I have been working with Sloan filters for some time and occasionally post i'r'g' images mapped to RGB. I calibrate each channel using AMASS Sloan magnitudes for field stars - and UCAC4 provides plenty to use. This allows for a much more detailed calibration than using a single G2V star.
Sloan filters have advantages and disadvantages over normal imaging RGB filters. The advantage is that the i'r'g' spans a wider part of the spectrum and pulls in signal from the infra-red. They also split into well-defined non-overlapping spectral bands so there is no cross contamination of signal into two channels. Finally they can be calibrated accurately using available sloan magnitudes - so there is no conversion from sloan to BVI magnitudes.
A disadvantage is that imaging rgb filters have a notch or gap in the spectral coverage designed to block light pollution. Sloan filters - and other photometric filters - would not have this gap - so they would possibly be more noisy. Another issue that I don't think is important is that some filter sets are designed to be "balanced" so the exposure can be the same in each channel. But this is done just by blocking out part of the spectrum received by some of the filters - to reduce its signal. This makes no sense to me since the "signal" may have an unknown spectrum - and either way you can "balance" just by taking fewer sub-exposures in a given channel.
For nebulosity, Sloan i'r'g' images will have both Nii and H-alpha emission in the green - which is a very different "look" from the way they are normally presented. But it shouldn't be too odd since H-alpha is mapped to green in the Hubble palette - and in fact the eye would be most responsive to the H-Beta in the green.
The main reason I do i'r'g' is that I can get calibrated, deterministic colors that have physical meaning. For stars you can immediately relate them to the H-R diagram, and for galaxies it allows direct comparison of one galaxy's color to another. Two things that show up are the number of faint, very red foreground stars that show prominently in red - while in normal rgb images they would not look very different. Another is that distant galaxies are recognizable by a fairly consistent yellow-orange color.
Example i'r'g' images are at the top of: http://www.astrogeek...ages/index.html
. The m83 image shows the benefit of Sloan filters in distinguishing the features near the nucleus - showing the color difference between the Hii regions in green and the stellar nucleus in yellow orange.
I think Sloan measurements are more common, but BVI measurements are still being made. I mainly like the fact that i'r'g' breaks the spectrum into well defined pieces - but the overlapping curves of BVI may be a better fit to visual response - and would allow more blending of hues since emission lines can appear in two channels at the same time.