 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
| PhotoKB HomeDiscussion Groups Digital Photography Digital PhotoDSLR CamerasZLR CamerasPoint & Shoot CamerasFilm Photography 35 mmLarge FormatMedium formatDarkroomFilm and LabsOther EquipmentPhoto Technique Nature PhotographyPeople PhotographyTechnique GeneralGeneral Photo Topics General TopicsAustralian PhotographyUK PhotographyDirectoryPhoto Clubs |
Photo Forum / Digital Photography / DSLR Cameras / August 2006New DSLRs still with 12-bit conversion only
There are people scanning 4x5 film at 48 bit depth in order not to ruin the images when they are manipulated digitally. Isn't it time for consumer DSLRs to at least go to 14 bits? Most professional CCD cameras provide 14 bit conversion or higher now. > There are people scanning 4x5 film at 48 bit depth in order not to ruin > the images when they are manipulated digitally. Isn't it time for > consumer > DSLRs to at least go to 14 bits? Most professional CCD cameras provide > > 14 bit conversion or higher now. No real need to go past 14 but 14 would be very nice, even 13 would help a lot. Scott The fact that you cannot see those 14 bits of color information on a monitor or in any printing process is beside the point. That being the case you should hold out for 16 bit color. Those idiots are ruining their 4x5 masterpieces with only 14 bit color. > The fact that you cannot see those 14 bits of color information on a monitor > or in any printing process is beside the point. That being the case you > should hold out for 16 bit color. Those idiots are ruining their 4x5 > masterpieces with only 14 bit color. The dynamic range of a DSLR is limited by the 12 bit converter, 14 would give it a greater range. This would allow to pull out more detail in the shadow areas, or keep from blowing out highlight as soon. Scott > The fact that you cannot see those 14 bits of color information on a monitor > or in any printing process is beside the point. That being the case you > should hold out for 16 bit color. Those idiots are ruining their 4x5 > masterpieces with only 14 bit color. Well, "those idiots," if we're to trust the original poster's sources, aren't ruining their masterpeices with only 14 bit color... they're using 16 bit per channel color (48 bit simply being another way of talking about 16 bit per channel color). 42 bit would be the same as 14 bit per channel. I'm all for 16 bits per channel out of our cameras... the software supports it, so it seems a waste that the cameras don't. PhaseOne's latest backs are 16 bit, so it certainly isn't beyond our technology. I wonder if Canon's next big announcement this month might make that jump, at least to 14. After all, all early reports suggest that the Digic III is quite a capable image processor... Regardless of who moves the DSLR segment to it, we really need to jump to 16 bit per channel color... more than megapixels, at this point bit depth is becoming very important (as part of dynamic range, which is the next great frontier for advancing digital images if you ask me). Will > There are people scanning 4x5 film at 48 bit depth in order not to ruin > the images when they are manipulated digitally. Isn't it time for > consumer > DSLRs to at least go to 14 bits? Most professional CCD cameras provide > > 14 bit conversion or higher now. Doesn't the Leica digital back give 16 bit? If you have a spare $10K. Tom >> There are people scanning 4x5 film at 48 bit depth in order not to ruin >> the images when they are manipulated digitally. Isn't it time for [quoted text clipped - 4 lines] > >Doesn't the Leica digital back give 16 bit? If you have a spare $10K. Yes, it has always been 16bits. > There are people scanning 4x5 film at 48 bit depth in order not to ruin > the images when they are manipulated digitally. That's nuts. There's no way that most of those bits can be anything but noise. >>There are people scanning 4x5 film at 48 bit depth in order not to ruin >>the images when they are manipulated digitally. > > That's nuts. There's no way that most of those bits can be anything but > noise. Technically, if you state that "most" of those bits are noise, that would mean 9 out of 16 are noise. I scan 4x5 film with a 12-bit/channel scanner and find I can do a better job of extracting details, especially in highlights than I can with 8-bit scans. I need to work less with graduated neutral density filters too, and