Date: Fri, 05 Apr 2002 08:57:15 From: "neil barnes" <[EMAIL PROTECTED]> Subject: [OT] RE: Digital cameras (danger Matt/Ray length exposition :)
> >Date: Fri, 05 Apr 2002 09:11:07 +0800 >From: Raymond <[EMAIL PROTECTED]> >Subject: RE: Digital cameras (was [LIB] New Website >OK lemme get this right in my head because now I'm starting to get confused >... heh. A slow lens means you need to use either a slower shutter speed or >a larger aperture. But doesn't a larger aperture mean LESS focusing depth? Yes. ObInfoDump: 'As you know, Bob,' A slow lens is one with a small maximum aperture - i.e. a larger f-number. As the aperture is increased (the f-number reduced) then the depth of field *at that aperture* is decreased and the exposure time *for the same light* is reduced. In practice, a difference of f1.8 to f1.4 is pretty pointless unless there are other overriding considerations (edge and centre resolution, weight etc) but a difference of f4 to f1.8 is well worth having. But - f-number is proportional to the length of the lens, so my 4*5 camera's 135mm 'standard' lens (it's actually slightly wide angle compared to a normal SLR 50mm/48 degree lens) has a maximum aperture of f4.5, while the OM 1-n standard lens is f1.8 - for about the same lens diameter. Theoretically, with both set to the same f-stop, they should have about the same resolving power in lines per inch, and the same depth of field, but the 4*5 camera has an awful lot more inches of negative... Having said that, depth of field is (obviously) proportional to the length of the lens, and the length of the lens is a product of the field of view and the size of the image sensor - whether it's a 10*8 inch film plate or a 2*3 mm LCD sensor. Cheap electronic cameras have low resolution, which only needs a small (i.e. cheap) sensor, which needs a short focus lens, which means that the depth of field isn't good unless it's stopped well down, which means the lens is slow. I would expect any photographer to be intimately familiar with the differences in viewing angle and depth compression between zooming in and moving closer. An professional will be able to consider every part of the image in respect to the brightness of the final output, and will be aware of the effects he can get by pushing/pulling development, different development processes, different grades of printing paper, different finishes. Most of these effects are *not* equivalently available in electronic media, and will not become available until sensors producing direct hi-resolution (not interleaved) images with a usable pixel depth of at least sixteen bits per colour are available - and also, displays which can show these images or printers which can print them. Bit depths are not well understood: the common 24M colour display cards use eight bits per colour - 256 different levels of illumination in each colour. This number was picked because using a colour monitor (twenty years ago) in normal viewing, this was sufficient so that people in general could not tell the difference between adjacent colours differing by one level. Which is fine, although monitors are now better and there may be a case for nine or ten bit displays, eight is so ingrained in computer hardware we're unlikely to see a change. But... if you want to do any change on the image - e.g. gain or brightness or gamma or colour changes, you are performing mathematical processes on the eight bit data which has two disadvantages: First, that the result may exceed the maximum or minimum possible values, causing parts of the image to clip to either black or white (and remember with a colour image, only one of the colours may be affected, cusing a colour shift in high or lowlights). Secondly, because you are multiplying by a constant which is very rarely a multiple of two, the results fall in the cracks between the numbers - causing quantisation errors (posterisation). If you are going to provide further changes to the image, then these errors are cumulative - you can end up, if you're not careful with an image, with significantly less discrete values than should be there, with odd colour shifts, all in the effort of improving the picture. If you start to apply geometrical changes as well - rotations, scaling etc then you have further problems of aliasing and mispositioning of the image. The only way to avoid these visible effects is to have a lot more information in the original image - even if you can't see it - than the final image will display. The bit equivalents for B+W images using modern emulsions are on the order of 15-17 bits for the negative (more bits as the size of the neg increases) and around 10-11 bits for a glossy print. Resolution on 35mm negs is around 3000lpi. Colour prints are a little lower both in bit depth and resolution, but a colour slide (diapositive) can reach 14-15 bits colour depth. I bet that was more than you wanted to know! Of course, the vast majority of people only want a camera to point and click and produce something recognisable - this discussion/rant is not aimed at them. Cheers, Neil _________________________________________________________________ Chat with friends online, try MSN Messenger: http://messenger.msn.com ************************************************************** http://libretto.basiclink.com - Libretto mailing list http://www.silverace.com/libretto/ - Archives -------TO UNSUBSCRIBE------- Reply to any of the list messages. The reply mail should be addressed to: [EMAIL PROTECTED] - Then replace any text on the message's subject line: cmd:unsubscribe --------TO UNSUBSCRIBE DIGEST------ Do above but with this on subject line: cmd:unsubscribe digest **************************************************************
