8 bit 16 bit

I read that the 'EOS-1Ds Mark II produces images with outstanding colour rendition and dynamic range. It has sufficient resolution to produce files which convert to 50MB uncompressed TIFF at 24 bit colour depth'

Does the 20D or 5D do that as well??

Im not sure people understand what you are asking, but I will take a stab at it.

No, the TIF files resulting from the 20D or 5D would not be 50 meg at 24 bits cd - the total file size would be smaller than that.

Generally speaking, most shooters wont be using 24bcd anyway - 8 or 16 is far more common. (Eight is pretty much the standard.) It might help if you explained why you are asking.

With the 20D/300D, and probably the 1 series, If you shoot RAW it's 12 bits, when you convert it to TIFF in photoshop you can choose 8 or 16 bit. Your eye can't see the difference, but it's good to have the extra information if you're doing a lot of manipulations on the image. JPG are 8 bit.

I thought it was 8 bits per channel for jpeg, so 3x8 would be the 24 bits you've referred to.

Yes.. 24bit is what we cal 8 bit.. (8X3)

A 20D or 1D2 tiff @ 8it (24bit) is roughly 24MB
A 20D or 1D2 tiff @ 16 bit is (you guessed it) twice the size at roughly 48MB

Wait a minute, you're confusing some things.

With 24 bits per pixel, people usually mean 3 channels (R, G, B) x 8 bits per channel.

Almost all digital cameras record 12 bits per channel (except some medium format digital backs which cost $20,000 or $30,000 - they record 16 bits per channel).

Now the JPEG file format doesn't support 12 bits per channel - it only supports 8 bits per channel. So if you save your photos as JPEG, you throw away the four least significant bits.

The TIFF file format supports 8 or 16 bits per channel. If you want to keep all the 12 bits per channel of information that the sensor produced, you should convert your RAW image to a 16-bit TIFF. Note that the four least significant bits of those 16 bits will simply be set to zero.

Conclusion: You have more dynamic range if you shoot in RAW and convert your RAW files to 16 bit per channel TIFF files. Look here for an in-depth description: Tonal quality and dynamic range in digital cameras

If you want to calculate the uncompressed file size, it's very easy:
(width of image in pixels) x (height of image in pixels) x (bytes per channel) x (number of channels)

Remember 8 bits = 1 byte, so bytes per channel = 1 for 8 bits per channel, or 2 for 16 bits per channel.

1Ds Mark II / 8 bpc: 4992 x 3328 x 1 x 3 = 49,840,128 bytes (47.5 MB), 16 bpc: 95 MB
5D / 8 bpc: 4368 x 2912 x 1 x 3 = 38,158,848 bytes (36.4 MB), 16 bpc: 72.8 MB
20D / 8 bpc: 3504 x 2336 x 1 x 3 = 24,556,032 bytes (23.4 MB), 16 bpc: 46.8 MB

Jesper wrote:
Wait a minute, you're confusing some things.

With 24 bits per pixel, people usually mean 3 channels (R, G, B) x 8 bits per channel.

Almost all digital cameras record 12 bits per channel (except some medium format digital backs which cost $20,000 or $30,000 - they record 16 bits per channel).

Now the JPEG file format doesn't support 12 bits per channel - it only supports 8 bits per channel. So if you save your photos as JPEG, you throw away the four least significant bits.

The TIFF file format supports 8 or 16 bits per channel. If you want to keep all the 12 bits per channel of information that the sensor produced, you should convert your RAW image to a 16-bit TIFF. Note that the four least significant bits of those 16 bits will simply be set to zero.

Conclusion: You have more dynamic range if you shoot in RAW and convert your RAW files to 16 bit per channel TIFF files. Look here for an in-depth description: Tonal quality and dynamic range in digital cameras

If you want to calculate the uncompressed file size, it's very easy:
(width of image in pixels) x (height of image in pixels) x (bytes per channel) x (number of channels)

Remember 8 bits = 1 byte, so bytes per channel = 1 for 8 bits per channel, or 2 for 16 bits per channel.

1Ds Mark II / 8 bpc: 4992 x 3328 x 1 x 3 = 49,840,128 bytes (47.5 MB), 16 bpc: 95 MB
5D / 8 bpc: 4368 x 2912 x 1 x 3 = 38,158,848 bytes (36.4 MB), 16 bpc: 72.8 MB
20D / 8 bpc: 3504 x 2336 x 1 x 3 = 24,556,032 bytes (23.4 MB), 16 bpc: 46.8 MB

Thanks I think that answers it, and thanks to everyone who had a stab at it.