Sorry, I don't read this board as often as I used to before the R1 Yahoogroup was created.
I checked the dulcimer recording you mentioned and I get an approximate S/N ratio at the beginning of the take (before the instruments play) of approximately 48DB. This is measured from the highest peak during the portion where the music is playing to the higest peak in the portion before the music begins. HOWEVER, the predominant noise is what I believe is a combination of low frequency noise (traffic rumble, etc) and probably some strings still vibrating from a previous start (a dulcimer does not become totally silent right away). I do not believe this is noise from the R1 input section. If I remove all noise below 200Hz, the S/N ratio becomes about 57DB, but that still includes noise that does not look to me to be from the R1. Note that I record in my living room in a quiet neighborhood but it is still in an urban setting so there is distant traffic rumble and various household noise that sneaks in. Plus, I do not always wait for all strings to be totally silent since they are completely masked when everyone starts playing.
To get a better picture of S/N ratio we need a quieter room and a definition of the average SPL of the sound source. In the mean time, I still think the S/N ratio is around 56DB for the kinds of music that I record.
In one area I do agree with you.... the R1 is probably not suitable for recording very quiet noises. But for most recordings I feel it is at least adequate for a sorta pocket sized unit. Especially considering what else is available.
Try some simple powered mics. I think they probably have a higher output and would make the R1 noise floor coorespondingly lower. I tried a simple stereo lapel mic from CompUSA and found that it had much hiher output than the internal mics. Unfortunately, it also sounded truly awful.