a Question about 16bit RAW

[fstop11.net]

I do alot of printing and I wondered would the image quality improve if I was opening Files from a Canon 20D in 16Bit mode ? This might not be the right place to post this but I understand theres a lot of people here who seem to know what they are talking about regarding 16bit.

Thanks Johnny.

[munchy]

You may notice less banding in areas with a fairly constant colour (a blue sky for example). I have noticed that pics are perhaps a little more forgiving in 16bit mode when you start to post process them.

AFAIK the 20D records in 12 bit. When you open a file in 8bit mode, some information is discarded. When you open in 16bit, you are retaining all the information taken by the sensor at the time of exposure (compared with 8bit mode).

The answer is almost definately try it and see. You may see very little difference.

HTH, Munchy

[fstop11.net]

You may notice less banding in areas with a fairly constant colour (a blue sky for example). I have noticed that pics are perhaps a little more forgiving in 16bit mode when you start to post process them.

AFAIK the 20D records in 12 bit. When you open a file in 8bit mode, some information is discarded. When you open in 16bit, you are retaining all the information taken by the sensor at the time of exposure (compared with 8bit mode).

The answer is almost definately try it and see. You may see very little difference.

HTH, Munchy
Thank you Munchy; I will :)

[vjack]

Most of what I've read on the subject recommends using 16bit, especially if you are doing any post processing. I always do 16bit RAW and convert to 8bit only if I know I'm going to use filters that require 8bit.

[tim]

As with most things in photography, the best advice is to "try it and see". Take a photo, open it in 8 and 16 bit, change levels, change contrast, and change the color of the whole photo in some small way. Convert to JPG (which is 8 bit but non-linear) and print both at a reasonable size. If you can't see a difference then 8 bit is fine.

I work in 8 bit unless I know i'm doing a lot of color manipulation. Since all my white ballancing is done in RAW that's very rare, so I work in 8 bit more 99% of the time. My prints look great.

[elTwitcho]

I print in 8 bit and the colours look incredible. My web versions are 8 bit and the colour looks no worse than they did before I converted them from 16 down to 8. Having said that though, I always leave my photos in 16 bit while I'm editing them, because there is alot more latitude for adjustment before you run into posterization or other weird effects than if you were editing in 8 bit.

[RodBarker]

Its best to stay in 16bit as long as you can , why , it carrys the most data information about your image , why throw heaps of pixel information away before you start serious processing, It might not hurt a 6x4 print much but it will with a 16 x 24 .

Rod

[tim]

Its best to stay in 16bit as long as you can , why , it carrys the most data information about your image , why throw heaps of pixel information away before you start serious processing, It might not hurt a 6x4 print much but it will with a 16 x 24 .

Would you care to explain how it hurts the quality of a 16 x 24, Rod, and how a 16 bit workflow will help your typical picture?

[blue_max]

Using 16bit rather than 8bit is rather like using raw over jpg.

If you have a series of greys from 0-100 in 100 steps and wanted 50 to be jet black, you would end up with 50 steps. If you started with 1000...

Graham

[tim]

Using 16bit rather than 8bit is rather like using raw over jpg.

If you have a series of greys from 0-100 in 100 steps and wanted 50 to be jet black, you would end up with 50 steps. If you started with 1000...

8 bit gives you 16 million possible colors (256 cubed), 16 bit gives you billions. If you do a lot of color transforms sure you might be able to see the difference, but I doubt it.

The advantage of 16 bit, to me, is keeping data in the shadows, where the difference between 0-256 and 0-65535 might be important. This is similar to the difference between RAW and JPG, though JPG are 8 nonlinear bits and RAW is 12 linear bits. For your average photo though, you're just taking up double the disk space. Since I shoot RAW and keep them I tend to do a lot of processing in there, and don't need TIFF/PSD unless I make huge changes to the image.

One day, when the 20D RAW format looks like it's going to be dropped, i'm going to have a rather large batch process to run... unless I want to stick with my processed JPGs, which is also quite possible, given some people only ever shoot JPG.

[robertwgross]

I've converted from RAW to TIF at 16 and at 8.

As a general rule, I try to do as much of my tweaking as possible at the RAW stage. If I think I might have to do major tweaking after the TIF conversion, then I will try to stay with 16 as long as possible. Also, if it an extremely important image, then I will likely try to stay with 16 as long as possible. You can still make major changes to TIF 16 in terms of exposure, white balance, etc, or you can do it to TIF 8, but in some images you will get slightly smoother results with 16. In other images, it doesn't seem to matter. Once all possible editing is finished, it gets stored as TIF 8, no matter what.

---Bob Gross---

[blue_max]

8 bit gives you 16 million possible colors (256 cubed), 16 bit gives you billions. If you do a lot of color transforms sure you might be able to see the difference, but I doubt it.

I won't argue with your figures, but on that basis, one may as well shoot jpg.

Raw and 16bit gives me the best chance to ensure I get the best possible shot from my badly exposed pictures! I have much more disk space and time, than photographic ability.

Graham

[tim]

I won't argue with your figures, but on that basis, one may as well shoot jpg.

Not quite. If you shoot RAW you can fix an awful lot there, which reduces the need for 16 bit TIFF. What Bob does above is a good idea, process in 16 bit and save in 8 bit.

I'm not against 16 bit, and i'm not on a crusade, i'm just pointing out that 16 bits isn't always necessary. The key is understanding what's going on behind the scenes, and knowing when you do and don't need to use it. The last huge prints I had done I opened in Photoshop in 16 bit mode, pro-photo color space. I did my color manipulations (which I had a few to do) and sharpening in there, and I still have the 16 bit PSD file. When I sent them to be printed I converted them to Q12 JPG. The prints look fantastic.

Raw and 16bit gives me the best chance to ensure I get the best possible shot from my badly exposed pictures! I have much more disk space and time, than photographic ability.

Yeah I agree, personally I take 10-20GB of photos per month doing weddings, so if I converted them to TIFF that'd be unmanagable. Hundreds of gig!

[blue_max]

That's the rub - if you have a huge amount to do it fails to be fun. I would never keep a file at 16bit either - I would keep the raw intact though in case I got better at photoshop one day.

Fundamentally, it does depend on what you are doing as to whether 16bit is any use. I have no compelling reason not to, so it's part of my workflow.

Hopefully all the arguements for and against have been covered. You just have to weigh up the pros and cons for you.

Graham