FROM: "Guillaume Dargaud" SUBJECT: How to use 48 bits images...? DATE: Fri, 16 Feb 2001 13:42:11 -0700 ORGANIZATION: CSU/DAS NEWSGROUPS: comp.periphs.scanners OK, I'm shopping for a slide scanner and most of the latest models can scan 3*10 bits, 3*12 bits or better, thus saving to 48 bits images. I know that VueScan can save 48 bits TIFF images and that 48 bits PNG also exist. So, practically, how does one use a 48 bit image ? I think that Paint Sho Pro reads them but then converts them to 24 bits. If this does not include any kind of color/contrast processing, then it's useless. Somebody else told me that PhotoShop can handle 48 bits images in a limited way. So my question is: those of you who have a more than 3*8 bit scanner, how do you use it practically ? -- Guillaume Dargaud Colorado State University - Dept of Atmospheric Science http://rome.atmos.colostate.edu/ "Give a man a fire, and he'll be warm for a day. Set a man on fire, and he'll be warm for the rest of his life." - John A. Hrastar. FROM: fulton@directlink.net (Wayne Fulton) SUBJECT: Re: How to use 48 bits images...? DATE: Fri, 16 Feb 2001 22:05:22 GMT ORGANIZATION: . NEWSGROUPS: comp.periphs.scanners >So, practically, how does one use a 48 bit image ? >I think that Paint Sho Pro reads them but then converts them to 24 bits. If >this does not include any kind of color/contrast processing, then it's >useless. >Somebody else told me that PhotoShop can handle 48 bits images in a limited >way. Adobe Photoshop, Corel Photo-Paint, and Ulead PhotoImpact (5 & 6 anyway) can accept 48 bit images, and have 48 bit contrast tools (histogram, curve, color balance). The purpose is to retain full data for the those tone altering adjustments. After that tonal adjustment, you would convert to 24 bit color to continue, or to save it. Operations like resampling or sharpening have no use for 48 bit data, they dont shift the pixel tonal values. -- Wayne fulton@scantips.com http://www.scantips.com "A few scanning tips" FROM: "Guillaume Dargaud" SUBJECT: Re: How to use 48 bits images...? DATE: Fri, 16 Feb 2001 15:19:05 -0700 ORGANIZATION: CSU/DAS NEWSGROUPS: comp.periphs.scanners > After that tonal adjustment, you would convert to 24 bit color to continue, or > to save it. Operations like resampling or sharpening have no use for 48 bit > data, they dont shift the pixel tonal values. Yes, that makes sense. Thanks for the answer. I'm used to Paint Shop Pro. I'll have to see if I can do the same with it. -- Guillaume Dargaud Colorado State University - Dept of Atmospheric Science http://rome.atmos.colostate.edu/ "Hollywood is where they shoot too many pictures and not enough ors." - Walter Winchell. FROM: fulton@directlink.net (Wayne Fulton) SUBJECT: Re: How to use 48 bits images...? DATE: Fri, 16 Feb 2001 22:34:24 GMT ORGANIZATION: . NEWSGROUPS: comp.periphs.scanners >Yes, that makes sense. Thanks for the answer. >I'm used to Paint Shop Pro. I'll have to see if I can do the same with it. PSP doesnt do 48 bits. -- Wayne fulton@scantips.com http://www.scantips.com "A few scanning tips" FROM: goober SUBJECT: Re: How to use 48 bits images...? DATE: Fri, 16 Feb 2001 19:21:43 -0600 ORGANIZATION: NewsDudes - http://www.newsdudes.com NEWSGROUPS: comp.periphs.scanners On Fri, 16 Feb 2001 13:42:11 -0700, "Guillaume Dargaud" wrote: -->OK, I'm shopping for a slide scanner and most of the latest models can scan -->3*10 bits, 3*12 bits or better, thus saving to 48 bits images. -->I know that VueScan can save 48 bits TIFF images and that 48 bits PNG also -->exist. Interesting, I had always heard that 36/48 bit scanners use the extra channels to insure colour accuracy and always output as 24 bit images to the imaging software. Maybe I'm wrong or I've misinterpreted things. -->So, practically, how does one use a 48 bit image ? -->I think that Paint Sho Pro reads them but then converts them to 24 bits. If -->this does not include any kind of color/contrast processing, then it's -->useless. -->Somebody else told me that PhotoShop can handle 48 bits images in a limited -->way. I think PSP uses images in the RGB format (max 24 bit) and as such its colour values are completely different from PhotoShop or CorelPaint which can use RGB or CMYK colours values. It is one of the features that prevents PSP from accepted by knowledgeable graphic artists. I found by actually editing images in both products and the colours were always different and it didn't dawn on me until I read an article on it. -->So my question is: those of you who have a more than 3*8 bit scanner, how do -->you use it practically ? -------------------------------------------------------------------- "Quick, look useless! The Borg are coming!" ---Mr. Ryba's quote board FROM: zoomr@mr.net (Alan) SUBJECT: Re: How to use 48 bits images...? DATE: Sat, 17 Feb 2001 09:58:45 GMT ORGANIZATION: Onvoy NEWSGROUPS: comp.periphs.scanners If a 48 bit image is converted to a halftone with a sufficient number of DPI, it will print a greater number of brightness levels in each channel. The short answer is that the quality will show up on a high resolution print. If it were the only deciding factor between two scanners, more bits may indicate a more recent design. It also gives much more flexibility in making a transfer curve from the scanned image to the final image. For instance, you could put more emphasis on both the highlights and shadows without "posterizing" the midtones as often happens with 24 bits. Unfortunately, I don't know of a program that can actually display 48 bits of depth on the screen. Back when VGA only had 256 colors they had dithering software that converted 24 bit images to a dot patterned 8 bit image, but I don't know if applications can do the equivalent with 48 bits. FROM: Kennedy McEwen SUBJECT: Re: How to use 48 bits images...? DATE: Sat, 17 Feb 2001 13:26:22 +0000 ORGANIZATION: A simple twist of fate NEWSGROUPS: comp.periphs.scanners In article <3a8e46a5.5155881@news.mr.net>, Alan writes >If a 48 bit image is converted to a halftone with a sufficient number >of DPI, it will print a greater number of brightness levels in each >channel. The short answer is that the quality will show up on a high >resolution print. If it were the only deciding factor between two >scanners, more bits may indicate a more recent design. > >It also gives much more flexibility in making a transfer curve from >the scanned image to the final image. For instance, you could put more >emphasis on both the highlights and shadows without "posterizing" the >midtones as often happens with 24 bits. > >Unfortunately, I don't know of a program that can actually display 48 >bits of depth on the screen. Back when VGA only had 256 colors they >had dithering software that converted 24 bit images to a dot patterned >8 bit image, but I don't know if applications can do the equivalent >with 48 bits. > Nor will there ever be an application that will display 48bit images, because there are no display adapters that can accept them. There will never be any display adapters that will accept them because there are no displays with sufficient dynamic range to show 48bit images. There will never be any displays to show them because there are no eyes that can distinguish between individual tones on 48bit images! In short - you can't differentiate much more than a million or so colours at any particular level of background illumination. 24bit colour coding already exceeds the ability of your eyes to differentiate adjacent colours, so nobody is going to invest megabucks developing technology that has no benefit over the existing capabilities. Even Billy Gates claims there are advantages for every edition of Windoze - you might not use them but they are there if you need them. 48-bit scanners are a bit like that. You can't see that range of data yourself, but if you need to manipulate the brightness, contrast, gamma or colour then the dynamic range is there so that the final result will reproduce without loss on a 24bit display - ensuring that it still looks as tonally complete and natural as if you had a perfect scan in 24bit colours. -- Kennedy Yes, Socrates himself is particularly missed; A lovely little thinker, but a bugger when he's pissed. Python Philosophers (replace 'nospam' with 'kennedym' when replying) FROM: "Michael Sifft" SUBJECT: Re: Farbraum Dia - RGB DATE: Mon, 26 Mar 2001 21:56:12 +0200 ORGANIZATION: RZ-Online NEWSGROUPS: de.rec.fotografie "Frank Heukelbach" schrieb im Newsbeitrag news:3ABF865D.F8DEF465@t-online.de... > Im Dia-Magazin 4/2000 war auf Seite 70 u. a. ein Beitrag über Farbräume. > Dort wurden genannt: > -Druck/CMYK und Monitor/RGB ähnlich groß > -Film YCC wesentlich größerer Farbraum als die o. g. > -"Auge" als größter Farbraum > > So richtig wurde das nicht erklärt.... Hallo, auf http://www.creativepro.com/story/feature/8582.html ist ein Artikel, der die "weiten" Farbräume ganz anschaulich macht. Es gibt offensichtlich sogar einen eigenen Farbraum ("Ektaspace"), um alle Farben eines (Ektachrome-)Dias zu beschreiben. Dieser Farbraum geht zum Teil sogar über die vom Auge sichbaren Farben (~ CIE-Lab) hinaus. Gruß Michael Sifft FROM: "Dirk Kadler" SUBJECT: Re: 16 Bit linear DATE: Wed, 24 Apr 2002 23:21:42 +0200 NEWSGROUPS: de.rec.fotografie "Helmut Holtrup" schrieb im Newsbeitrag news:aa6ops$h88$07$1@news.t-online.com... > Hallo, > > ich hatte heute Morgen schon mal gefragt. Aber die Fachleute waren wohl > nicvht da:-) > > Beim Scanprogramm von Minolta kann ich u.a. 16 Bit und 16 Bit linear > einstellen. Was ist das für ein Unterschied? Also dann versuche ich das mal: Dein Scanner kann wahrscheinlich keine 16 Bit je Farbkanal scannen, sondern nur 12. 12 Bit je Farbkanal, also 36 bei RGB kann Photoshop aber weder einlesen, noch bearbeiten. Daher werden die 12 Bit auf 16 Bit "gestreckt" Der minimale Wert bei beiden ist "0" Der maximale Wert bei 12 Bit ist 4095 Der maximale Wert bei 16 Bit ist 65535 Wenn Du nun "16 Bit" einstellst, dann wird jeder gescannte Farbwert mit 16 multipliziert Wird "0" gescannt, dann wird "0" als Wert in die Datei geschrieben. Wird "1000" gescannt, dann wird "16000" in die Datei geschrieben. Wird "4095" gescannt, dann wird aber "65535" in die Datei geschrieben. Einfacher gesagt - der zu scannende Helligkeitsereich wird "gestreckt". Dabei gibt es aber ein Problem. Durch das multiplizieren entstehen Löcher. Denn es kann nur Farbwerte geben, die ein Vielfaches der Zahl 16 sind. Bei der "16 Bit" Variante kann es z.B. keinen Farbwert "12" geben. Denn der bei 12 Bit gescannte Wert von "0" ist "0", der von "1" ist dann schon "16". Bei "16 Bit linear" wird nicht multipliziert. Es wird einfach nur der untere Bereich des 16 Bit-Bereiches genutzt. Wird "0" gescannt, dann wird "0" als Wert in die Datei geschrieben. Wird "1000" gescannt, dann wird "1000" in die Datei geschrieben. Wird "4095" gescannt, dann wird aber "4095" in die Datei geschrieben. Die Werte von "4096" bis "65535" werden nicht benutzt. Dadurch sehen die so gescannten Bilder sehr dunkel aus und weiss oder sehr helle Farben existieren in der Datei nicht. Um nun ein "normales" Bild zu bekommen muss auf jeden Fall nachbearbeitet werden. Dazu können verschiedene Funktionen von Photoshop und anderen Programmen genutzt werden. Photoshop multipliziert aber nicht nur so einfach wie die Scannersoftware, sondern kann auch Zwischenwerte errechnen. Dadurch entstehen keine "Löcher" und es sind feinere Abstufungen, vor allem in dunklen Bereichen des Bildes möglich. Scanne einfach mal das gleiche Foto mit "16 Bit" und "16 Bit linear". Du siehst einen grossen Helligkeitsunterschied. Ändere dann bei dem dunkleren Bild (das linear gescannte) die Helligkeit mit den Gradationskurven. Ich hoffe meine Erklärung ist verständlich. Gruß Dirk