 |
 | 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 2005Is RAW usage/support just overblown, elitist hype?
The reason I ask this is that someone on another forum said that it's impossible to compare two cameras to each other (even if they are radically different cameras) by using only JPEG, that you have to use RAW images. Is this the case? Because the differences are clear in most cases from what I've seen comparing cameras with JPEG output, and I've never seen RAW images used in a test between two cameras. I read a couple magazine pieces that seemed to suggest the people who say only RAW is suitable for high quality output (to print or a monitor) are basically crackpots, though the mags didn't use that terminology, and that the format is used by some to lord over the heads of the uninitiated and that statements like "JPEG is only suitable for small vacation pictures" (quoted) are nonsense. The only thing I've heard said about RAW when dealing with the specifics of images is that it can allow the recovery of shadow or highlight information because the format better preserves dynamic range of an image. So, when people make the far-fetched claims that JPEG (no matter what compression level) is no where near as good, on what do they base it? Under what circumstances does RAW really come to the fore? -Rich > The reason I ask this is that someone on another > forum said that it's impossible to compare two [quoted text clipped - 22 lines] > Under what circumstances does RAW really come to the fore? > -Rich The above statement is quite wrong and comes from an uninformed person. A RAW file is not an image file. You can't print a RAW file or 'see' it on a monitor until the data is converted to an image. Once you have converted it, it no longer is camera RAW data. There is nothing incredibly wrong with JPEG format except that it uses a throw away compression process and camera makers have decided to use it to compress their images. Herein lies the pitfalls of JPEGs. Those camera makers who choose to offer TIFF or JPEG files, recognize that some people need to/ want to enlarge their pictures and the high compression ratios used with JPEGs, do not provide very good enlargements. Otherwise, the JPG format absolutely throws away highlight and shadow data during compression in the presumption that all you really "need" is the mid range data. Camera RAW not only keeps vital, unprocessed information for later retrieval but it does not "throw away" data. I believe it does get compressed but not to the level of a JPG camera image. I frequently save a RAW file after processing in Photoshop as an *UNCOMPRESSED* JPG. These uncompressed files are as good as TIFF files when enlarging or printing images. It is only when you compress a JPG it starts to lose quality, get artifacts and other image flaws.  SignatureDouglas, Saying of the day: Never trust a skinny cook. > I frequently save a RAW file after processing in Photoshop as an > *UNCOMPRESSED* JPG. These uncompressed files are as good as TIFF files > when enlarging or printing images. It is only when you compress a JPG it > starts to lose quality, get artifacts and other image flaws. I am under the impression that if you save an image as a jpeg, make a modification, then save a again as a jpeg, and on and on... That each time you save as a jpeg the process of saving as a jpeg (algorithm) loses info (it's not lossless) so the picture degrades some. This is why some use tiff or psd for many edits and many saves. I have not heard of a truly lossless jpeg. BTW pse3 has jpeg2000 format (and you can select option of lossless jpeg200) so you can get a lossless form of jpeg 2000 file. One advantage is that it's (the jpeg200) a smaller file than a tiff file (but it wont preserve layers). > > I frequently save a RAW file after processing in Photoshop as an > > *UNCOMPRESSED* JPG. These uncompressed files are as good as TIFF files [quoted text clipped - 7 lines] > or psd for many edits and many saves. I have not heard of a truly lossless > jpeg. Well, every time I "save as" an unedited jpg in Photoshop with the best quality it gets bigger. Either way, it's not the same file as before. Greg >The above statement is quite wrong and comes from an uninformed person. >A RAW file is not an image file. You can't print a RAW file or 'see' it >on a monitor until the data is converted to an image. Well, actually, you can. It is just not traditional to do so, nor is a normal-looking image. A Sigma RAW bitmap is already an RGB image, but has not had color extrapolation calculations applied yet, so it is rather drab looking. A straightforward RAW bitmap is greyscale, and has a checker-pattern due to the fact that each pixel represents one of three primary colors in a Bayer camera, and have different levels even in a solid-colored and illuminated subject. You can also colorize the RAW bitmap so that the blue pixels are zero in the red and green channels, etc. The green is overwhelming, because there are twice as many green pixels, in an RGB Bayer camera. You can also interpolate each color channel to the full megapixel value in a Bayer camera, independently, as in the middle sample here: http://www.pbase.com/jps_photo/image/47061260 Green may predominate here, but not because of the number of green pixels, but because most Bayer DSLRs have green-filtered sensels that are more sensitive than the other two colors. All of these options are more RAW than any conversion with color and tonal tranfer curves applied, and demosaicing. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< > The reason I ask this is that someone on another > forum said that it's impossible to compare two [quoted text clipped - 22 lines] > Under what circumstances does RAW really come to the fore? > -Rich With a jpg (and I believe, tif) image, the camera firmware does a lot of image processing before saving the file to the flash card, like applying any preset parameters such as contrast, saturation, sharpening, color balance ( as the camera saw it, not necessarily correct), color space, and jpg compression to the image. The image characteristics are then permanently set, and only a relatively limited amount of manual processing can be done. Compression artifacts, blown whites, etc. can't be altered. RAW files have none of the above processing done, leaving the various parameters to be done in the RAW converter, and then in Photoshop or equivalent. For comparing images from two cameras, jpg's may well be treated differently between different cameras, and so the resulting images cannot be accurately compared. Of course, so may the RAW files, but there's a better chance of equal treatment there. Colin D. > For comparing images from two cameras, jpg's may well be treated > differently between different cameras, and so the resulting images > cannot be accurately compared. Of course, so may the RAW files, but > there's a better chance of equal treatment there. Agreed this is the reason to compare with raw. If everything is set up properly though and contrast/saturation set to equal settings, there is very little difference between jpeg & raw but cameras all have different defaults and that's more or less hidden. P&S cameras usually use some pretty strong saturation & contrast increase as well as sharpening etc. SignaturePaul Furman http://www.edgehill.net/1 san francisco native plants >The reason I ask this is that someone on another >forum said that it's impossible to compare two [quoted text clipped - 4 lines] >cameras with JPEG output, and I've never >seen RAW images used in a test between two cameras. Well, the test is very simple: try to produce for each camera the best