specify the style file, GPSToday.style.  Give the output file name as say geolog.gpx and specify the gpx output format.  Then run the translation process to get your location data into a standards compliant format you can use with a photo geotagging program.

Applying the geotags

The final step is to feed both the gpx file and your photos to a program which can do the timestamp matching and write the geotags to the photo files.  I picked a free open source program called GPicSync.

It is fairly self explanatory in use.  If you managed to cope with the format conversions then running it will not be too taxing.  It creates new versions of your photos, with the geotags written into the EXIF data, and saves the untouched originals in a backup folder.  Job done.

Proving the geotagging has actually worked

As an independent check I downloaded Microsoft’s Pro Photo Tools 2 suite, again free.  If you open one of your geotagged photos in that application you can use the built-in map function to locate where the geotag is saying the picture was taken.

I did that and was amazed to find that my pictures were all correctly geotagged.  Having done the hard work I now have a proven workflow for geotagging photos taken with my DSLR, and it did not cost me a penny.

Using the geotagging solution in practice

It’s simply a matter of making sure you take your phone on your photo trips and the geologging function is in operation from before you take the first picture in a batch, and until after the last picture in the batch.  You can keep starting and stopping the logging function.  You don’t have to change the logging file name, GPSToday just appends new time/location records without overwriting.

It is important to make sure the GPS is logged in to some satellites before starting shooting.  Best to turn the GPS on in persistent mode (see GPSToday help) and just leave it on for the day.  And check periodically it is still locked on.  You don’t want to be caught out by finding the GPS is off or not latched onto a healthy bunch of satellites as you are about to start taking pictures.

One important point: time zone.  The workflow will only work if the timestamps on the photos and the phone are in sync.  Make sure, particularly if you are travelling abroad, that the time settings on both camera and phone are in sync.

Oh, and there’s this photo which may or may not be photoshopped.

Apart from the 17.2 MP resolution (obviously) and the claim it supports live view, there is little else known. Not even whether it can shoot video.

It has been suggested that Sony would be reluctant to offer a DSLR that shot video for fear of damaging their valuable camcorder sales. Well Canon sell a fair few camcorders themselves and that hasn’t stopped them.

No, I think the video-enabled DSLR and camcorder play to quite distinct markets and offering a video-capable DSLR would overall be positive for Sony’s revenue, even though the CEO might need to twist a few arms at the camcorder division.