Does 2.5x gain mean 2.5 times louder than unity gain?
Yes.
Is having 12x better sound wise? I'm only going to use this as a summing device, not to boost anything beyond it's regular output levels).
Soundwise they would sound the same, the gain is just nice insurance, in case something is real quiet, like some piezo equipped instruments or mic or something. You could make R13 4.7K and have just a little more than unity gain if you don't want extra gain available.
Not sure I follow on the Vref thing. So there's basically a trace missing from pin1 of IC2 going a bunch of places?
the 4.5v Vref biases all the opamps to half the supply voltage, so the negative cycle on the audio doesn't get cut off by the amplifiers not being able to produce a negative (lower than 0v) voltage. Yes, it is supposed to go a bunch of places, everywhere on the schematic that has VREF in a diamond shaped border tied to it.
Lastly, is there some automated way to turn a schematic (specifically that one) into a perfboard layout?
No, but luckily elbow grease is the original automation.
I don't bother with a layout, just start by making the board the size it needs to be, drilling holes in it so it can be mounted later, then stick the ICs on in a roomy location with plenty of pads around them for the discrete components, then start from the inputs and work through the circuit, one connection at a time. You're making a little copper road system, it doesn't matter the shape of the road, as long as it ends up in the right place and makes the right stops. With a roomy board, there's a million ways to lay it out right. Some are more clever than others, but the only thing that matters is if it works.
One could keep track of connections made on the schematic with a highlighter, if inclined. I just kind of triple check everything as I go and continually review the little area I'm doing, checking the connections are right, polarity of components is right, no unintended shorts, that kind of thing.