image using RAW and using jpeg. If there is almost no difference, you can use jpeg to demonstrate the capabilities of the camera. If the there is a big difference, you have to use RAW. In general, there is no way of knowing whether the in camera RAW conversion uses the same algorithm as a stand-alone RAW converter. >I read a couple magazine pieces that seemed to suggest the >people who say only RAW is suitable for high quality output [quoted text clipped - 3 lines] >statements like "JPEG is only suitable for small vacation pictures" >(quoted) are nonsense. Only equipment freaks ask about the details of the process before they look at the picture. If the pictures come out all right, the process is good enough. >The only thing I've heard said about RAW when dealing with the >specifics of images is that it can allow the recovery of shadow >or highlight information because the format better preserves dynamic >range of an image. Jpeg compression loses all kinds of information. Many RAW formats are independent of the white balance setting, so you correct for that later. With jpeg, if you get the white balance wrong, you may not be able to correct it. >So, when people make the far-fetched claims that JPEG (no matter what >compression level) is no where near as good, on what do they base it? It depends on the amount of post processing you plan to do. Jpeg allows for only small white balance and contrast corrections. SignatureThat was it. Done. The faulty Monk was turned out into the desert where it could believe what it liked, including the idea that it had been hard done by. It was allowed to keep its horse, since horses were so cheap to make. -- Douglas Adams in Dirk Gently's Holistic Detective Agency > The only thing I've heard said about RAW when dealing with the > specifics of images is that it can allow the recovery of shadow [quoted text clipped - 5 lines] > Under what circumstances does RAW really come to the fore? > -Rich With my D70, this is the difference I see: http://www.geocities.com/angels2000photos/raw.jpg I posted that once before, and got comments back that the images had different contrast, one had been sharpened, or the colour difference accounts for the apparent increase in acutance of the converted raw images. It doesn't matter what I do, the result is always very similar to those examples. I certainly did not try to distort those results. Jpegs were all at Large, High Quality, Normal sharpness. Nikon Capture produces almost exactly the same result as RSE. Shooting a B&W resolution chart shows very little difference, as the test patterns just create moire at the level where you would otherwise probably see the difference. Overblown elitist hype? I don't know. If you were shooting a lot of images, then concerting RAW could use a serious amount of time. If you tend to (as I do) shoot less images, then a minute or so to deal with the raw file is no problem. > With my D70, this is the difference I see: > [quoted text clipped - 9 lines] > test patterns just create moire at the level where you would > otherwise probably see the difference. How did you get a fine JPEG and a RAW of the same frame? It seems to me that you are saying that the JPEG was created in the camera. -Mike >>With my D70, this is the difference I see: >> [quoted text clipped - 13 lines] > It seems to me that you are saying that the JPEG was created > in the camera. No, they were consecutive shots, one raw, one high quality jpeg. You can with a d70 get a raw plus a basic jpeg of the same frame, but then there is an enormous difference - the basic jpg is compressed to the extent that I don't think you'd even get a good 6x4 print from it. I took all precautions to eliminate camera shake, all were shot at high shutter speed using a tripod, and I have repeated this many times with the same result. The colour difference was an exercise for me in how difficult it is to match colour balance in raw using a jpeg image for reference. I cannot get it spot on. >>> With my D70, this is the difference I see: >>> [quoted text clipped - 24 lines] > match colour balance in raw using a jpeg image for reference. I > cannot get it spot on. That's what confused me. The middle shot looks identical. The palm fronds have not moved at all. -Mike >That's what confused me. The middle shot looks identical. The palm >fronds have not moved at all. Maybe nobody's at the bottom of the tree shaking it? SignatureOwamanga! http://www.pbase.com/owamanga >> With my D70, this is the difference I see: >> [quoted text clipped - 13 lines] >It seems to me that you are saying that the JPEG was created >in the camera. The 20D will produce a RAW and a large JPEG from the same shot in the camera. ********************************************************* "It looked like the sort of book described in library catalogues as "slightly foxed", although it would be more honest to admit that it looked as though it had been badgered, wolved and possibly beared as well." _Light Fantastic_ Terry Pratchett >>>With my D70, this is the difference I see: >>> [quoted text clipped - 16 lines] > The 20D will produce a RAW and a large JPEG from the same shot in the > camera. Yes. I guess if you were to shoot hundreds of shots a day, and had big memory cards, then it could be useful to have RAW + high quality jpeg - and only bother to use the RAW file to rescue shots with exposure or w/b problems. I expect - but I don't know - if the comparison between post-converted Canon RAW and jpeg is less dramatic than for the D70. I suspect that either Canon's in-camera processing is better, or the effect is less evident because of the cameras having stronger AA filter, or a combination of both. > On Tue, 2 Aug 2005 21:00:48 +1000, "Mike Warren" >>How did you get a fine JPEG and a RAW of the same frame? >>It seems to me that you are saying that the JPEG was created >>in the camera. > > The 20D will produce a RAW and a large JPEG from the same shot in the > camera. The Maxxum 7D will also, albeit with the in-camera jpg processing rules. Digital photography, and the 7D in particular, are new to me - so my learning curve is ever-present. However, I have toyed around with this dual-image capability, and have been pleasantly surprised at the choices available. Signaturejer email reply - I am not a 'ten' >The 20D will produce a RAW and a large JPEG from the same shot in the >camera. Also, it *always* embeds a 3MP JPEG in the RAW file. This is what you zoom into in the review, if you shoot RAW-only. I thought my 20D was unsharp when I first used it, because of these 3MP reviews. