…rescuing this G10 shot of SF at IR, 1/1000s F4.5, 30.5mm on the lens, ISO80 @ 0eV evaluative…
Plenty of fine-detail, not much diffraction even at F5.6 with a 14MP 1/1.6" diagonal sensor…but not much color there, either. And the sky, sheesh I’ve seen that so many times with my G9. So uninspiring.
So I got the CHDK hack working with my new SX130 and while I couldn’t get it to shoot at ISO50 I did get raw (actually DNG) out of it so that’s a good start. Actually the camera jpegs are pretty decent, just a little overexposed/overcontrasted, which is not surprising. And the colors are definitely a little flat and gray compared to the C9100 at least based on my Best Buy test shots for the C9100. I still think that the SX130 has significantly better fine-detail with less flaring of the highights than the C9100, even in jpeg. And of course shooting raw most of those problems go away and then there is the lens-distortion and the noise. But giving it credit where credit is due, the camera does a good job of fixing lens-distortion and a decent job of dealing with noise. Where it falls down is in controlling the exposure, saturation and brightness if shot anywhere near 0eV evaluative on a sunny day. Much like the original version of this shot. [3 years of learned aversion to shooting digital cameras at 0eV evaluative led me to pretty-much waste many rolls of film when I started to shoot film. Shooting like this works well for film. It does not work well with digital cameras no matter how many dumb-asses tell you to "shoot to the right". That’s a very good way to ruin your shots, plus the camera and just about every raw-converter out there are going to push your shots to the right anyway. Until you figure out how to get them to not do that. Digital is not film. Do not shoot digital cameras like film…do not shoot film cameras like digital.
The SX130 shots look significantly sharper with less blur and much-better white-balance and much better headroom coming out of dcraw [I mean, really, they look quite good, very natural], but there is visible noise in the shadows even at ISO80 and significant lens-distortion, though it’s standard barrel distortion and easy enough to fix. The DIGIC4 processor in the SX130 does a great job of cleaning up the LD but in removing the noise (and surprisingly mildly) it adds significant blur (simply because there is so much noise even at ISO80) then sharpens to compensate. And it does a good job overall. Not great but good. Given a low-enough exposure the shots look quite-decent. If you underexpose enough and you don’t have a problem with the shadows, and you didn’t know what was there before the NR step, you’d probably be perfectly happy with the output, at least for a Canon. It’s actually not bad, clearly one of the better point & shoots that I’ve ever shot, much better than the A710 with the same-size sensor and fewer MP, with more headroom than the A610 and certainly better IQ than the A650IS, and, LOL, it certainly runs rings around a whole bunch of Panasonic and Nikon point & shoots that I’ve owned and is not hugely different from the G9 itself but without the lens limitations of the G or A ranges. I’d say that it’s more like a midget Rebel with middling IQ, decent AF and a decent superzoom than a p&s.
In my opinion the NR is the difference between the SX130 and a credible D-SLR/superzoom combo. The G10-G12 lack even the superzoom. If you’re going to give up the lens, you *have* to give up the NR. In 2011 especially going forward a camera with a 4x zoom simply cannot have a slow lens on top of that.
The only thing that makes such a short-throw lens worthwhile is that it’s fast, allowing the shooter to stay at or near the lowest ISO at all times (without resorting to the use of a tripod).
Otherwise you have to crank up the ISO or crank up the resolution which effectively does the same thing, then start with a real image and run NR on it and that means you end up with a "digital photograph". NR is where the mice scamper while the men walk down the road. One of the main reasons that I still shoot film is that it does not require NR. Yes, I’m sure that SF looks something like this, but I’m sure that it looks a lot better in person. The difference is that due to NR and low SNR both this is just not sharp and the colors are not natural. It’s well-sharpened and the brightness, gamma and contrast have been adjusted to give it a good overall dynamic-range.
There is a difference between the two.
And you can see that very clearly in this result.
This is why people still shoot raw, still shoot film and still spend $1500 for a good camera and lens. Yes the G10 might be light and easy to carry, and yes it might be much easier to deal with jpegs than to go through the hassle of raw-processing. But you won’t drive hours from home and then trudge 2 hours up a hill and take shots of great scenes like this and come home and pop the SD card into your computer and get results out of it like this (and the original jpeg was WAY worse than this) too many times before you stop even trying. That happens enough times you’ll be happy to spend good money on good gear, happy to buy heavy gear & carry it around, happy to buy a real good point & shoot that can shoot raw…and then you’ll shoot raw all the time…and then you’ll be happy to shoot film.
Because the main problems here are the Bayer-blur, the dynamic-range, the noise and the color. Not the resolution. Sure a slight misfocus will make it worse, a little diffraction will too. But diffraction is not the problem here, especially not at 4×6". It’s the noise, the blur, and the color. And to some extent the DR because to keep the buildings and sky under control yet still give it good overall contrast I had to drive the shadows way down.
