Wednesday, April 6, 2011


It finally smells like spring, and the sakura is flowering right on time too. We are hitting the brief, pleasant few weeks between the raw, wet cold of winter, and the oppressive summer heat. We're not doing hanami this year, but I took a couple of quick phonecam shots of flowering cherry blossoms at work this afternoon.


Cherry Blossoms, Ikoma.

One day somebody will release a phone with an actual good camera in it. One day.


mercen said...

And it's software that will make that camera good. Witness apps like SynthCam.

The days when the hardware had to be good are over. Many cameras rely on software corrections for imperfections in their lenses and I can't think of any reason why that trend would slow down, especially on smartphones, where apps are all the rage.

Whether that is a good or a bad thing is a common topic of debate in photography circles...


Fernando said...

The picture is not bad at are just too picky hahaha

Janne Morén said...

Well, yes and no. This is a big issue, and it would have been worth a whole separate blog post if I had been confident enough in my knowledge of the subject.

First, it's software that makes this camera bad; you're right that even small improvements in software would make for a much better result here. It overdoes noise removal (I have to actually add noise myself in post-processing), and does it with a flawed or incomplete algorithm. You need to bring up the black point after noise removal or you'll end up with lowered contrast, and this camera neglects to do that.

But software can only do so much. You can't improve information that is never there to record in the first place. There are real, hard, physical limits to how much resolution (for instance) you have available, depending on the size of the camera. For something the size of a smartphone cam the optical limit is around 5-6Mp or so. No amount of processing can improve on that.

So while software can certainly improve the current state of things, it will never make a tiny (1cm^3 - it's amazing, really!) phone cam unit perform like, say, a current mid-range 15mp DSLR. And no, stuff like stacking multiple shots doesn't count - you're trading resolution for lousy shutter speed (things have to basically stay where they are for the entire shot sequence).

Also, format matters. Even given the same resolution, lens angle of view and equivalent aperture, a small-sensor image and an MF shot simply look different. A 5mm lens and a 90mm lens will give you different results, same angle or not.

mercen said...

While I agree with you that a small camera will never do what a bigger one does, software can do things that seemed impossible. And yes, that includes stacking. That may not be an option with moving subjects but many subjects don't move.

Don't get me wrong, I'd rather have a great lens in front of a great sensor than replace them with a great algorithm (though I'd like a good algorithm to massage the data I got from that great lens and sensor ;). But I do see pictures from small cameras getting better and better (granted, partly thanks to advances in hardware) and I expect even phone cameras will give great results in the near future. Heck, they're already amazingly good when you take their size into consideration (which you've already alluded to in your reply).

Of course, Phase One and Hasselblad need not worry about it. But for common folks, I don't know how much longer a "real" camera will be needed.


Janne Morén said...

It is absolutely true that for more and more people their only camera is going to be their phone. I even blogged about it here. Many phone cams are good enough for some uses already. With a bit more development (mostly hardware, actually, and mostly a modest increase in unit size) they'll be good enough to largely replace the lower-end p&s market.

But this size cameras will never be "great". They just can't. They'll be "decent", and that will be good enough for casual users, which are most people.

For hobbyists the situation is different of course. Any improvement - in hardware or software - will improve cameras at all levels, so the goalposts for what is considered good are moved. Never mind that the 35mm cameras and films of the 1970's already produced images good enough for the vast majority of today's amateurs; we all still want more and better, just because we can, not because we need it.

Until, that is, we no longer want more. This happens all the time, in all kinds of fields. There was a period camera makers competed to make the fastest lenses. The peak was reached when Canon, Leica and others started making f/0.95 lenses - at which point the buying public balked. The lenses got too big and too expensive and the trend stopped.

And some people seem to think current films have reached some limit on resolution, grain and sharpness. They haven't, and there are films for sale today (like ADOX CM20) with much higher resolution than mainstream films. But they never caught on; you need excellent lenses, rigorous setup and careful focusing to make use of it, and for most people, and most subjects, it isn't worth it. Current mainstream films are "good enough".

And this will happen with any current trend in photography too. You may notice that the big "pro" 35mm full-frame DSLR's aren't exactly flying off dealers' shelves, and no maker is scrambling to convert their low-and midrange customers to full-frame. Phone cams will get better right up until people stop caring, at which point they'll stop. Which will happen - may already be happening - much earlier than enthusiasts may want.

Hmm, maybe I should have made thsi a blog post after all. ^_^

Chris said...

Nice blog. I used to live in Osaka. I can't wait to sit under my sakura tree this Spring.

mercen said...


I think we agree on all points, really.

Your mentioning the demands that come with higher resolution films reminded me of what I've read about medium format digital backs: the resolution in those cameras has gotten so high that photographers who use them are finding they need extreme caution when using them. They require extremely stable tripods, of course, but the lenses have to be exceptionally good too. Not only that, but the tolerances in camera mounts are apparently not quite up to the task anymore! I don't have personal experience with those but that seems to be the consensus amongst people who write about them on the net and it does make sense so I guess it is true.

Maybe there are a couple of blog posts hidden in all this indeed ;)