How "lossy" is the Canon 300D's JPG implementation

[vpi]

Although I shoot mostly in RAW (so let's not make this in a

discussion about RAW vs. JPG), a friend of mine asked me this

question. Does anyone know how "good" or "bad", or how "lossy" the

JPG algorythms are that Canon uses? If I set the 300D to "L

superfine", is that, e.g., the equivalent to storing a JPG in

Photoshop at quality level 8 (High) or 12 (Excellent)?

[alexdi]

I could guess, but you could more easily find out. Shoot a same scene in RAW, and then in superfine JPEG. Import the RAW file in Photoshop and save it as level 12, or whichever size you intend to compare.

 

The larger the resulting JPEG, the less information it has thrown away.

 

Or, if you're feeling particularly indolent, you could browse DPReview and Imaging Resource. Both have JPEG comparisons. Suffice to say, the 300D shows no artifacts set to superfine.

 

DI

[eric bond]

For most people, it is extremely difficult to tell much difference between a RAW output file and a .jpg at Large/Fine.

[kin_lau]

Just shot the screen in RAW, and then use the RAW converter to get an uncompressed tif, a "L superfine", then use Photoshop to convert the tif to a jpeg @ the 12 - 8 settings, and compare.

[tommyinca]

I think the content matter. Simple scenery will not stress low compression JPEG. Try the comparison(RAW & JPEG) on a busy scene with high dynamic range.

[wilfred_m_rand]

No data is 'lost' on the initial exposure, but you will experience is the step down from 16-bit (RAW) to 8-bit (jpg) - which is a loss of color gamut, not pixel data. It's when you open in jpg and resave in jpg that you begin tossing away data. So: after opening your jpg files, resave in tiff with lossless LZW compression.