I'd like to comment on a couple of things that I've found make a big difference in this process.
First is that I Shoot RAW. RAW offers some significant image quality advantages over jpeg. Not the least of which is the ability to correct for lens distortions prior to stitching. I import all my RAW files into Lightroom and use the automatic lens correction built in. It's painless. I have a preset that applies the appropriate corrections based on the lens identified in the EXIF data. The image is corrected for distortion, CA and vignetting automatically. I find that when distortion is corrected for images made with short focal lengths they stitch better. Images made with longer focal lengths stitch well corrected or not.
I always bracket. Most of the time I shoot -2, -1, 0, +1, +2 even if the histogram on the camera says I'm good. I find that Shadows render much better when they are exposed longer and I very very often blend in highlight detail that would have been lost on a normal exposure.
Next is that I use Autopano Giga from Kolor (www.autopano.net
). I have not used PTGui or Arcsoft so I can't comment on how well they work or how easy they are to use but I can tell you that Autopano builds magnificent mosaics whether you are stitching 2 frames or 2000. It can also handle bracketed images seamlessly. With one click Autopano will separate all the bracketed frames into separate stacks and then allow you to output a separate file for each bracket All stitched identically. This is great for combining for HDR or layering in Photoshop. Autopano also has tools for easily correcting distortions such as convergence. It's an easy program to install and learn and I believe worth a small investment in time to try out their trial version.
For HDR processing I haven't found anything that works better than Oloneo PhotoEngine. PhotoEngine can take in 7 100megapixel images and let you process them in real time without skipping a beat. I just did an image that was 5 separate 10,600 x 16,300 pixel files and Oloneo read them in as cleanly as if they were single frames. Try those file sizes with Photomatix or Nik.
I love large format film. I worked with large format for 30 plus years and can say that the view camera experience has always been thing of joy. I won't pretend for a moment that stitching is easier or replaces large format film. But it does have some very real advantages.
I can carry one camera and use it for casual spontaneous photographs or for deliberate calculated stitched mosaics. I don't need to make two trips to the lab and spend $5 for film and processing for each press of the shutter release. When I leave the location I know that I have captured the image and I don't have to worry about film being ruined going through airport security or having someone open the film box not understanding the consequences. For any focal length you can increase the field of view by stitching additional frames without changing the perspective. I no longer need to send my negs off to be drum scanned or mess with oil mounting on my flatbed scanner. I can easily capture wide subject brightness ranges by adding additional exposures. But the most compelling reason I have found is that I am able to produce stunning prints at my favorite sizes (20 x 24 and larger.)
In many ways shooting stitched mosaics exercises the same brain functions as shooting with a view camera. You really have to pay attention to the details of what you are doing. You can't see the final image through the viewfinder so you have to use your mind to visualize what you are shooting. It becomes a slow methodical experience that I truly enjoy.
One issue with using a DLSR for stitching is that many lenses don't stop down to those extreme apertures that you have on large format lenses. Along with that there are perceived diffraction issues with using apertures smaller than say f/16 on my D300. I say perceived because you have take the entire system into account when considering diffraction. That means the print size and viewing distance. When stitching images the magnification of the final print is reduced and thereby reduces the effects of diffraction.
But depth of field is a real issue. Where I could easily stop down to f/45 or smaller on a large format lens I can only stop down to f/16 on my 85mm. That makes the plane of focus critical. I used to make assumptions about actual depth of field and just stop down to cover it. Now I use a DOF application on my iPhone and a laser distance meter to measure everything. It adds more time and complexity but even changes in the plane of focus of 6 inches can sometimes mean success or failure when f/16 is your minimum aperture. A laser distance meter is an excellent tool to add to your bag.
Below is an image that is from 125 stitched frames (5 rows x 5 columns x 5 exposures.) I found that I was not able to capture the field of view I needed using my normal 3x3 matrix because I was already against the wall behind me. Adding a couple of columns and a couple of rows solved my problem. The length of time to capture all 125 frames was 38 minutes. Seems like a long time but I always found I averaged about one image per hour with the view camera.
Sorry for the long winded post but this is a subject that is very dear to me.
All the best.