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< On Tue, 02 Aug 2005 21:24:45 +0000, JPS wrote: >>The 20D will produce a RAW and a large JPEG from the same shot in the >>camera. > > Also, it *always* embeds a 3MP JPEG in the RAW file. This is what you > zoom into in the review, if you shoot RAW-only. I thought my 20D was > unsharp when I first used it, because of these 3MP reviews. No, it's unsharp because it's a Canon... =8^) >> The only thing I've heard said about RAW when dealing with the >> specifics of images is that it can allow the recovery of shadow [quoted text clipped - 24 lines] >tend to (as I do) shoot less images, then a minute or so to deal with >the raw file is no problem. One thing about the images you've shown puzzles me; I understood RAW to be an image unprocessed yet they appear sharper than the JPEGS. I thought the in-camera processing into a JPEG always resulted in some kind of sharpening while RAW didn't. -Rich >>With my D70, this is the difference I see: >> [quoted text clipped - 4 lines] > I thought the in-camera processing into a JPEG always resulted in some > kind of sharpening while RAW didn't. He must have sharpened them in the raw conversion but you can get a sharper image from raw. Here's some tests I did: <http://www.edgehill.net/1/?SC=go.php&DIR=Misc/photography/raw-vs-jpg> SignaturePaul Furman http://www.edgehill.net/1 san francisco native plants >>> With my D70, this is the difference I see: >>> [quoted text clipped - 8 lines] > sharper image from raw. Here's some tests I did: > <http://www.edgehill.net/1/?SC=go.php&DIR=Misc/photography/raw-vs-jpg> The top image (of the wine label) *may* have been sharpened during the raw conversion, the checkbox in RSE to "apply sharpening" was checked, but the sliders for sharpening and detail extraction were set at 0. That should have applied no sharpening - but I don't guarantee it. If my comparisons show more difference than yours, then I suspect that part of that may be in that I used an extremely sharp 105mm macro lens in all of the shots. I was looking at your shots comparing noise and noise reduction, raw vs jpg. I also see more noise on converted raw files than ex-camera jpegs, unless I use noise reduction in RSE, in which case I'm pretty sure that I can get much less noise but still better detail retention and acutance than jpeg. Colour noise reduction isn't needed IMO. Most of the noise on the D70 images seems to be luma noise, not chroma. I print my own using an Epson R1800 printer. Trial and error is required - the printer is unforgiving on mistakes. The printer resolution in dpi is very high. I can certainly see a difference between images shot jpg ex camera and raw-post converted when printed at 10 x 7. I never apply sharpening or detail extraction using RSE. With the D70, for me this seems to result in significant aliasing and artifacts. But I always apply some USM to the tiff before printing. In fact, my impression is that what looks like slightly overdone USM on screen removes the tendency of my prints at 10 x 7 inches to look "too good" - a sort of unexpected smoothness and plastic appearance. In this example of a print: http://www.geocities.com/angels2000photos/print.jpg some USM was applied to the tiff before printing. At 300dpi/300ppi, aliasing seen in the crop is not visible in the print. If I could show you the print at that size, compared with any output I have ever seen from any wet-process print using any film or digital file, then you would see why I am very happy with the D70 / R1800 combination. The prints are stunning. I want a D2x - the possibility to produce prints at 15 x 10 at that quality is very appealing. > With my D70, this is the difference I see: > [quoted text clipped - 4 lines] > images. It doesn't matter what I do, the result is always very similar to > those examples. Looking for contrast, sharpness, and color differences betweeen a RAW and JPEG isn't going to tell you about the merits of RAW vs. JPEG, it's just going to show you the merits of the camera's conversion algorithm vs. your conversion algorithm. The fact that your RAW shots look sharper and with more contrast just means that the 150+ MB Photoshop program, on your multi-gigahertz processer with hundreds (or thousands) of megabytes of RAM is doing a better software conversion (possibly with your input on what you want to see) than the software in your camera, which is likely a few hundred kilobytes (or LESS) in size, running on a CPU with a couple of hundred megahertz, with no input, and having only a fraction of the RAM which you have just on your video card. Not such a big surprise when you look at it like that, is it? : ) steve >> With my D70, this is the difference I see: >> [quoted text clipped - 10 lines] > going to show you the merits of the camera's conversion algorithm vs. your > conversion algorithm. And in the end, the difference is? If the computer can get a better end result than the camera, use the computer to process the file. SignatureStacey I guess if you were shooting your friend at a wedding (the photographer is in the bathroom) and you got a killer shot of him and his bride, but somehow the camera didnt seem to get the wb correct for the shot and its a little underexposed, you may have wished you shot it in raw if your a good raw editor. As there's more hope you can correct it and get it just perfect, than if its shot in jpeg. I usually shoot jpeg (smaller files and quicker workflow for me as I'm not an expert at batch processing etc.) > The reason I ask this is that someone on another > forum said that it's impossible to compare two > cameras to each other (even if they are radically > different cameras) by using only JPEG, that you have > to use RAW images. Is this the case? You can't compare RAW images (at least not by looking at them), because RAW files as such cannot be rendered as visual images. They must first be converted to a format suitable for rendering, such as JPEG or TIFF. However, if you capture in RAW, and convert to JPEG (or TIFF) out-of-camera using the *same software* with the *same settings* you have a starting point for making comparisons that is not influenced by in-camera processing. If you compare JPEG-images created by the camera, different settings for USM and colour saturation can produce huge differences in perceived quality that has nothing to do with inherent quality. > Because the differences are clear in most cases from what I've seen > comparing cameras with JPEG output, and I've never seen RAW images [quoted text clipped - 3 lines] > who say only RAW is suitable for high quality output (to print or a > monitor) are basically crackpots, I would say they are crackpots too, because RAW isn't suitable for output (of any quality). If, however, someone told me that RAW was a better /starting point/ for high quality post-processing than JPEG, I would nod in agreement. > though the mags didn't use that terminology, and that the format is > used by some to lord over the heads of the uninitiated and that [quoted text clipped - 5 lines] > highlight information because the format better preserves dynamic > range of an image. Yes, about 1.5-2 stops in my experience. It is also non-lossy. That matters if you do a lot of post-processing. JPEG-artefacts is pretty invisible if you shoot H-Q and keep your camera's output unchanged, But they tend to become more noticable if you play around with curves and the USM - even if you take care and never re-save in JPEG. > So, when people make the far-fetched claims that JPEG (no matter > what compression level) is no where near as good, on what do they > base it? If you don't post-process, H-Q JPEG is just as good as RAW, and comes in a much smaller package. If you post-process, RAW gives you better dynamic range, more leaway with the W-B, and zero artefacts. The bottom line for me, however, is that by shooting RAW I always have the /option/ to post-process. I have a workflow where I first batch convert all RAW into JPEG using default camera settings (giving me the same JPEGs that the camera would have produced if I had shot in JPEG in the first place). This conversion runs in the backround, when I do other things. Then, if I spot images that I think would benefit from a more manual touch - I grab the RAW and redo the conversion "by hand". > Under what circumstances does RAW really come to the fore? If you ever want to post-process your images. Signature- gisle hannemyr [ gisle{at}hannemyr.no - http://folk.uio.no/gisle/ ] ------------------------------------------------------------------------ Kodak DCS460, Canon Powershot G5, Olympus 2020Z ------------------------------------------------------------------------ snipped >I would say they are crackpots too, because RAW isn't