Posted by Tiger-Heli on 23, 2001 at 9:08 AM:
I have been following with interest the threads on 4-way vs. 8-way sticks and Derrick Re’s IC converter/emulator and have some surprising observations:
First, I don’t believe Jack Tripper’s suggestion for wiring the NC contacts in series with the NO contacts will work. I have not tried his idea, but the theory does not look practical.
I think you could make a circuit work with latching relays (not sure if the time delay would affect things), but Derrick Re’s IC is cheaper than the relays and should work better. I think Derrick Re’s idea should work at least as well as a four-way joystick and possibly better than a true four-way. (This is based on my observations from Derrick’s web pages that a true four way may have dead spots at or near the diagonals if the stick is not adjusted properly. An eight-way joystick will always hit the diagonals and Derrick’s software interprets how to pass this information on to the game.)
However, all this is moot because MAME AUTOMATICALLY uses what Derrick describes as “Fast Sticky” emulation. All skeptics can verify this for themselves. Here is what I tried and the results:
First, I am using MAME (DOS) V37B6 on a Pentium 200MMX under a Windoze ’98 DOS Box. I think the ‘Fast Sticky” mode works from V37B4 to present, but I have not verified this. I chose QIX to run the test on because many posts have mentioned the QIX freezing as the joystick passes through the diagonals. I think that “Fast Sticky” mode is used on all 4-way games, but I did not verify this.
Since I have only a standard (gameport, analog) joystick (soon to change ;-) ), I decided to use the keyboard arrow keys to verify this. Actually, I recommend using these keys for anyone attempting to duplicate my results. It is easier to know exactly when a switch is depressed and the behavior should be identical to that of an arcade joystick hooked up to a keyboard encoder or hack. The inputs in Qix were set to the default (UP = UP or JS Y+, etc.)
Here is the test:
Start Qix. Hold down the Fast Draw (L Ctrl button) throughout the test. Hold down the UP arrow key.
RESULT: The Qix starts moving up.
While continuing to hold down the UP arrow key, depress and hold down the RIGHT arrow key. (Up-Right Diagonal depressed).
RESULT: The Qix immediately starts moving right. There is not flashing, stuttering, or hesitation. (This is what Derrick described as “Fast Sticky” mode.)
With the diagonal depressed and the Qix moving right, there are two options: Release the UP arrow. RESULT: The Qix continues moving right. OR, Release the RIGHT arrow. Result: The Qix immediately switches to moving up. There is no hesitation. (Both of these results are what would be expected for “Fast Sticky” mode.)
And now a very surprising result: I remapped all the inputs to mask the diagonals (UP = UP NOT RIGHT OR UP NOT LEFT, etc.) and repeated the test. THE RESULTS WERE IDENTICAL!
Conclusions: Eight-way joysticks should work fine for 4-way games in MAME, because the software is already filtering out diagonal inputs. Derrick’s IC should not be required for MAME (although other programs may benefit from it, or the Turbo Fire or Sticky Mode features may be useful for some games). Keyboard input assignments to exclude diagonals do not change the way MAME responds and are no longer required.
I appreciate all the other posts on this subject and encourage everyone here with an interest in this to try it out for yourself and respond.