Episode 028: Are 8 Bit enough?

Download the Video!
Download the companion file! (…if there is one…. 😉 )

This is a new thing on Meet the GIMP!, an interview. I talk to Joel Cornuz from the Linux Photography Blog about the differences in 8 and 16 Bit postprocessing.

Where is Swizerland?But before that you’ll learn a bit about the Confoederatio Helvetica (Latin) Schweizerische Eidgenossenschaft (German) Confédération suisse (French) Confederazione Svizzera (Italian) Confederaziun svizra (Romansh) Swiss Confederation (English) and


View Larger Map

Geneva, Genève, Genf, Ginevra or Genevra. You’ll learn a surprising fact about how the Swiss cope with living an a country with four different mother tongues and nothing about Chocolate, Cheese and Swiss Army Knifes. Neither about CERN, the Red Cross and all the other institutions in Geneva. As we talk all the time in a video podcast I’ll show you a slide show of Joel’s images from his town and country.

I went into the interview knowing that I am not missing much with 8 Bit only in GIMP. But now I miss the 16 Bit option and hope that GEGL will solve this problem soon. 8 Bit is not a problem with images which contain a lot of colours and which get not much post processing. But if you tweak the curves of a nearly monochromatic image too much you are in trouble because you loose colours by interpolation and rounding errors. And so you end with 100 instead of 256 different shades – and that’s not enough for a smooth image.

As you perhaps recall I have taken ways around this problem by using two differently converted files from one RAW image. This would often not be necessary with 16 Bit.

Joel shows another way around this problem – post processing in cinepaint. That’s a hypertuned image processing engine in an old and a bit rusty GIMP 1.9x chassis. I’ll look into that and do a video about it in the future.

The TOC

00:23 Welcome
00:38 Interview with Joel Cornuz from the Linux Photography Blog
23:43 Linux Photography Website
25:25 The End
TOC made by paynekj

Contact me!

Did you like this interview? Shall I make more of them or do you feel this was a waste of time? Please tell me your opinion. You can leave your comments on this blog or write me a mail.

Creative Commons License This work is licensed under a
Creative Commons Attribution-Share Alike 2.0 Germany License.

17 thoughts on “Episode 028: Are 8 Bit enough?

  1. Pingback: I’m on meetthegimp.org « Linux Photography

  2. Perhaps I dozed off or something but I failed to appreciate the differences between 8 bit and 16 bit. Did I miss some examples showing the difference or what? There were some super images though.

  3. Perhaps we were not clear enough about that.

    8 bit are enough for looking at. Screens and JPEG don’t have more and a lot of printers work also in 8 bit.

    But if you start to mangle the colours you do integer math in 8 bit. And that leads to losses due to rounding errors. You get less than 256 different shades in each channel and that starts to become visible as flat areas. Just try to tweak a curve in an image back and forth repeatedly and observe.

    If you mangle the colours in 16 bit you can loose lots of shades and still have more than you’ll ever need. Convert that to 8 bit and you have a fine image.

    According to an interview with Sven Neumann GIMP will get 16 bit internal engine parts in the near future. The curves tool would be a good start. 😉

  4. Hi Rolf,

    I haven’t seen the video yet (I’m currently downloading it). I just wanted to share a recent tutorial I read that explained the 16-bit vs 8-bit issue VERY CLEAR:

    http://www.photoshopessentials.com/essentials/16-bit/

    The tutorial is based on Photoshop but it doesn’t matter in this case (if you want to have a clear understanding of why 16-bit is better).

    All the best,
    Jorge

  5. Rolf

    Thanks for the great interview. Just a quick thumbs up for Joel’s website – a great resource. It was invaluable when trying to set-up my Spyder2Express in Linux. Lots of valuable information.

    Cheers
    Ed

  6. I do understand the maths involved and have seen how using the curves tool reduces the number of shades available. I suppose the number of shades available becomes significant if big enlargements are required to be printed or projected but, for the amateur whose images are generally viewed either on a computer or some other digital device, what is the likelihood of such differences being noticeable? I would like to see an image, produced by an amateur photographer, where it is necessary to make such big changes that the reduction in available shades becomes significant.

  7. Thanks Rolf for another great tutorial/interview. The 8 bit problem is why I mostly shoot raw and then use UFRaw for my initial edits to keep edits mostly in 16 bit. Then using Gimp for just fine adjustments so I don’t knarl up the shot with changes. I’m looking forward to a 16 bit gimp. I think photoshop will get a big hit from Gimp if it goes to 16, as that is the biggest argument against using gimp.

  8. Hi there,

    I think Rolf did a good job a interviewing and editing. I was less than average at explaining 🙂

    @Norman: this is all part of your workflow: if you start with a RAW file and edit / print in 16 bits, you will see a difference from a JPEG from camera edited in 8 bits.
    Now HOW MUCH of a difference will vary on how tricky the image is and how much post-processing is necessary.

    Take care,

    Joel

  9. Thanks for the interesting interview. Even though I doubt I’ll ever really need the 16-bit, I’m keeping an eye on Krita and PAINT.NET, both of which are 16-bit app.s, if I understand correctly.

  10. Well, I don’t NEED 16 bit (it worked up to today with 8), but I WANT 16 bit to have the leeway. But I’ll don’t switch programs – I’ll wait for GIMP to catch up.

  11. Pingback: Meet the GIMP « Ang Pilipino GIMP

  12. Hi there, great episode, I have truly enjoyed the last episodes I have seen (I started a couple of weeks ago with episode 002, and gonna get soon up to date). Great explanations Joel. Great episode -again- Rolf

    I’ve just downloaded cinepaint, I’ll compile it and give it a try, all the same I’ll be waiting for the Gimp to be 16 bit (I love it!)

    In the meanwhile, the point of Rolf seems good enough tho me; shooting RAW pics and working on RAW, using the Gimp for the fine details…

    Saludos,

    Jaims

  13. If I start my ubuntu instead of winxp, do not need to compile a thing, just to apt-get the .deb. I don’t get the last version of the thing, but it works fine.

    I’ve seen that the ‘add layers’ and ‘add mask layers’ tools don’t work like in the Gimp; has anyone found good tutorials or resources for cinepaint?

    Not that I’m gonna to switch to cinepaint from Gimp, but learning a bit can’t be harmful 🙂

    Jaims

  14. Pingback: The p-Code Machine » Meet The GIMP

  15. Pingback: Цифровая фотография - » MTG: Обзоры выпусков за весь 2008 год!

Anything to add from your side of the computer?