Wikipedia:Picture peer review/Gants Hill stn interior concourse.JPG

Gants Hill station interior

Strong visual impression and good example of London Underground interior architecture

Nominated by: Sunil060902 (talk) 13:58, 14 July 2008 (UTC)[reply]

Comments

Interesting composition, but looks quite noisy. --Fletcher (talk) 14:24, 14 July 2008 (UTC)[reply]
Lighting is perhaps not ideal. Strong shadows in the centre of the picture - the people, and are they light posts, are very dark. As Fletcher says, also seems noisy and also appears pretty soft. I also get the feeling it's tilted (look at the tiles along the bottom edge). I do like the smaller versions though. --jjron (talk) 08:49, 15 July 2008 (UTC)[reply]

Unfortunately not permitted to use flash on the Tube. Would the original have benefited from higher resolution? This was on "2M" (1600 x 1200). Might be tilted but only marginally (the ceiling spans seem in synch). best, Sunil060902 (talk) 14:38, 15 July 2008 (UTC)[reply]

It's more a question of image quality than size. I think the shot would benefit from a lower sensitivity, which entails a longer exposure. At that focal length you may be able to do so without needing a tripod, but you might want to find a location where there is more ambient light to obviate the flash, or you will get those shadows. Acknowledge it's a tricky shot. Fletcher (talk) 16:42, 15 July 2008 (UTC)[reply]

Basically agree, but to get back to the question of size, all I can say is generally use the highest settings you have available. I'm not sure of the terminology used on Fuji cameras (i.e., 2M), but it sounds like what you've put up here is the 'fullsize' image, but you are able to shoot at considerably higher res. For sure shoot at a higher res; you can always crop and downsize later, but you can never add back in what you've lost by shooting at lower res. Also agree it sounds a tough shot - a longer exposure to get into the 'shadows' is likely to uncomfortably overexpose the lighting in the roof. I'm also guessing that you'd need a tripod to get a long enough exposure without blurring, but if flashes are banned, I'm willing to bet tripods are too. Hmmm... --jjron (talk) 09:35, 16 July 2008 (UTC)[reply]

Yeah I can feel my heart palpitating just thinking of it. If you don't get mugged you're going to get shot by some paranoid guard who thinks photographer = terrorist. ("Everybody down -- he's got a 'pod!!!") On a more optimistic note, isn't the rule of thumb one over the focal length for the maximum exposure time? And his focal length was only at 8.8mm. So I was thinking maybe he could get down even slower than 1/105. Or maybe not. That's a good point you make about overexposing the ceiling; I wonder if that is the kind of thing where you could use a graduated ND filter.... Hmmmm, indeed. Fletcher (talk) 23:41, 16 July 2008 (UTC)[reply]

Thanks for the advice but Tripods are banned without written permission! Also not sure how I can change the exposure time - I have a point and click camera. Anyway, I put a higher resolution version up, but it may be just as noisy as the original! best, Sunil060902 (talk) 00:21, 17 July 2008 (UTC)[reply]

Yup, that one's even noisier than the first. Point and shoots can take great pictures, but may be challenged in situations like these. Modern P&S are cramming a ton of megapixels onto a small sensor which results in a noisier image. (Don't let anyone sell you a camera on megapixels alone). Also in low lighting with no flash, the camera needs to use a higher sensitivity (ISO) to get a good exposure, which results in even more noise. Fletcher (talk) 01:13, 17 July 2008 (UTC)[reply]

Somewhat off topic: Sorry, but that website has one of the most misleading illustrations I've ever seen. The second image isn't noisier that the first; to be comparable, they must be presented in equal sizes, i.e. the second must be downsampled to the resolution of the first. Ceteris paribus, cramming more pixels onto a sensor never increases noise (or more accurately S/N ratio), and in practice, the reverse actually happens. Manufacturers put more pixels into higher-end cameras, and generally, along with those pixels, newer technology, better processors, and higher quality optics. If you try to buy a 6 megapixel camera today, you're going to get a piece of crap because no manufacturer is willing to inverst higher-end components in a low megapixel camera. And since higher megapixel cameras generally have better components, they generally have lower noise when compared downsampled equal to lower MP cameras. Oh, and besides noise, higher res often brings more detail, so it's better to shoot higher MP and downsample later if you want a lower MP image. Thegreenj 02:46, 17 July 2008 (UTC)[reply]

The argument is not that adding pixels is no improvement, nor that new generations of cameras aren't more refined, but that adding pixels to a lower-end P&S without also improving its lens and sensor may harm image quality. This has been widely reported.[1] Indeed, to me the higher resolution image he provided looks worse. Still, I agree it's better to shoot initially in high res if you can. And just for the record, I have a year-old 6 MP camera, and while less than professional quality, no, I don't think it's a piece of crap. Any crappy images produced most likely have something to do with the operator. :-) Fletcher (talk) 04:13, 17 July 2008 (UTC)[reply]

Gotta agree with Fletcher. Somewhere along the line 'megapixels' became the selling point in the P&S sales wars, and manufacturers regularly just cram more onto the sensor without improving the optics, etc (in fact, often by scrimping on the other stuff that makes a good camera). At the high end of the market they may be doing the right thing, I have no argument with the Mark III Canon 1Ds at 21MP for example, but this is not the case at the lower end of the market, where the crammed little sensors are putting out awfully noisy pictures. Case in point - my almost 4yo Canon A95 at only 5MP still takes better pictures than any newer 'improved' high MP point-and-shoot that I've used. I'm not saying it's better than anything currently available, but I get consistently better results than on various more recent and supposedly improved equipment that I've used. Another issue of course is that the higher end P&Ss are now getting pushed out of the market by the decent psuedo-SLRs, the continually dropping SLR prices. They have to keep those P&Ss cheap, and often do so by cutting corners, but generally don't want to let the MP count drop because that's what so many people shop on. --jjron (talk) 09:26, 17 July 2008 (UTC)[reply]

Somewhat on topic: Is there another one of those benches that you could use to steady your camera? If there is, set the camera on ISO 100 or Lo-ISO (if possible), set it on timer, and see how it turns out. Re:Fletcher, rule of thumb shutter speed is 1/(35mm equ. focal length), which is about 1/50 here. Thegreenj 03:05, 17 July 2008 (UTC)[reply]

<==Good point; forgot about that. Fletcher (talk) 04:13, 17 July 2008 (UTC)[reply]

Hey, calling the D40 out is cheating :). Really, though, I get what you and jjron are saying, but I think it's blown out of proportion by both sides, manufacturers and Ralph Naders of photography. Basically, if you have 4x the pixels on a sensor, all other things equal, each pixel will have to be amplified 4x and will have 4x the noise, but when downsampled 4x (to a comparable resolution to the original) it will have roughly the same amount of noise. Of course, this doesn't work perfectly (it doesn't take into account the size of the "grain", for example), and jjron makes a point about cutting corners. Thegreenj 03:25, 19 July 2008 (UTC)[reply]
- Not sure I follow the logic of this technique. When you downsample the image you may lose the noise but don't you also lose detail? I don't see the point of buying all those shiny new megapixels if you can't use them. Fletcher (talk) 17:02, 19 July 2008 (UTC)[reply]
  - No, don't downsample, but don't say that fewer megapixels is better, because in the end, it's more complicated than that. A higher MP image with more noise at 100% than a lower MP image doesn't necessarily make it worse. Thegreenj 00:40, 22 July 2008 (UTC)[reply]

Seconder