that is with velvia! While I don't believe 16-bits are needed for film scanning, 12-bits are better than 8 in my experience. Roger photos at: http://www.clarkvision.com >>>There are people scanning 4x5 film at 48 bit depth in order not to ruin >>>the images when they are manipulated digitally. >> That's nuts. There's no way that most of those bits can be anything but >> noise. > Technically, if you state that "most" of those bits are noise, > that would mean 9 out of 16 are noise. Umm, if you mean 16 bits each for three colors, you have a point. But that's not what the post said. > >>>There are people scanning 4x5 film at 48 bit depth in order not to ruin > >>>the images when they are manipulated digitally. [quoted text clipped - 7 lines] > Umm, if you mean 16 bits each for three colors, you have a point. But > that's not what the post said. When someone talks about 48 bits of depth you can assume they are talking total bits for all three colors, or 16 bits per color. Scott > There are people scanning 4x5 film at 48 bit depth in > order not to ruin the images when they are manipulated > digitally. I do not believe that many scanners would have 16-bit converters. Could you name some make/model? Many/most/all scanners can output the final image data in 48-bit RGB mode, and so can many/most/all the dSLRs too. >Isn't it time for consumer DSLRs to at least go to 14 bits? No, since they only have something about 9 stops of dynamic range. So 10 bit conversion is enough, 12 bits is an overkill already. Note that 12-bit conversion results image data that is said to be 3*12= 36 bit deep. >Most professional CCD cameras provide 14 bit conversion or higher now. What are these professional CCD cameras, could you please name some Make/Model? Timo Autiokari http://www.aim-dtp.net > > There are people scanning 4x5 film at 48 bit depth in > > order not to ruin the images when they are manipulated [quoted text clipped - 10 lines] > No, since they only have something about 9 stops of dynamic range. So 10 > bit conversion is enough, 12 bits is an overkill already. Do this simple task, take a shot that has a pretty deep shadow at iso 100, now take the same same shot at iso 400, the iso 400 shot will have more detail in the shadows. Note that the CCD has the same exposure in both cases. A good 14 bit converter would pick up this detail with out the need for switching to ISO 400. In fact since changing the ISO of camera is just a gain change a camera that has enough bits in the converter would have no need for anything but the lowest ISO setting. BWT if you are only getting 9 stops out of your DSLR there is a real problem there, are you shooting raw or JPEG, note that not shooting raw will rob you of at least two stops of range if not a bit more. Scott > Do this simple task, take a shot that has a pretty deep shadow at iso > 100, now take the same same shot at iso 400, the iso 400 shot will have > more detail in the shadows. Note that the CCD has the same exposure in > both cases. A good 14 bit converter would pick up this detail with out > the need for switching to ISO 400. the sony does it already. > In fact since changing the ISO of camera is just a gain change a camera > that has enough [quoted text clipped - 6 lines] > > Scott > > > There are people scanning 4x5 film at 48 bit depth in > > > order not to ruin the images when they are manipulated [quoted text clipped - 16 lines] > both cases. A good 14 bit converter would pick up this detail with out > the need for switching to ISO 400. Would it mean less noise if done this way? > > There are people scanning 4x5 film at 48 bit depth in > > order not to ruin the images when they are manipulated > > digitally. > > I do not believe that many scanners would have 16-bit converters. Could > you name some make/model? Pretty much all higher end scanners are 16-bit, e.g.: Nikon 5000 ED http://www.nikonusa.com/template.php?cat=1&grp=98&productNr=9238 > Many/most/all scanners can output the final image data in 48-bit RGB > mode, and so can many/most/all the dSLRs too. [quoted text clipped - 3 lines] > No, since they only have something about 9 stops of dynamic range. So 10 > bit conversion is enough, 12 bits is an overkill already. Haven't you gotten your test charts to work yet? If you did get a quality setup working, you would see this is a false statement. > Note that 12-bit conversion results image data that is said to