And if you could switch off preflash on the K10D, you could do this with any old gun on a three quid slave adaptor

Well FWIW, and I expect it's too late... but I preferred the second one for composition. I didn't like the way the foreground path in the first led the eye to nowhere, nor the way the tree emerges rather randomly without a trunk.

And I preferred the first one for the burnt out sun. Any time you see the sun as a pale disc (or random egg shape like in the others), it's very low, and you're looking at it through thick haze.

Other times, guess what - you look (albeit very briefly) at the sun, and it IS burnt out (and then your eye is).

I obviously haven't got the hang of AF yet then.

I find pointing the camera accurately enough for the AF to see the bird (obviously, only if it's flying... ) and focus on it (this is at a minimum of 200mm) harder than I do focusing manually.

If I miss it a bit, the AF hunts, the whole thing goes way out of focus and I can't see the bird any more either. Much easier so far, to follow-focus as I pan.

Any hints for better AF technique?

Good grief. Ok, so I chose a bad example maybe. But 'replaced' in this sense doesn't have to mean 'replaced in every instance'. And if I'd said "replaced in the consumer camera market" you'd probably have been less exercised over it.

CDs have replaced vinyl albums; transistors have replaced valves; fuel injection has replaced carburation - but doubtless you'll find a few that still use the former in each case.

The point is, technology moves on. And at every evolution, you find those that are resistant to it, not because of genuine objections, but because of an emotional attachment to the technology they've grown up with. This isn't even all that unreasonable - it's a pain to have to keep learning new things sometimes, especially if it seems that what you already have does the job fine.

Of course we will, if they're compatible with our M-series lenses...

I think we will not.

I wonder if there was the same sort of debate when 120 roll film was gradually replaced with 35mm.

But you don't have to look at every frame out of the millions there are to find the one you want. You know roughly where it is, and then looking through the 100 or so frames that are relevant can be very quick. I scrolled through some of the bursts of shots I took today in Picasa, which I use for previews. It's very quick. Video wouldn't be much slower, given the right tools, and these are evolving all the time.

It depends what you mean by 'near'. Something between two years and twenty I'd guess. The quality of video I can get from my little Sony still camera is not great, but it's not terrible. I wouldn't use the video to extract stills yet... but five years ago I wouldn't have bought a digital SLR at all.

I think your view of what is a video camera is too narrow. Imagine a K10D, able to take 100 frames a second. Now with all your creative skill, set up a fantastically composed and lit portrait. Press the button a shade earlier than you would have...

... now you have 100 frames, and you can pick the perfect one.

I have an emotional objection to the way all the unskilled muppets will jump on the bandwagon, since it will seem that all the years I've spent acquiring the skill to get the one frame I want will have been wasted. But even though I'm better than I was, I still lose good ones because my timing is imperfect. It will remain the case that the best photographers will get better results than the ones that point and hope.

It won't create the great results in itself. But it might mean missing fewer great results for the sake of pressing the button a fifth of a second late or early.

I think the difference that the technology makes is that it removes, or at least reduces, the need to be a technician as well as an artist. In the middle of the quality spectrum, there will be those that attempt to use the technology as a replacement for the art - I've criticised some of PeteKD's pictures on that very basis. But at the top end, the best artists will still create the best art.

Maybe I'm not understanding you. What is this human bottleneck?

In that article, all that cobblers about seismic this and that is just theatrics that can be ignored, but that aside, all I see him saying is that video will be increasingly used to acquire the raw material from which stills are extracted, and that as the technology improves, it'll get easier and better.

And all I'm suggesting is that if the quality of the video is very high, and the tools available for the extraction of the frame you want very quick to use, more people will use that medium, in order to avoid having to acquire the skill required to press the button at just the right moment.

I took lots of pictures at the London Marathon today. They were just snaps, aimed at capturing a friend as she ran. Mostly because of people in the way, I couldn't even look through the viewfinder, since I had to hold the camera up high. I used the K10D's continuous mode, and took about 150 shots, of which about a dozen were any good. If I'd had video, I would have had the choice of a lot more moments from which to select.

I don't see the article as any much more profound than that. If you want to take beautifully lit and composed portraits, or landscapes, the compositional skill will be just the same, whether you use the future high quality videos as the source of stills or a DSLR, but even then, with video, you might stand a higher chance of capturing that perfect expression, or that perfect moment of light as the sun sets.

Well I didn't take the article as saying that the need for stills will go away, just that the means of acquisition - in photojournalism particularly - will change.

The technology isn't irrelevant at all, since it lets you do things quickly, once you've made a decision what to do - and the decision support available is increasing all the time. We already have automatic red eye removal and face recognition - it's only a matter of time before they'll be able to apply standard optimisations to video, just as Photoshop lets you do Auto smart fix, and have them run blindingly fast.

Try telling a video editor that can remember physically splicing film strips that the technology is irrelevant!

Agreed. As I said, the time taken to view video will not change (until we have bionic brain implants that work like coprocessors, and allow us to assimilate information at different speeds). So the need for stills won't go away in the forseeable future. I don't think the article was suggesting that it would.

But just as darkroom skills have given way to software skills, when it's possible to extract the perfect high quality still from a piece of video, the skill required of the photojournalist to perfectly anticipate the moment will be lessened.

People were saying that digital would never be good enough to replace film and not all that long ago at that.

The need to have stills will probably be with us for the lifetimes of all those reading this forum. But the means of acquiring them and displaying them will continue to evolve. And when the 60fps, 30 Mega pixel per frame, auto-multi-element-metered auto-smart-fix video camera is the same price as a K10D, people will be buying them and using them. The skill needed to press the shutter-release the moment before the ball hits the racquet will become a thing of history.

Wasn't 64K more memory than anyone would ever need in a personal computer?

I think it's the time to view, not the time to make, which is the really significant thing here, and if you can pull (and quickly enhance) a few really high quality stills from a video, even this objection disappears.

Not that I've ever done it, but the darkroom skills of dodging and burning are redundant now, in the context of layers and HDR.

To say what technology will or won't make possible (for which, read 'affordable'), is a little premature, methinks.

And in any case, quality is not related to how many people want to see it. The quality of most of YouTube is truly awful, but...

My apologies. I didn't realise you didn't already know the technique, I thought you were trying it and it wasn't working.

It's exactly as Don says.

Well, almost. The MF lens needs to start a bit out of focus, not in focus. Otherwise it'll fire straight away.

The point is that when the camera's set to MF, it'll release whenever you press the button. When set to AF, it only releases when it thinks it's in focus.

Incidentally, you have to be moving the focus ring very slowly, otherwise you'll have moved it a bit too far by the time the shutter fires.