suitable for >output (of any quality). If, however, someone told me that RAW was a >better /starting point/ for high quality post-processing than JPEG, I >would nod in agreement. No, TIFF for high quality print and produce them from RAW and use PSD for post processing so you can use layers. ********************************************************* "It looked like the sort of book described in library catalogues as "slightly foxed", although it would be more honest to admit that it looked as though it had been badgered, wolved and possibly beared as well." _Light Fantastic_ Terry Pratchett >> The reason I ask this is that someone on another >> forum said that it's impossible to compare two [quoted text clipped - 6 lines] >They must first be converted to a format suitable for >rendering, such as JPEG or TIFF. So when I use a program like Olympus Studio or Silkypix that displays the RAW file (I haven't save it to TIFF or JPEG) what is it I'm seeing, what kind of file is it on the screen? -Rich >>> The reason I ask this is that someone on another >>> forum said that it's impossible to compare two [quoted text clipped - 10 lines] >the RAW file (I haven't save it to TIFF or JPEG) what is it I'm >seeing, what kind of file is it on the screen? I can see his point, but I think he's being a bit anal. For all intents and purposes, you are looking at the RAW file. ..but more accurately you are looking at just one of many different possible interpretations of the RAW file. Each converter/importer will show you something different. SignatureOwamanga! http://www.pbase.com/owamanga >>>> The reason I ask this is that someone on another >>>> forum said that it's impossible to compare two [quoted text clipped - 17 lines] >possible interpretations of the RAW file. Each converter/importer will >show you something different. Ah, even with digital there is no uniformity! -Rich >>>>>The reason I ask this is that someone on another >>>>>forum said that it's impossible to compare two [quoted text clipped - 20 lines] > Ah, even with digital there is no uniformity! > -Rich You are clueless. When compressing a 12-bit of information from the sensor to 8-bit TIFF, you are bound to lose information. And you don't linearly compress the data but selectively using curve to decide you want to compress; the highlight or the shadow or special curve for your situation. With RAW, you make your decision later. With JPEG, the camera makes all the decision by itself. Also, when a scene has huge dynamic range and you want to preserve them, you can make two pictures from the same RAW file, one a bit underexpose to preserve the highlight and one a bit overexposed to show the shadows. Then blend them in Photoshop. If in case you severely underexpose a JPEG picture, you would notice that your picture would be very poterize if you have to adjust the level over half a stop. With RAW, the same picture would still look fine and will still be good enough up to 2 stops. This simply demonstrate that there are a bit more information from the sensor which JPEG/TIFF would need to throw away. >Also, when a scene has huge dynamic >range and you want to preserve them, you can make two pictures from the >same RAW file, one a bit underexpose to preserve the highlight and one a >bit overexposed to show the shadows. Then blend them in Photoshop. It is really ridiculous that you need to do this; this should all be done by the RAW conversion process, or with something like PS CS2's "HDR" mode, you should be able to open or save in HDR mode from ACR, with white balance applied (but some of the controls greyed out), to do the RAW conversion without transfer curves, to the HDR format. I don't know how Adobe missed this idea; I automatically assumed that CS2's HDR would be a conversion option for ACR. Even 16-bit could be used for saving, as 16-bit-gamma2.2-adjusted images have plenty of precision and dynamic range possible; all you would need is an adjustment layer that put your own desired transfer curve onto the image, possibly even doing contrast by zones (modulate the high frequency content of compressed zones to emphasize detail. We are in the dark ages of software. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >> Also, when a scene has huge dynamic range and you want to preserve them, >> you can make two pictures from the same RAW file, one a bit underexpose [quoted text clipped - 3 lines] > It is really ridiculous that you need to do this; this should all be > done by the RAW conversion process, With the Curves interface in CS2's camera raw, the need to blend two conversions of the same RAW file is pretty much in the past; if the data is in the RAW file you can get the right tonal curve to use it in one conversion now. > I don't know how Adobe missed this idea; I automatically assumed that > CS2's HDR would be a conversion option for ACR. Not necessary; if a RAW file only has 12 bits of data, going to 16 bit in Photoshop is enough, no? HDR has *way* more range than you can ever get from one RAW file. I've tried a seven-shot bracket sequence and that still leaves a vast amount of unused headroom. SignatureJeremy | jeremy@exit109.com >>> Also, when a scene has huge dynamic range and you want to preserve them, >>> you can make two pictures from the same RAW file, one a bit underexpose [quoted text clipped - 8 lines] >data is in the RAW file you can get the right tonal curve to use it >in one conversion now. Yes, that's a good way of opening, but the "Curves" tool is rather crude; it is spatially inefficient. I frequently find that the pixel-spacing of the interface is way too coarse to make a proper curve; the most important part of the curve to do accurately often falls within the height of only 2 or 3 pixels, and even if you can see a curve bending away from the top or bottom line, quite a bit of it winds up at 0 or 255. I am not the type of person to be pleased for long with software features; I can always see a much better way to do things, and PS "curves" are posterizing, pixelized junk, IMO. >> I don't know how Adobe missed this idea; I automatically assumed that >> CS2's HDR would be a conversion option for ACR. >Not necessary; if a RAW file only has 12 bits of data, going to 16 bit >in Photoshop is enough, no? Why did you clip the part where I said even 16-bit gamma-adjusted data is pretty sufficient for a RAW? >HDR has *way* more range than you can ever get from one RAW file. I've >tried a seven-shot bracket sequence and that still leaves a vast amount >of unused headroom. HDR goes into the hundreds of thousands, or millions, with 1.0 representing 255 in the 8-bit mode. It is a bit flaky in the info tool, though, as values will go negative as you increase brightness with the "exposure" tool. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< > Yes, that's a good way of opening, but the "Curves" tool is rather > crude; it is spatially inefficient. I frequently find that the > pixel-spacing of the interface is way too coarse to make a proper curve; It is; the scale is too large for normal use, probably in order to accommodate extreme use. I'd rather have a button to choose the scale. Still, I have yet to encounter a situation where this prevented me from getting what I wanted out of it, so the imperfection doesn't really bother me terribly. >>> I don't know how Adobe missed this idea; I automatically assumed that >>> CS2's HDR would be a conversion option for ACR. [quoted text clipped - 4 lines] > Why did you clip the part where I said even 16-bit gamma-adjusted data > is pretty sufficient for a RAW? I was replying to the part where you assumed that HDR would be a conversion option in ACR, which doesn't really make sense since there's not enough data in ACR to use for HDR. SignatureJeremy | jeremy@exit109.com >I was replying to the part where you assumed that HDR would be a conversion >option in ACR, which doesn't really make sense since there's not enough data >in ACR to use for HDR. Yes, there is. 16-bit gamma-adjusted loses some highlight precision, and HDR would allow you to make 1.0 the normal clipping point, and keep the highlights. