HeyMikey, I am using Amarra mini at this stage but also ran the full version for a month's trial before deciding. I may upgrade to the full version later, but its only advantages for me would be the higher sampling than 96kHz and I am currently limited to 96kHz by both my DAC and the Offramp. I note that Dan Lavry holds strongly to the view that upsampling to 96kHz is still better than upsampling to 192kHz, so I am not fretting about that one.
I converted my files using Sample Manager, which you can find at http://www.audiofile-engineering.com/samplemanager/. Use the iZotope options for resampling and redithering. iZotope is the best sounding set of algorithms that I have tried, and it seems that many in the industry use it.
The redithering to 24 bit provided the most significant improvement, more so than the upsampling.
The hassle with Sample Manager is that you can only batch convert around 100 files at a time or some cache fills up and the program has to be restarted.
It was recommended to me that the best process is to redither to 24 bits first, and resample second - which I just accepted and didn't test myself.
I recommend you make yourself a 24 bit version of a few songs, a 96 kHz version and a 24/96 version and then compare them with a 16/44.1 version of the file. Trying this out for myself with a DAC that upconverts to 96kHz, I preferred the 24/96 file. Trying this with a NOS DAC where the sampling rate is switchable between 44.1, 48, 88, and 96, I also preferred the 24/96 file.
For those that are interested in such things, I do employ double blind tests routinely, but in a way that is helpful as opposed to confusing - but to describe it here would be a long post.