30 minutes to write all that up about this one shot. 3 years of experience. Your eyes can see it and your brain can feel it but without the technical details it can’t connect what it sees to how it feels as a result. It can only say, "this just doesn’t look good". And your wallet just can’t fix that. You cannot buy a camera that is really capable of taking really good shots all by itself unless it can stumble across a scene and a set of lighting conditions and a lens which matches up well with its hardware capabilities and with the settings that an expert photographer would use, that’s been programmed into it. But obviously they can’t do that for every scene and set of lighting conditions out there not to mention for every budget. Sure, you can spend $5k on say a 5DMk2, a Canon 24-70 F2.8 and a Canon 70-200 F2.8 EF IS or the 28-300 EF-IS and get a whole lot of good shots that way, but that’s not only prohibitively expensive it’s also a hell of a lot of gear to carry just to take some good shots. You’d be a nut to buy $5k worth of gear to take this shot, I’m sorry (and if you got it free through a tax-rebate, then welcome to the Land of Opportunity). To cross that bridge you have to master (or at least get good at) post-processing, raw-shooting and probably film as well. Because you won’t know what you’re missing by shooting only digital until you master film, just like you *definitely* won’t know what you’re missing by only shooting jpegs. It’s like the difference between having a wheelchair that can take you 10 miles in 2 hours vs being able to run 10 miles in 2 hours on your own legs. Trust me: producing really good pictures is way more than just buying a decent camera and taking it out shooting. Any dope can point a digital camera at a scene like this and take a shot. Try producing a good-looking shot of a scene like this. It ain’t easy. But these guys should be embarassed to have this up on their site as a "demo". Unless, I don’t know, maybe if you’re trying to sell DSLRs or something.
The art of photography is not necessarily in getting a great shot of a great scene or an "interesting" scene. It’s in taking the mundane and turning it into art by capturing and presenting it in a way that makes it look at least as good, and interesting, or appealing…as eye-catching and mesmerizing as it does in real life, hopefully even more so. And if you can do that with a *great* scene, an "interesting" scene, then you’ve got a true work of art on your hands. These guys have used a $400 camera to turn a good scene into dogshit like it’s no big deal whatsoever. And they call themselves "running a photography enthusiasts’ website". I guess they are saying that if you’re cheap enough to only spend $400 on a camera then you shouldn’t expect shots that are better than that but I would *never* put my name on something like that if I could do so-much better with just a little effort. I’m tempted to fly out there and take this same shot with film and my $200 500si and Tamron 28-300 just to see what the fuck is up with it. The thing that I love about film is that not only is it cheap and good gear is cheap but if you’re even halfway-decent as a photographer and in post-processing, you are going to get good results as long as your film isn’t hosed in development. You don’t have to spend five hundred dollars to get a shot that is far better than this.
…if your solution to your IQ problems is to spend a ton of money, you’re doing something very, very wrong. If you feel that you need to spend a ton of money to get good shots, you’re making major mistakes somewhere between your shot-selection, your technique and your post-processing. I would say that if you think that the answer is to spend X, then you need to spend X/4, maybe even X/10, and work on improving your game to make up the difference. You want to take this shot really well, go ahead and get a G10 now that they are well-used and sell for about $200-$300 on eBay, but learn how to shoot it raw and post-process the shots right. You should be able to easily do better than this with the original raw files and not just some idiots’ blown-out jpegs.
You want to really learn something useful, get a good used SLR and a Tamron 28-300 to fit it, some decent ISO200 color negative film, a Promaster multicoated UV filter and strap it around your neck, march out and take this shot, get the film developed at CVS for $2/roll and buy a good Epson V300 from Amazon for $60 delivered and scan the film and post-process it yourself until you can beat this shot. If I can take the poles and cars and grass and crap out of this by hand, then you can deal with the dust in a film shot like this. Or go ahead: spend $2500 on the camera and another $700 on the lens just to get a good shot of SF from the top of a hill. And when that still doesn’t happen, maybe you can try a pano, or an HDR tool, or a 64-bit copy of Lightroom, or maybe a $10k Pentax 645D and a $1500 35mm D* lens. I swear in the next year or two I’m going to drag my tired old shit out to take this shot with film and post it with a link right here. And if by the time I’ve worked on the scan for 5 minutes I haven’t beaten that same-old faded washed-out pinkish fake-looking digital "red" on the buildings in the afternoon sun, not to mention that noxious pink-blue sky, I’ll eat the fucking lens-cap. With all that detail in the foreground this would be a piece of cake with film, at 4×6" you wouldn’t even notice most of the dust.
"No, it’s gotta be digital, it’s gotta be digital, it’s gotta be digital…who cares if it looks like utter shit, it’s gotta be digital…"
I ask you, which one looks worse, this one or this one? Do I get paid to run around taking shots with new cameras? No. Am I an experienced pro? No. Was I even an experienced AMATEUR when I took this shot? No. I barely knew what a fucking DSLR even *was*. And only because I actually had one. I wasn’t a goddammed "pro photographer and hardware reviewer". And I would *still* beat shots like that. Even *I* could take a decent sunlit shot. I wouldn’t put up work on my pro website that I would be embarrased to show to a 5 year old. "Hey look, Johnny: this is how Daddy earns a living to support you and Mommy". My kid would look at that and go running from the room in terror.
…actually this looks pretty decent at 4×6" on my monitor 🙂