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >>I was replying to the part where you assumed that HDR would be a conversion >>option in ACR, which doesn't really make sense since there's not enough data [quoted text clipped - 3 lines] > and HDR would allow you to make 1.0 the normal clipping point, and keep > the highlights. I have not played with Photoshop CS2's HDR mode. Is it similar to CaptureOne's High Contrast Mode? Neverthless, like curve, the setting is done globally which I have no problem to archieve. But there are situation where you want to brighten the foreground and leave the background exposure as is (forgot to use fill flash). Or doing something like "digital graudate filter." I have not yet seen a RAW program that can do that automatically. >>>I was replying to the part where you assumed that HDR would be a conversion >>>option in ACR, which doesn't really make sense since there's not enough data [quoted text clipped - 6 lines] >I have not played with Photoshop CS2's HDR mode. Is it similar to >CaptureOne's High Contrast Mode? No; it's a data/file format, not a rendering curve. >Neverthless, like curve, the setting is >done globally which I have no problem to archieve. But there are >situation where you want to brighten the foreground and leave the >background exposure as is (forgot to use fill flash). Fixing that, in a literal way, would require painting a mask. You could fake it by masking based on levels, but this, like "fill light" functions, gives strange results if overdone. >Or doing something >like "digital graudate filter." I have not yet seen a RAW program that >can do that automatically. That's why you need a precision data format, and layers. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >>>>I was replying to the part where you assumed that HDR would be a conversion >>>>option in ACR, which doesn't really make sense since there's not enough data [quoted text clipped - 23 lines] > > That's why you need a precision data format, and layers. Masking is indeed labor intensive. Perhaps you should try this technique (The Layer Mark). It works for me. http://luminous-landscape.com/tutorials/digital-blending.shtml All I have to say is with RAW, I don't need to take two pictures with different exposures as if using 8-bit JPEG. I think you know that anyway. You can find the digital graudate filter tutorial via google. It's easy and I like the result. >>>>> I was replying to the part where you assumed that HDR would be a >>>>> conversion [quoted text clipped - 35 lines] > You can find the digital graudate filter tutorial via google. It's easy > and I like the result. I looked at Photoshop CS2's HDR (32bit). It requires 3 and more pictures taken at actual different exposures (different output from the same RAW file won't work). I'd do that if I am taking landscape photos with tripod setup. Neverthless, I discover that RAWShooter seems to be the best of the bunch. > I looked at Photoshop CS2's HDR (32bit). It requires 3 and more pictures > taken at actual different exposures (different output from the same RAW > file won't work). I'd do that if I am taking landscape photos with > tripod setup. I suppose renaming the RAW file twice and using exposure compensation on each one has not occured to you yet? SignatureDouglas, Zero care factor for negative responses from anonymous posters. > I suppose renaming the RAW file twice and using exposure compensation on > each one has not occured to you yet? I suppose limiting your posting to times when you actually know what you're talking about has not occurred to you yet? SignatureJeremy | jeremy@exit109.com > > I suppose renaming the RAW file twice and using exposure compensation on > > each one has not occured to you yet? [quoted text clipped - 4 lines] > -- > Jeremy | jeremy@exit109.com A bit harsh, Jeremy, tho' Doug's wasn't exactly kind either. Here's a quote lifted from Luminous Landscape's treatise on the subject: "Also popular, when multiple exposures aren't possible (for example when there is movement in the scene), is to process the RAW file twice, once for the highlights and once for the shadows. This can't, of course, extract information that isn't in the actual file, but it can do a better job than using the usual raw processing tools currently available. The same blending techniques as are used for merging multiple exposures are also used here, with similar issues arising." So Doug's question, while not diplomatic, is in fact valid, and your put-down seems to indicate that you don't quite grasp it all either. Colin D. > A bit harsh, Jeremy, tho' Doug's wasn't exactly kind either. His condescending attitude does not invite kindness. > So Doug's question, while not diplomatic, is in fact valid, and your > put-down seems to indicate that you don't quite grasp it all either. Go try renaming a single raw file 3 times, changing the parameters in each copy, and doing a Merge To HDR on it. Go on, I'll wait. SignatureJeremy | jeremy@exit109.com > > A bit harsh, Jeremy, tho' Doug's wasn't exactly kind either. > [quoted text clipped - 8 lines] > -- > Jeremy | jeremy@exit109.com I use Canon's Zoombrowser, not the fastest piece of software I own, but I like the way I can alter the parameters in the RAW decoder. I didn't actually do it, but I figure I can copy and rename the RAW file (1 original, 1 copy) in less than a minute, and decode them with the same parameters except for exposure adjust in another two or three minutes. It's not as if one would have to do a bunch of files this way. But, that's a red herring, of course. The argument was whether it was valid to do that with one RAW image, and according to LL it is. Colin D. >>> A bit harsh, Jeremy, tho' Doug's wasn't exactly kind either. >> [quoted text clipped - 19 lines] > But, that's a red herring, of course. The argument was whether it was > valid to do that with one RAW image, and according to LL it is. Why has nobody tried this? I just made 3 psd files from a RAW that had very dark shadows with 3 stops of exposure difference between them. PS CS2s Merge to HDR gave an error message: "There is not enough dynamic range in these photos to construct a useful HDR image". I then proceeded to get an acceptable image by playing with curves. This process did take a bit longer than using HDR. -Mike >>>>A bit harsh, Jeremy, tho' Doug's wasn't exactly kind either. >>> [quoted text clipped - 30 lines] > > -Mike The problem with HDR is twofold. First it reads the EXIF data for exposure and decided the images are identical and spits out the warning. Next it is limited by what Adobe have decided is acceptable for their program's output so what usefulness it might have for an experimenter, can only be harnessed if you make some alterations to the RAW file's EXIF data before conversion. I don't think this tool is very useful in it's present guise even though it has great potential. The "curvemister" claims the whole secret to running a digital darkroom is in curves. I find myself, that if an image is worth the time it takes to edit it, the time it takes is totally irrelevant. I used to make contrast masks to achieve wider range prints with Cibachrome processes. As electronics matured, I scanned tricky images at 