Thanks to all for your comments. As I already stated, the first conclusion of my test is that even my home printer can show a difference between sRGB and aRGB. Former tests had already shown that with 6 inks, it could compare very well with sRGB from web labs. And in many cases I tend to prefer the subtler rendition of aRGB.
The second one is that when shown to non expert viewers, many cannot see any difference, and if they see it their choice may vary... Remember that this result should not be generalized as I am far from the best testing conditions. I have always been convinced that a wider gamut is not necessaritly a better one. If you try an analogy between color and tones, a wider gamut of tones (HDR) results in flatter images if you don't use tone mapping, S-curves or other ways to regain local contrast.
What is more striking is that other factors are far more important, even if understimated for colour accuracy or richness:
- drying time of the ink
- quality of viewing light
- precision of white balance
Beside the well-known influence of monitor and printer calibration for paper and ink, any variation in in-camera settings or camera profiles in ACR will yield much more different color results than the choice of color space. For Canon user, choosing the camera profile in ACR (neutral, faithful, portrait, landscape...) produces drastic changes!