I have not noticed that
I have not done any conversions back to the original format.
I was more interested in the final result (does it play, does it sound good when converted from DSD256 to DSD128... etc).
Are you sure you've converted to the same file type (i.e. .dsf or dff)? Usually the .dsf is a couple of kB larger than it's .dff counterpart (as I've noticed when using tools to convert .dff to .dsf - since .dff does not support metadata tags).
Yes, I am quite sure. Tbh, it would have surprised me if the results had shown a 100% lossless conversion as I don't think that is possible. Still, the difference was less than 3MB on a 580MB audio file so I didn't even try to find out if I could hear a difference.
A DFF file can be repackaged into a DSF file 100% losslessly (not talking file size here), but dsd64 and dsd128 etc have noise moved into different ultrasonic frequencies...
What also concerns me is that when converting PCM to DSF, JRiver produces smaller files than the Tascam Editor does. My concern is that JRiver devs may have chosen more short cuts to favor functionality at the cost of quality. Of course I have no proof of this but their refusal to fix a design flaw and unwillingness to favor quality over functionality has made me suspicious.
This just adds to your bug report which they seem unwilling to acknowledge. Because of that bug I stopped using MC to play DSD months ago. Every new MC version I test to see if it was fixed. My Calyx M has basically unlimited storage capacity (SD card and micro SD card slot) and I just moved all my DSD audio onto SD cards, using the Calyx exclusively to play DSD audio.
Not here to bash MC as I see plenty potential but there are some design flaws and bugs (that have been around for a long time) that need to be addressed.
-R.