2 extremes of EI and merged the results on a computer to obtain wide contrast range pictures. With digital images I frequently develop a RAW file at extremes of EI and obtain exceptional results with masking and merging. I use mask Pro to mask the image(s) and blend the results of up to 4 images, depending on what I seek to achieve. Before I posted the question about using differently exposed RAW files, I had experimented with this process and despite what my pet troll and Jeremy Nixon seems to think, it can be accomplished relatively easily. Like walking, you have learn how to edit photographs and I doubt you'll do that by posting in a camera newsgroup and taking the advise of trolls. SignatureDouglas, You never really make it on the 'net until you get your own personal Troll. Mine's called Chrlz. Don't feed him, he bites! > Before I posted the question about using differently exposed RAW files, > I had experimented with this process and despite what my pet troll and > Jeremy Nixon seems to think, it can be accomplished relatively easily. Yeah, I've done it, too. But "Merge to HDR" doesn't work like that, so your condescending, insulting reply was 100% off-base. SignatureJeremy | jeremy@exit109.com So, Douglas, you've read up on it *now*. Well done - you almost sound as if you know what you are doing.. BUT.... Please explain WHY, when LEO said: > I looked at Photoshop CS2's HDR (32bit). It requires 3 and more pictures > taken at actual different exposures (different output from the same RAW > file won't work) YOU replied: >I suppose renaming the RAW file twice and using exposure >compensation on each one has not occured to you yet? Please, Douglas, explain that to us. In simple language. Then, explain this: >.. it can be accomplished relatively easily. Jeremy and I were clearly, and correctly, referring to the use of Merge to HDR. So, to enlighten us and the group, please give us an example of how you have used a single RAW file to achieve a useful result in Merge to HDR. We all *already* know about curves/layers/blending.... I'm waiting. > But, that's a red herring, of course. The argument was whether it was > valid to do that with one RAW image, and according to LL it is. The Curves interface in Camera Raw makes it unnecessary now, though. SignatureJeremy | jeremy@exit109.com >> But, that's a red herring, of course. The argument was whether it was >> valid to do that with one RAW image, and according to LL it is. > >The Curves interface in Camera Raw makes it unnecessary now, though. It could be much better, though. It is totally ridiculous that you have to under-expose the image on the first tab to squeeze in more highlights! Why not just let the curve tool go past "255" in the x axis? Perhaps the curve tool is applied after clipping, but this is a poor design, IMO. The curve tool should apply to the linear data before it is gamma-adjusted; then the input (x) axis could go past "255". Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< > It is totally ridiculous that you have to under-expose the image on the > first tab to squeeze in more highlights! Why not just let the curve > tool go past "255" in the x axis? Perhaps you're actually complaining about the poor UI decision to make ACR's "highlight recovery" feature invisible? ACR tries to regenerate highlight detail when one or two color channels are clipped, and this is triggered by simply lowering the "exposure" slider, in what seems to be an attempt to simulate having the highlight detail there in the first place. It does this without ever telling you whether the highlights are real or recovered. So, lowering the "exposure" slider actually is different from anything you can actually do with the "curves" UI. Too, there is a naming problem. The "exposure" slider's purpose is actually to set the white point of the data. Then, 255 on the curve is that white point. > Perhaps the curve tool is applied after clipping, but this is a poor > design, IMO. The curve tool should apply to the linear data before it > is gamma-adjusted; then the input (x) axis could go past "255". The curve tool actually does apply to the linear data; that it appears to work on gamma-adjusted data is a UI representation. It does not, however, go higher than "255", which is probably (a) good for the "normal" user, and (b) bad for the advanced user. Perhaps you should send a feature request? "Curves" was added in the first place as a result of requests. SignatureJeremy | jeremy@exit109.com > So Doug's question, while not diplomatic, is in fact valid, and your > put-down seems to indicate that you don't quite grasp it all either. > > Colin D. Sadly, Jeremy has a grasp on something... Anatomical. He is only the second person to go in my kill file in over 14 years of Internet participation. I did that because in the time this group has existed, he has contributed not a single post of any value. All he does is fire a few shots to start a rumble. Quite pathetic, when you think about it. Seems to run common for other from his neck of the woods too. Either I am getting less tolerant of idiots or the number of them now able to use a keyboard is increasing. Perhaps I could have worded my post a little differently but it fits in well with the general line of sarcasm this group has developed. On the subject of Raw Shooter... I spent many days with that developer and every other I could find. I was attempting to develop a file in which about 9 stops of EI were to be found. Raw Shooter does a nice job of developing basically correctly exposed images with maybe 1 stop in the shadows. When you try and pull 2 stops or more from the shadows, it fails miserably with the result showing major flaws in the image. Canon's own developer is not much better. The picture I posted a couple of weeks back of a nighttime wedding portrait has considerable detail in the shadows and correctly exposed highlights. The real prize came when I made differently developed files and obtained an abnormally wide contrast range. No need to have shot different exposures, just develop one image for highlights and one for shadows. SignatureDouglas, Never trust a Photographer with a Guide dog! > He is only the second person to go in my kill file in over 14 years of > Internet participation. Ooh, killfiled by a barely-competent hack with an ego problem for daring to point out that most of what he posts is plainly incorrect. I'm just crushed. SignatureJeremy | jeremy@exit109.com >"Also popular, when multiple exposures aren't possible > (for example when there is movement in the scene), >is to process the RAW file twice.... Colin, did you notice the words "when multiple exposures aren't possible"? And did you read it in context - the quote was referring to simply blending two files with different compensations. It was NOT about using the Merge to HDR function. So, like Mr Nixon, I'm waiting for your response - did you try it, and did it work? >So Doug's question, while not diplomatic, is in fact valid, >and your put-down seems to indicate that you don't quite >grasp it all either. No, it *isn't* valid. (Let's ignore the diplomacy issue.) You can NOT use the same RAW file differently processed, in a merge to hdr process. All you will get is "There is not enough dynamic range in these photos to construct a useful HDR image". Now given that you have told *others* off for not grasping the topic - I think it would be appropriate for you to return and clarify this. And if you are wrong, be grown up enough to admit it. PS - A small addendum. You *can* in fact fool PS into trying it, but only by falsifying the EXIF data to indicate the exposures 'truly' differed. However, you will not