As I mentioned in a previous post, I recently transitioned from outsourcing my prints to a hybrid digital/silver service bureau, to printing in-house on my own archival pigment printer. I'd been using Digital Silver Imaging and Miller's, and was reasonably happy with the quality of the prints from both sources. The tones and gradients were very good (neutral and smooth), and the deep blacks were also excellent. The one nagging issue was that the prints were generally on the soft side, which worked fine for some subjects at some scales, but they could badly break down on some images in areas with high fidelity detail (as above).
The scan on the left is a digital/silver print. It's showing fairly extreme softness and a lack of fine detail in the vegetation. The scan on the right is a print from my Epson P800. It's showing far better detail, clarity, and separation of tones. These are magnified scans, and these differences might not be noticed by the average person at typical viewing distances, but for the type of work that I do (F64 style landscapes), this level of detail is a real benefit (and more satisfying on a personal level too).
Here's a related bit of information for my photographer friends who do their own printing. There's a long standing debate regarding how to handle print resolution when printing out of the Lightroom print module. No less an authority than Jeff Schewe (author of The Digital Print) recommends upsampling to the native resolution of the printer when printing out of Lightroom (360/720 for Epson, 300/600 for Canon). For images under 360ppi he's recommending upsampling to 360ppi, and for images between 360-720ppi he's recommending upsampling to 720ppi. On the other hand, both Martin Evening (author of The Adobe Photoshop Lightroom Book) and Julieanne Kost (Principal Digital Imaging Evangelist, Adobe) are recommending leaving the "Print Resolution" box unchecked, which is, in effect, setting the output resolution to "Auto". So, we have three of the top experts on Lightroom in disagreement on this critical point. On paper, Schewe's method makes the most sense and seems to provide the most control. Interestingly though, I've been running dozens of test prints using these and other output resolution settings, and in most cases, I have to agree with Evening and Kost; I've gotten consistently cleaner and more highly detailed prints with the resolution box unchecked. I can't (yet) explain technically why this is the case, but I'm a big fan of empirical evidence when it comes to things like the look-and-feel of prints.