Results 1 to 13 of 13

Thread: New to RAW Format

  1. #1
    Kevin Mersereau
    Guest

    Question New to RAW Format

    New to RAW format. I'm wondering if I am able to send my pics as raw or do I convert to JPEG for processing.
    Thanks for any help

  2. #2
    Dave Phillips
    Guest

    Default

    Hi Kevin and welcome.
    Keep in mind that RAW is not an image until it is processed in a RAW converter,
    either to .tiff or .jpg. RAW is just the "raw" data for an image.

    The advantage is that you have much more in depth processing capability
    with RAW, highlight/shadow recovery, white balance etc.

    Generally only the fine tuning touch up is done after conversion to tiff or jpg

    this info is only generalized, but hth

  3. #3
    Kevin Mersereau
    Guest

    Default

    Thanks for the help Dave.

  4. #4
    Lifetime Member Doug Brown's Avatar
    Join Date
    Jan 2008
    Location
    Albuquerque, NM
    Posts
    11,879
    Threads
    917
    Thank You Posts

    Default

    Hi Kevin. Your question is a good one. RAW files, as Dave points out, contain much more information than JPGs do. RAW image processing programs such as Lightroom and Adobe Camera RAW let you manipulate white balance, exposure, saturation, contrast, sharpness, etc. in a non-destructive way. In my workflow, I export processed RAW files as PSD files into Photoshop; this preserves the full bit depth of the RAW file. Converting to JPG for editing in Photoshop drops you down to 8 bits.
    Upcoming Workshops: Bosque del Apache 2019, Ecuador 2020 (details coming soon)
    Website -
    Facebook - 500px

  5. #5
    Connie Mier
    Guest

    Default

    Not to hijack your thread Kevin, but after reading these comments and remembering some previous ones, I am a bit confused on which is a better format for post processing and have some questions of my own. Someone posted a while back that sharpening is best performed in jpeg, not RAW. Dave's comments suggest some processing is performed in RAW and others (not sure what specific fine tuning he is referring to) are done in jpeg. While Doug suggests that RAW is the way to go for all editing, if I read him correctly.

    I do all my PP in RAW, but now wonder if I should convert to jpeg for some things. Can the experts kindly clarify on what works best and why it works best. Thank you and all comments are appreciated.
    Connie

  6. #6
    BPN Viewer
    Join Date
    Jan 2008
    Location
    Orlando, Florida
    Posts
    993
    Threads
    166
    Thank You Posts

    Default

    Connie, I do most of the PP in RAW in NX2, then convert to a Tiff for any other processing I want to do in CS3. I save the Tiff for any print output and then create a jpeg in sABG for web use. RAW and Tiff loose no information when processing with them. Jpeg's on the other hand loose some information as you PP and save. The RAW file has 4000 more bits of information than a Jpeg. Hope that helps.

  7. #7
    BPN Viewer
    Join Date
    Nov 2008
    Location
    sacramento california
    Posts
    500
    Threads
    57
    Thank You Posts

    Default

    Hi Nancy

    What is sABG?
    What PP do you do in CS3 that you don't do in NX2? ( I use NX2 also}
    sorry to continue the hijack of the original thread.
    Thanks Ray

  8. #8
    BPN Viewer
    Join Date
    Jan 2008
    Location
    Orlando, Florida
    Posts
    993
    Threads
    166
    Thank You Posts

    Default

    Quote Originally Posted by Ray Rozema View Post
    Hi Nancy

    What is sABG?
    What PP do you do in CS3 that you don't do in NX2? ( I use NX2 also}
    sorry to continue the hijack of the original thread.
    Thanks Ray
    Ray, sRGB, sorry I wrote it wrong:p, is a color space. It is the color space mostly used for the web. The other mostly used color space is RGB. RGB has more color gamut than sRGB. This is a nice link for more info http://www.drycreekphoto.com/Learn/color_spaces.htm . As far as working in CS3. I do my cropping and print from CS3. Plus I have several of Niksoftware plug ins for CS3 that I use, http://www.niksoftware.com/index/usa/entry.php .

  9. #9
    Banned
    Join Date
    Feb 2008
    Location
    Colorado
    Posts
    3,949
    Threads
    254
    Thank You Posts

    Default

    Some clarifications (and probably more than you wanted to know about raw?).

    Raw format out of a digital camera actually is an image format, as much as any other image format, it is just a little strange (and there are even stranger ones in scientific imaging). A digital camera raw file is like three images interleaved. You can separate them and have 3 separate images, for example.

    Raw files record a linear response to light and are typically 12 or 14 bits, covering a range from 0 to 4095 (12-bits) or 0 to 16383 (14 bits).

    When you convert to "standard" image formates, including tiff, jpeg, psd, etc, a variably gamma curve is applied which compresses the high end and boosts the low end. At the high end there is technical loss in any of these formats, including 16-bit tif and 16-bit psd (which is really 15-bit). I say technical because there is actual loss of information, but all consumer digital camera output contains enough photon noise to mask that loss.

    The point of the above is that if you have saturated the high end, the act of conversion saturates it some more. If you convert to 16-bit tiff with a linear conversion, you have a better chance of recovering more highlights. Note that photoshop does not have a linear conversion option. (Not sure of other photographer type raw converters; I use DCRAW and ImagesPlus when I want linear output).

    If you convert to 8-bit tif, you lose a fair amount of information, particularly in the highlights and a little in noise at the low end (mid-tones too, but less noticeable). If you do an 8-bit jpeg conversion, you lose more than the 8-bit tif, and it depends on the compression factor. Jpeg losses are greatest at the low end where you see the effects less (it is a pretty impressive algorithm).