get a useful result, and are likely to get that same message anyway. Using the same RAW file defeats the entire point of having the function. > >"Also popular, when multiple exposures aren't possible > > (for example when there is movement in the scene), [quoted text clipped - 4 lines] > simply blending two files with different compensations. It was NOT > about using the Merge to HDR function. Yes, chrlz, I read the entire treatise, and yes, I saw the qualifier you mention. I personally haven't tried it, I merely quoted from the LL article, which appeared to offer different compensations from a RAW file as input to the HDR function. I would expect that the two compensations would have to be quite different, one with the highlights not much brighter than middle gray, and the other with shadows not much denser than middle gray. > So, like Mr Nixon, I'm waiting for your response - did you try it, and > did it work? No, I just read the article, as I said above. > >So Doug's question, while not diplomatic, is in fact valid, > >and your put-down seems to indicate that you don't quite [quoted text clipped - 4 lines] > All you will get is "There is not enough dynamic range in these photos > to construct a useful HDR image". Ok, I accept that. But the article does not say that, so in the absence of actual experience, one would be misled. And I was referring to the article, and did not say or imply that I had tried it. As for validity, in my book *any* question, whether or not badly put, is valid. Statements may or may not be valid, but Pixby's post I read as a question. > Now given that you have told *others* off for not grasping the topic - > I think it would be appropriate for you to return and clarify this. > And if you are wrong, be grown up enough to admit it. I suppose I deserve that for working from reading/hearsay and not actual experience, tho' I suspect I've got a lot of company in that regard. You win. > PS - A small addendum. You *can* in fact fool PS into trying it, but > only by falsifying the EXIF data to indicate the exposures 'truly' > differed. However, you will not get a useful result, and are likely to > get that same message anyway. Using the same RAW file defeats the > entire point of having the function. Aaahh, there's where experience shows. Who would guess that PS would check the exifs for exposure? In hindsight, it's obvious. Thanks for being polite about it. Colin D. > Thanks for being polite about it. No problems, and you just gained my respect by returning (I wonder if Douglas will?)! I don't want to labour the point tooo much (O:, but if you check the article again, you will see that the discussion about the use of two adjusted RAW files is in the first section ('Dynamic Range'). In that section he is clearly discussing alternative methods, ie blending, *before* he introduces the topic of using HDR in the next section ('What is HDR'). And to quote the man himself, he says of HDR: "And, to the inevitable question - no, you can't use different outputs of the same RAW file. They must be separate exposures.." I happily confess I haven't had the opportunity to actually use this function yet myself, but I plan to, and I have done a lot of homework beforehand - I tend to over-research terribly before I make a move... I'll post some results, if/when I ever get around to it.. (O; >"And, to the inevitable question - no, you can't use different >outputs of the same RAW file. They must be separate exposures.." IMO, the fact that people even have to try to get two different EIs of the same RAW capture blended or merged together just illustrates how archaic current software is. Now, we have HDR in PS, but Adobe couldn't think of a very basic way to use it; export from ACR, for a single RAW file. Why should anyone suffer the artifacts of clipping from one image, and blocky shadows from the other, blended together? The best way is to render the full dynamic range of the RAW capture, white-balanced, demosaiced, maybe even saturated, rendered to a linear data format, with a curve-adjusting visual aid layered above it. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< > In message <1123235178.727536.306040@z14g2000cwz.googlegroups.com>, SNIP > The best way is to render the full dynamic range of the RAW > capture, white-balanced, demosaiced, maybe even saturated, > rendered to a linear data format, with a curve-adjusting visual > aid layered above it. Yes, the real challenge is in tonemapping (in a creative attempt to squeeze a high Dynamic Range into the 8-b/ch output range). The Shadows and Highlight adjustment was a first (useful) PS step in that direction. Histogram adjustment is another relatively simple approach but that would need a cumulative histogram with slope limiting to implement it, and could be used on existing images as well (not really a new technique by the way, http://radsite.lbl.gov/radiance/papers/lbnl39882/tonemap.pdf). Bart >> I looked at Photoshop CS2's HDR (32bit). It requires 3 and more >> pictures taken at actual different exposures (different output from [quoted text clipped - 3 lines] > I suppose renaming the RAW file twice and using exposure compensation > on each one has not occured to you yet? That won't work. They have to have a very large difference. -Mike >> I looked at Photoshop CS2's HDR (32bit). It requires 3 and more >> pictures taken at actual different exposures (different output from [quoted text clipped - 3 lines] > I suppose renaming the RAW file twice and using exposure compensation on > each one has not occured to you yet? You must be drunk. We were talking about Photoshop's 32-bit HDR. As for blending, which I said already, that I use two pictures extracting from the same RAW file. >> I looked at Photoshop CS2's HDR (32bit). It requires 3 and more pictures >> taken at actual different exposures (different output from the same RAW >> file won't work). I'd do that if I am taking landscape photos with >> tripod setup. >I suppose renaming the RAW file twice and using exposure compensation on >each one has not occured to you yet? People have done that and been told by the program that there is not enough dynamic range, and the program quits trying to merge them. Of course, it is really ridiculous that you even have to try to do this, when *ONE* raw conversion sans luminance transfer curve has all the same information, without extra posterization and clipping artifacts. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >Neverthless, I discover that RAWShooter seems to be the best of the bunch. Well, its "highlight contrast" and "shadow contrast" sliders apply the type of curve that is very difficult to apply with Adobe Curve tools, with it's coarse interface posterized by the small number of pixels used in the interface, and the 256-value limit. One-pixel resolution makes it extremely difficult to get oft-needed curves that could be expressed much more easily with a different type of interface. The DirectX audio plugins for dynamics that I was using several years ago had dynamics curves that had controls like slope/knee and sharpness for each, as well as gain for the middle line, and endpoints. It would also help if you had an option to have a non-linear vertical axis, scaled like this: 255 254 252 248 ... ... 7 3 1 0 The curves should be algebraic formulas, generated either by direct input, or intuitive creation tools, not 256-entry tables for 15-bit images. Photoshop is very conservative with its tools, and IMO, doesn't have a lot of vision about making things more intuitive and useful. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< SNIP > I have not played with Photoshop CS2's HDR mode. Is it similar to > CaptureOne's High Contrast Mode? No, it is a method that allows to exceed the camera's dynamic range (needs to be done by actually capturing a larger range). The task that follows is then how to fit that huge range of luminances into the limited range of a monitor, or even harder the range of prints. http://www.luminous-landscape.com/tutorials/hdr.shtml Bart >No, it is a method that allows to exceed the camera's dynamic range >(needs to be done by actually capturing a larger range). I still don't see why it wouldn't render a single RAW to it, remaining linear in luminance. That could be useful. PS has too many electric fences put up for no apparent reason, except, "we haven't thought of a reason you might want to do that, so we won't let you". Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >>No, it is a method that allows to exceed the camera's dynamic >>range (needs to be done by actually capturing a larger range). > > I still don't see why it wouldn't render a single RAW to it, > remaining linear in luminance. I've read (not tested it myself yet) that since the Raw's EXIF info (obviously) shows the same exposure time / aperture, the HDR construction fails (assuming (!) too little exposure difference). The test I've yet to do is saving different push/pull Raw conversions as PNG, and using those as input to the HDR function. It all shouldn't do much more than letting the highlight recovery of the pulled conversion kick in, QED. The fact that Photoshop's HDR input is converted to 8-bit/channel input is somewhat of a bummer, so I make sure to use gamma adjusted 8-b/ch JPEGs or PNG's as input. Things work faster that way. Anyway, the real challenge is in tonemapping for true multiple exposure shots (as they have superior Dynamic range compared to Raws and all needs to be fitted in 8-b/ch output modalities). Bart >>>The reason I ask this is that someone on another >>>forum said that it's impossible to compare two [quoted text clipped - 10 lines] > the RAW file (I haven't save it to TIFF or JPEG) what is it I'm > seeing, what kind of file is it on the screen? You are looking at a quick raw conversion with unknown default settings. SignaturePaul Furman http://www.edgehill.net/1 san francisco native plants >>> The reason I ask this is that someone on another >>> forum said that it's impossible to compare two >>> cameras to each other (even if they are radically >>> different cameras) by using only JPEG, that you have >>> to use RAW images. Is this the case? >> You can't compare RAW images (at least not by looking at them), >> because RAW files as such cannot be rendered as visual images. >> They must first be converted to a format suitable for >> rendering, such as JPEG or TIFF. > So when I use a program like Olympus Studio or Silkypix that > displays the RAW file (I haven't save it to TIFF or JPEG) what is it > I'm seeing, what kind of file is it on the screen? If you're using Windows, most likely a BMP. You might think that this is just about the equivalent to a RAW-file - but it isn't. Unless you have a very primitive RAW converter (I'm not familiar with Olympus Studio or Silkypix), you can change the settings in he RAW converter (set a different W-B Kelvin, use a different amount of USM, set different saturation, etc.) - and what you see on the screen is changed as a result of changing these settings. The RAW-file doesn't change just because you change the settings in your RAW-conversion program, so obviously you what you see on the screen isn't the RAW file - but just one (of the many, many possible alternative) conversions of it. Signature- gisle hannemyr [ gisle{at}hannemyr.no - http://folk.uio.no/gisle/ ] ------------------------------------------------------------------------ Kodak DCS460, Canon Powershot G5, Olympus 2020Z ------------------------------------------------------------------------ >So when I use a program like Olympus Studio or Silkypix that displays >the RAW file (I haven't save it to TIFF or JPEG) what is it I'm >seeing, what kind of file is it on the screen? It's a .NOT file. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< >> The reason I ask this is that someone on another >> forum said that it's impossible to compare two [quoted text clipped - 11 lines] > you have a starting point for making comparisons that is not > influenced by in-camera processing. But then you are comparing how this specific software deals with a certain cameras RAW files, not how well the camera works. To be fair IMHO you need to use the RAW software that works the best for each camera, then compare that. > If you compare JPEG-images created by the camera, different settings > for USM and colour saturation can produce huge differences in > perceived quality that has nothing to do with inherent quality. If you work both files to get the "best image quality" and compare that, seems fair to me! SignatureStacey > The reason I ask this is that someone on another forum said that it's > impossible to compare two cameras to each other (even if they are > radically different cameras) by using only JPEG, that you have to use RAW > images. Is this the case? [..] It depends on what you're trying to compare. You'd only get a useful comparison of anything where the in-camera JPEG processing would ruin the comparison. So, say... If you can put the same lens on the two different cameras, you could compare their sensors by shooting RAW, thus avoiding the JPEG processing. Same for lenses... Put two different lenses on the same camera, shoot RAW, compare the differences without worrying about the camera's JPEG compression decisions. If you're trying to compare two cameras, then you'd want to see their JPEG output. If one JPEG is good while the other is trash, then you've made a useful comparison. Whether that matters to you or not is another story. And this is the point at which I wish, out loud, yet again, that camera makers would include PNG as a output choice. 16 bits per channel, lossless compression, and no patent licensing fees. >The reason I ask this is that someone on another >forum said that it's impossible to compare two >cameras to each other (even if they are radically >different cameras) by using only JPEG, that you have >to use RAW images. Is this the case? Well, to me it is logical that if a camera has both JPEG and RAW, that anyone interested in maximum quality would shoot RAW, so RAW should be compared when you're looking for the best the cameras can do. JPEGs clip out-of-gamut colors, reduce mathematical precision in the midtones and highlights, clip away highlights, flatten/darken shadows, and have compression artifacts. Sometimes the difference is not very visible, usually for tame colors in moderate dynamic-range scenes, optimally exposed. Signature<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm> ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< If you do not know why you should use raw you should not use raw. If you understand why you would want to shoot in raw format you will never want to capture images in any other format (unless your memory card is running out of space!). >If you do not know why you should use raw you should not use raw. >If you understand why you would want to shoot in raw format you will never >want to capture images in any other format (unless your memory card is >running out of space!). You'd still "want" to shoot raw though, no? :-) ---------- Ed Ruf Lifetime AMA# 344007 (Usenet@EdwardG.Ruf.com) See images taken with my CP-990/5700 & D70 at http://edwardgruf.com/Digital_Photography/General/index.html |
Enable EMail Alerts
Start New Thread
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads |