First post, by ElBrunzy
My understanding is that in version 1.2 of DOOM 1 they fall back on software mixing on the GUS for the DMX music. Back then I had a slow 386sx16 with a GUS so, I was happy to benefit from the it's hardware mixing until I upgraded my doom version. Unfortunately I encountered a bigger problem with my board incompatibility with NMI who would make doom run very slow in some version, it was barely playable. So I was wondering, does anybody experienced the passage to software mixing when upgrading DOOM version on a compatible motherboard ? For instance, having to reduce a bit screen size, or changing to low quality level.
I was curious about that software mixing, so I grabbed a copy of doom 1.1 and 1.9 to compare. Now my retro computer is too fast to notice any performance penalty, but I was focused about sound quality. I have to admit I could not notice any very significant and obvious difference other than a louder volume in software 1.9 and instrument change that are probably introduced in the 1.4/1.5 instrument mapping improvement. For instance, you have bells in the beginning of level 18 music on 1.9 and they seem absent on 1.1 (using parameter -GUS or -GUS1 on 1.9 changed nothing). I also noticed that the maximum number of FX voices passed from 3 to 8 in the setup. Does anyone have information on the software rendering specification of the DMX library ? Rendering midi in software can be quite CPU taxing and I think we all agree that doom was pushing our computer to the limit so I'm quite curious about how they managed the transition from 14 voices at 44.1khz 16bit stereo to software rendering in such a smooth way that it seem many people did not noticed it.