    If you are doing raw processing, select 16-bit output, which will give you the least losses. Note you can do jpeg 2000 and 16-bit output which gives a more precision than 8-bit jpeg but not as much precision as16-bit tif or 32-bit floating point.

    32-bit floating point would be the ultimate in precision.

    Photoshop: if you saved your raw file as a 16-bit tif with a good raw converter, when you open it in photoshop, it is read in as 15-bits. Photoshop uses what is called signed integers, so one bit is used for negative numbers. A 16-bit photoshop file has a max value of 32765, not 65535 for true 16-bits. So saving and resaving photoshop 16-bit psd or 16-bit tif files will degrade the numbers, although probably not anything you would ever notice as a photographer (it's not good for scientists trying to have accurate data).

    Further reading:

    Digital Camera Raw Converter Shadow Detail and Image Editor Limitations:
    Factors in Getting Shadow Detail in Images

    http://www.clarkvision.com/imagedeta....shadow.detail

    Personally, I convert all my digital photo images to 8-bit jpeg for preview (or shoot jpeg +raw). The jpegs will always be there for future quick reference and are not dependent on a particular raw converter. I archive copies of DCRAW, including source code so I can be sure I can always read my raw files in the future (it is a simple format and I have written my own extractor of raw data too before raw converters caught up with most recent models).
    For work on an image I intend to print or publish, I always save my raw converted images as 16-bit tiffs or greater (32-bit floating point FITS, a scientific format, for special applications). I process photographs with 16-bit work flow until the very end, making the file for printing. That means all sharpening and editing in 16-bit (once again, in photoshop, that really means 15-bit). Here is my digital work flow:
    http://www.clarkvision.com/photoinfo/digitalworkflow

    Roger
    Last edited by Roger Clark; 10-02-2009 at 08:37 PM.

  10. #10
    BPN Viewer
    Join Date
    Nov 2008
    Location
    sacramento california
    Posts
    500
    Threads
    57
    Thank You Posts

    Default

    thanks for the clarification, I think.

    Ray

  11. #11
    BPN Member
    Join Date
    Sep 2009
    Posts
    3,867
    Threads
    169
    Thank You Posts

    Default

    Roger, thank you for the very instructive reply.

    Some follow-up questions please if you don't mind.

    Quote Originally Posted by rnclark View Post
    Some clarifications (and probably more than you wanted to know about raw?).

    Raw files record a linear response to light and are typically 12 or 14 bits, covering a range from 0 to 4095 (12-bits) or 0 to 16383 (14 bits).
    Why is there a loss when converting to a 16-bit (or 15-bit for PS) tiff or psd which is wider than the
    12 or 14 bits of the raw format?
    Or is the loss because of the gamma curve you mentioned?
    Are there any algorithms for a lossless coversion?

    Quote Originally Posted by rnclark View Post

    Photoshop: if you saved your raw file as a 16-bit tif with a good raw converter, when you open it in photoshop, it is read in as 15-bits. Photoshop uses what is called signed integers, so one bit is used for negative numbers. A 16-bit photoshop file has a max value of 32765, not 65535 for true 16-bits. So saving and resaving photoshop 16-bit psd or 16-bit tif files will degrade the numbers, although probably not anything you would ever notice as a photographer (it's not good for scientists trying to have accurate data).

    Roger
    Just out of curiousity, does a signed integer serve any purpose here?
    Does a negative value mean anything in a pixel representation?
    Also, I'm not sure I understand why resaving a photoshop psd would degrade further (the damage has already been done by the one bit loss in the first save)


    Thanks again.

  12. #12
    Banned
    Join Date
    Feb 2008
    Location
    Colorado
    Posts
    3,949
    Threads
    254
    Thank You Posts

    Default

    Quote Originally Posted by Sidharth Kodikal View Post
    Roger, thank you for the very instructive reply.

    Some follow-up questions please if you don't mind.

    Why is there a loss when converting to a 16-bit (or 15-bit for PS) tiff or psd which is wider than the
    12 or 14 bits of the raw format?
    Or is the loss because of the gamma curve you mentioned?
    Are there any algorithms for a lossless coversion?
    Sidharth,
    Yes, you are correct, it is due to the gamm curve applied. The gamma curve compresses the highlights into too few bits. Here is more detail:
    http://www.clarkvision.com/imagedetail/raw.versus.jpeg1
    Look at the lines for linear to 16-bit standard curve conversions. Any value above 1 is a loss.


    Quote Originally Posted by Sidharth Kodikal View Post
    Just out of curiousity, does a signed integer serve any purpose here?
    Does a negative value mean anything in a pixel representation?
    Also, I'm not sure I understand why resaving a photoshop psd would degrade further (the damage has already been done by the one bit loss in the first save)
    Photoshop uses some math tricks to speed processing using 15-bit numbers, witch was much more important 10+ years ago when computers were much slower.

    There is a technical advantage for negative numbers because for a zero signal, you always have noise, and you want the average for no signal to be zero, but any one pixel will fluctuate due to noise. Most camera manufacturers get around the problem by adding a small number to the digital values. For example, a camera may add 128 to 12-bit numbers. Nikon does not add an offset, so their noise floor gets truncated at zero. It makes it hard to detect a true zero level because with the truncation and noise, any average you do is always above zero. This is important in astrophotography where one adds many exposures together to detect faint detail like that in galaxies. For ordinary photography it is hard to notice the difference.

    Quote Originally Posted by Sidharth Kodikal View Post
    Thanks again.
    Glad to help.

    Roger

  13. #13
    BPN Member
    Join Date
    Sep 2009
    Posts
    3,867
    Threads
    169
    Thank You Posts

    Default

    Thanks Roger.
    Your website it filled with excellent articles!
    I will have to chew on this a bit more before I understand it better.

    Best,
    Sidharth

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Web Analytics