be 3*12= > 36 bit deep. You are stating color bit depth but what is important in dynamic range is the bit depth per color. > >Most professional CCD cameras provide 14 bit conversion or higher now. > > What are these professional CCD cameras, could you please name some > Make/Model? Most of the models on this page are 16-bit: http://www.britastro.org/vss/ccdtable.html Roger > Haven't you gotten your test charts to work yet? If you did get > a quality setup working, you would see this is a false statement. I have made test shots with both D60 and MK2, both have just slightly over 9 stops range. I have not had the time to finish a page about them, there is plenty to do with result handing, most of the DR evaluations that I have seen manage the results incorrectly. >> Note that 12-bit conversion results image data that is said to be >> 3*12= 36 bit deep. >> > You are stating color bit depth but what is important in dynamic > range is the bit depth per color. That is very so. I merely tried to explain to RichA that 48 bits is the same as 16 bits/channel. And that those 16-bits are just the A/D converter bits (or the bit-depth of the file format) not the DR of the image data. >> What are these professional CCD cameras, could you please name some >> Make/Model? > > Most of the models on this page are 16-bit: > http://www.britastro.org/vss/ccdtable.html RickA mentioned both "DSLRs" and "professional CCD cameras", from that I understood that the question was about common color digital SLR cameras. Are any of the cameras mentioned on the abve web page color digital SLRs? I'm very well aware that there are scientific digital cameras with 16-bit converters (and with the required cooling). Timo Autiokari > > Haven't you gotten your test charts to work yet? If you did get > > a quality setup working, you would see this is a false statement. [quoted text clipped - 3 lines] > there is plenty to do with result handing, most of the DR evaluations > that I have seen manage the results incorrectly. I guess you are defining dynamic range differently than others, as many other tests come in at around 11-stops. After all, if DSLR cameras test in at 50,000 to 80,000 full well electrons and read noise of less than 4 electrons, 80,000/4 = 20,000 dynamic range = 14.3 stops. With 12-bit converters, cameras become bit limited to just under 12 bits, and that is what people are observing in good cameras. Do the test Scott outlined in this thread. > > Most of the models on this page are 16-bit: > > http://www.britastro.org/vss/ccdtable.html [quoted text clipped - 5 lines] > SLRs? I'm very well aware that there are scientific digital cameras with > 16-bit converters (and with the required cooling). The fact that scientific cameras are 16-bit and cooled are unrelated. Cooling is to reduce thermal noise in long exposures. 16-bit A-D converters are used to adequately digitize the signal over the entire dynamic range of the sensor. Digital cameras use a crutch to avoid using 16-bit (or 14-bit) converters: an analog gain step before the A-D known to the user as the ISO setting. If you have an 80,000 electron full well and read noise below 4 electrons, you need 14+ -bit converters to digitize the signal at low levels. 12-bit converters do not do an adequate job, and that is simply illustrated by the test Scott outlined: 1) meter a scene with dark shadows and take a picture at the lowest iso of the camera. 2) increase the iso (e.g. to 800 or 1600) and take an image of the same scene with the same f-stop and exposure time as in #1. Do you see more detail in the shadows? For good DSLRs, the answer is yes (every DSLR I have seen tested shows this to be the case). That means the 12-bit A=D is limiting the information recorded. Sensor manufacturers and professional electronics engineers use a published test method to measure sensor dynamic range. That method and results for a DSLR are explained here: Procedures for Evaluating Digital Camera Sensor Noise, Dynamic Range, and Full Well Capacities; Canon 1D Mark II Analysis http://www.clarkvision.com/imagedetail/evaluation-1d2 (where you will see in Table 1, column G that the 1D Mark II has a dynamic range at ISO 100 = 3190 or 11.6 stops, and that the A-D limits read noise to 16.6 electrons, where the read noise is only 3.9 electrons when adequately digitized at ISO 800). Roger > > There are people scanning 4x5 film at 48 bit depth in > > order