That first initial estimate you tried looks a little on the low side...maybe because of stock heads or maybe some data entry error. Here are some examples with the Edelbrock RPM FE heads which have the flow data posted to run good numbers. I like to select small tube header with open exhaust since I find it more closely approximates real world tests by siding on the conservative side.
Crane Cam 340321 (the one you spec'd above as stock)
http://www.cobralads.com/dyno2000/427bb1.gif
Crane Cam 340471
http://www.cobralads.com/dyno2000/427bb2.gif
Crane Cam 341191
http://www.cobralads.com/dyno2000/427bb3.gif
I am not sure how difficult it is, but if you can go to a roller (solid or hydraulic) it opens up the door to cams with more aggressive profiles like
Crane Cam 349541
It looks like all the cams above are fairly close but there is a 33HP difference from the stock one to the roller and 33 is a good chunk of HP to leave on the table.
You probably already know this, but just in case, I find the easiest way to get the cam data is from the number over the black diamond on the page you gave. See that
3 in a diamond...that is the crane part number....it is easier to use than the grind number of C3AZ-6250-AA. You can enter the part numbers here and get the cam spec card
http://www.cranecams.com/spectest.htm. The hardest part of desktop dyno for me was learning how to properly take the data off that card and put the numbers in the right places. The key is to always watch the .050 vs seat timing...I always prefer using the 050. Here is a
tutorial that might help new dyno2000 users who are reading along.
I noticed on that edelbrock page that those heads can be milled to accept 2.19 and 1.73 valves. You might check around and see if that is a good idea. I tend to think it would make the heads flow a little better but it is best to check with a head guy on this.
hope this info is useful
Andy
http://www.cobralads.com