DigitalBrain

新進會員

2004/01/27 12:25 請先登入後才可回應

器材: 其他其他

Sorry for last post. I should specific more detail on the title.
Ok. Don"t talk about convenience and post pocessing.

If side by side, As I know, DSLR has better CCD which may be fair to compare to film. If using the same lens, same ISO, same aperture and shutter speed. Does DSLR can be better than film SLR in the field of image quality.

I don"t have money to make the judgement. If digital is better than film, perhaps I will give up my EOS 30.
Any sources to support your point?
Thanks!

*香港至NET超人* 網路會員	21) 2004/01/28 09:18 I have no problem on S2Pro to 8R. You will never get what you feel on film when printing by digital. Remember that film"s feeling is primary but not absolute perfect. Film has integral sign curve on softening highlight and shadow. It is actually "distorted".
jsklee 網路會員	22) 2004/01/29 12:27 Could I say that it is this "distorted" treatment of highlights and shadow makes film better than digital?
*香港至NET超人* 網路會員	23) 2004/01/29 12:29 If that is not truely recorded, I would say that this effect is suitable to one"s need.
jsklee 網路會員	24) 2004/01/29 12:57 I think that our eyes have the largest Dynamic Range! So we are keen to pursuing the recording media, no matter it is film or CCD, to be more or less similar to our eyes!
vincenthoy28 網路會員	25) 2004/01/29 13:03 但係DSLR D紫邊問題又點呢?? 既然超人兄覺得現在DSLR高低光位OK, 都想請教下係唔係有方法對付紫邊問題? 還有係高光位得白色一塊低光位又黑色一塊O既問題? 因為我只係試過直出, 係唔係用photoshop就可以解決個問題? 因為我都想學下野.
jsklee 網路會員	26) 2004/01/29 13:16 You can use PS to extract the details in bright and shadow area, except where it does not contain any data, which means that it is totally bright and totally dark.
jsklee 網路會員	27) 2004/01/29 13:20 http://www.photographical.net/canon_1ds_35mm.html
jsklee 網路會員	28) 2004/01/29 13:39 http://www.normankoren.com/Tutorials/MTF7.html
*香港至NET超人* 網路會員	29) 2004/01/29 13:54 唔係所有DC都有紫邊, 呢個問題亦因為DC的MOSAIC形式感色光及卓算問題而成. 至於高光位係點, 影RAW會好好多, 唔係冇DATA, 係點搞佢出來.
vincenthoy28 網路會員	30) 2004/01/29 17:49 Just futher to the web site: http://www.normankoren.com/Tutorials/MTF7.html It seem to me that the resolving power for DSLR still far away from film. i.e Max. resolving power of EOS1Ds is 66lp/mm but the Provis can go up to 150Ip/mm according to the information provided by Carl Zeiss.
mickey_mouse 網路會員	31) 2004/01/29 18:01 With the reference, >Just futher to the web site: >http://www.normankoren.com/Tutorials/MTF7.html IDS get higher resolution than film. When did you get the 66lp/mm figure? for the 10D. we alread need to have a lenses with resolution roughly = 3000/24 = 125 lpmm lenses. Not many lenses have such resolution power. CCD is no longer bottleneck for a 135 camera. For a 1Ds. we need roughly a lenses with 6000/36 = 167 lpmm
vincenthoy28 網路會員	32) 2004/01/29 19:51 I don"t know there you get the result of 6000/36 = 167 lpmm. But from the Digital camera resolution summary it write that the MTF is 66Ip/mm
*香港至NET超人* 網路會員	33) 2004/01/29 20:53 http://www.normankoren.com/Tutorials/MTF1A.html#Excellent_lens Of what the 66lpmm was reported by just this "Excellent lens", this test really produce by top end of Flintstone laboratory! With the original link: http://www.normankoren.com/Tutorials/MTF7.html Almost all the DSLR is out performance Provia 100 with 4000dpi in the MTF test....
vincenthoy28 網路會員	34) 2004/01/30 04:33 But I stil wonder as the test were based on the Canon 28-70mm f/2.8L lens with around 61 lines/mm. Actually, I understand that there some excellent lens like Contax 85f1.2 are able to provide the resolution upto 200lines/mm. Although the test just mention DSLR is out performance Provia 100 with 4000dpi in the MTF test, I think is not fair as the lens & scanner already limit the "real" max resolution of the film. Also, I think mickey_mouse may be right for a 1Ds has 167 lpmm if there is no any noise generated by the CMOS itself and no anti-aliasing (low pass) filters. But this is not happened in the real world! Upto this moment, I still discovered most of the comparsion that people make are based on "unfair" condtion like they only use the lens with "low resolution" (I think at least they should use 200f1.8L). Also, I just wonder if the film scanner is a good enough method to explain the "real" resolution of the film.
mickey_mouse 網路會員	35) 2004/01/30 05:23 For an "ideal" lenses of 200 lpmm, no matter how good is the film, it can at most record less than 20 million pixels for 24x36mm film. And most of the films are only around 150lpmm maximum. it can only roughly record less than 15 million pixels.