not to ruin the images when they are manipulated [quoted text clipped - 21 lines] > Timo Autiokari > http://www.aim-dtp.net Finger Lakes, SBIG, Mamamatsu, Apogee. I think they all offer 14 bit or better. >> > There are people scanning 4x5 film at 48 bit depth in >> > order not to ruin the images when they are manipulated [quoted text clipped - 24 lines] >Finger Lakes, SBIG, Mamamatsu, Apogee. I think they all offer 14 bit >or better. It's a pity that none of those companies offers a decent DSLR. Consumer, professional, or otherwise. TR > >> > There are people scanning 4x5 film at 48 bit depth in > >> > order not to ruin the images when they are manipulated [quoted text clipped - 29 lines] > > TR If they did, it would cost about 10x what a Canon 1DsMkII does. >If they did, it would cost about 10x what a Canon 1DsMkII does. Leica has 16bits and is cheaper than that. > >If they did, it would cost about 10x what a Canon 1DsMkII does. > > Leica has 16bits and is cheaper than that. The camera companies cited aren't consumer companies and make cameras for industry, science. They tend to cost a lot more than a consumer camera. Many of them are cooled for longer exposures as well. They also need to be hooked up to a computer/laptop to function, download images, etc. The pricing on them, like consumer cameras has come down. Five years ago, you could have paid $40,000 for a top flight 4-8 meg camera. They are generally less costly now. But, if you want one of them with a top-grade CCD (as opposed to what is used in even the best consumer cameras) you still pay a great deal. >> >If they did, it would cost about 10x what a Canon 1DsMkII does. >> [quoted text clipped - 10 lines] >But, if you want one of them with a top-grade CCD (as opposed to what >is used in even the best consumer cameras) you still pay a great deal. Really Rich, I'm a consumer and have an M8 on order. Sorry you can't afford to shoot with good equipment. >> >> > There are people scanning 4x5 film at 48 bit depth in >> >> > order not to ruin the images when they are manipulated [quoted text clipped - 31 lines] > >If they did, it would cost about 10x what a Canon 1DsMkII does. Timo asked you to identify a "professional CCD camera" that provides 14 bit conversion, and you gave him 4 that do not offer anything close to a DSLR. (This froop is rec.photo.digital.slr-systems, y'know.) While I will concede the point that these companies manufacture cameras for the professional and (very) serious amatuer markets, most photos, professional or otherwise, would not consider any of these cameras to be a suitable replacement for a cheap P&S, much less a DSLR. Too heavy, too slow, and even the smallest lens is much too big to be a walk-around lens. Rich, I doubt that even YOU consider a cameras specifically designed for astrophotography to be even remotely similar to a DSLR. TR > >> >> > There are people scanning 4x5 film at 48 bit depth in > >> >> > order not to ruin the images when they are manipulated [quoted text clipped - 45 lines] > Rich, I doubt that even YOU consider a cameras specifically designed > for astrophotography to be even remotely similar to a DSLR. Try to broaden you though process just a bit. My POINT was that there is a REASON they don't use 10 and 12 bit conversion and PERHAPS it would benefit DSLRs. I am not advocating using an astronomical camera attached to a frigging laptop as a DSLR replacement. Though it might be interesting to see what it could do! > TR >>> > There are people scanning 4x5 film at 48 bit depth in >>> > order not to ruin the images when they are manipulated [quoted text clipped - 27 lines] >It's a pity that none of those companies offers a decent DSLR. >Consumer, professional, or otherwise. Leica offers you 16bits in a great DSLR and soon in a great rangefinder. >>>> > There are people scanning 4x5 film at 48 bit depth in >>>> > order not to ruin the images when they are manipulated [quoted text clipped - 30 lines] >Leica offers you 16bits in a great DSLR and soon in a great >rangefinder. I know. I was challenging Rich's asinine response to Timo. TR Well I guess the Cheaper DSLR only have 12bit but the Leica DRM module for the R8/R9 has always had 16bits as will the new M8. http://www.leica-camera.us/photography/r_system/digital-modul-r/ Read the specs. |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2009 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.