FTL 網路會員	36) 2004/01/30 07:08 The matter is not how much XXX Lpmm or how many pixel, but rather the total raw information recorded. Of course Professional grade digital can do just as well for all the pre-press and printing need. So do film. If there is a need to quantify the comparison. We must thus set some guideline and prerequisite. Comparing film with resolution after digitizing it is certainly un-realistic as then it become a digitized image of the film, not the film itself. Let"s take a typical 6M-pixel ( normal hobbyist DSLR ) image as sample. With todays raw development, you can usually get 12 bit per channel bit depth, but JPG will have to thin down to the normal 24 bit ( 8 bit per channel ) and printing it on say dedicated printer ( nominal is 300DPI ) gives roughly a 8X10 bleededc on short side. But this is printed at 1:1. To get that same 1:1 Print with film, you need to take a picture with a large format 8X10 camera, and do contact print. And of course in this case the film will win right off as it will have vastly more recorded data. And any Analog recording will have far higher graduation than digital while the latitude is entirely different matter. See the point, Since they are essentially 2 different media that take 2 different way to achive a final product ( be it print or what ever ). Its always unfair to compare one with others in a single criteria. Ultimately the practical comparison would either be 1. by usage, how you want to use the system 2. by final imaage, is it a print, a web display, or a electronic pre-press processed file 3. by cost 4. by need .... there are just picture that demand certain media or capture method As far as normal consumer/hobbyist goes. Film with traditional D&P held sway over digital not because film is better, its just at the same price bracket, digital is not yet evolved long enough to have the quality filtered down. And besides, resolution / accutance is of almost no consquence in typical day to day shooting as today"s photgraphers usually use Zoom lens handheld. How much resolution can you get over that. Even the best of lens/camera cannot correct for that. So the comparison is mute. And 至NET超人 is correct, Film had certain tone and sensitivity curve which is unique which also make the style and taste of the film. If that"s what suit the photographer, then that"s the tool. Digital had that too but they can be post-exposure processed away.
*香港至NET超人* 網路會員	37) 2004/01/30 09:37 I know the tone are very much different, so I concentrated on resolution. Digital film is not fair may just becasue the scanner is poor and resolution is low. But in the real world, 10,000dpi drum scan was there. With such high resolution, it records even the shape of film particles. Do you still think that film is so fine and high resolution? If you are so strong on this point, try to find someone who can access lab with 100x, or large magnification, oil immerse microscrope to see details of your film, as well as lenses performance. You will cry immediately. FTL told something important, is how often you get the most of the film. I think this is NEVER! There is almost 99% lenses on the earth never reach 100lpmm or more. With finer the film, you get more resolution because the final resolution is: SQRT{1/[SQR(1/lens_res) + SQR(1/film_res)]} With the lens resolution = 60lpmm and original film resolution = 100lpmm, the final resolution is 51.45lpmm. If the new film got 150lpmm, the final resolution is just increased to 55.71lpmm. The news film got 50% more resolution but the final resolution is just 5% more.
AX 網路會員	38) 2004/01/30 13:02 dpi in current designation for resolution power is missleading. 1. they tell only B&W resolution, nothing related to daily photo targets. 2. dpi doesn"t tell the fuzziness or quality of the pixel. that is how well it represents the incident data in terms of color and spacial information. 3. correct rendition of color information is importance than ", resolution / accutance is of almost no consquence in typical day to day shooting as today"s photgraphers usually use Zoom lens handheld" since this is related to our BRAIN"s previous experience about the texture of targets like woods texture in a scence shot. == So a plastic look is a well description of current crop of DC or DSLR (MOSAIC形式感色光)with resolution <16Mp. when compared to 35mm slide film(quality 8Mp)
Francis Poon 網路會員	39) 2004/01/30 15:14 According to http://www.normankoren.com/Tutorials/MTF7.html the DSLR that exceed scanned 4000 dpi 35mm film are: Provia F 1.0 (as the reference) 1Ds 1.01 14n 1.15 D100 & 10D 0.72 SD9/10 0.73 Remember 4000 dpi is still a limitation of extracting all details from film. In my opinion, if we prefer scanned film film image at 4000 dpi, then you may get better resolution by 1Ds and 14n. The 6Mp DSLR are close to about 72%. What we prefer DSLR is their convenience. As I view through microscope or high magnification loupe, I found the the 4000dpi scanned image still not capture all the detail of Provia 100F, I guess 6000 dpi is more proper. Up to now, I"m not fond of scanning film, it"s boring and time consuming. For dynamic range, we are just comparing in the same way the tube amp to transistor amp, tube amp characteristic is somewhat like neg. film and DSLR is more like CCD or CMOS, which will crop the high light and pollute the dim light due to noise. People prefer the tube amp because it is more comfortable for hearing even though the non-linear distortion is greater when testing with instrument. The bayer CCD oe CMOS pixel pattern also require a better resolution lens in order to get similar image resolution as film. That"s why the SD9/10 with 3Mp pixel perform as well as the 6Mp 10D/D100 using bayer pattern imager.
vincenthoy28 網路會員	40) 2004/01/30 18:51 香港至NET超人, 所以我用定焦鏡. 如果用數碼, 我影o既相o既影像質素就會比CCD/CMOS限制住. 其實我唔抗拒用DSLR, 因為佢有preview等我可以好準確控制我張相要表達O既野. 可惜我有時會放大D相, 所以覺得現在重未係時候. 我想如果有10Mp o既Foveon"s X3 sensors, 咁DSLR真係come of age呀!