I think the output is the diagram labeled 2 just to the right of the of the photodiode on the right. It's the small rectangle that has a connection above the right diode and ground. Off to the left of it in charachters I cannot recreate here well are o10 with a diagonal line through the larger 0. That would make sense to me as it's the only place in the diagram that is open between a potentially variable voltage/current point and ground. I suspect you will measure output in milivolts. Changing the R2 3k resistor to a 10k pot would give you a degree of "calibration" or gain controll I think too, which may be useful if you were going to use the output for something besides measurement (I.e. The recording device you mentioned)
Disclaimer- my knowledge of schematics is amatuer at best and my previous experience is several years of building DIY audio stuff. I'm not too savvy with electronics but after years of building audio stuff from schematics and trying to learn what's actually going on with them in the process I think I have a pretty decent foundation in understanding simpler stuff like this. It's a very simple and cheap circuit to build and worse case scenario is you try it and waste a couple dollars in the process.
Which photo diode is used to measure the optical reflection? I'm not sure what you are exactly trying to accomplish with this circuit, can you give a more detailed description of your application? Are you trying to measure frequency response? You mentioned "mixed up in the laser diode cavity" which really throws me off. I don't see how that is going to happen. What I envision you doing with something like this is pointing your laser onto the moving object (audio driver) which will cause the reflected light to move around onto the field it is reflected to. I presume you would try to put one (or both) photodiodes onto this reflected field to gather some of the reflected laser light. I can see where that might give some variable output that would be proportional to the movement of the audio driver.
It looks like a novel idea but I cannot see how this would provide any useful information on frequency response of a driver that could be applied to designing a loudspeaker (which would be my interest in something like this) because it would not take into consideration the effects of the baffle, interaction between drivers, and the room. It would only give the response curve of the moving driver surface which does not equate to ultimate audio response of the driver (which is a system that includes baffle, room, etc).
Can you provide a link to the paper where you found this schema?
I will be very interested to see what you do with this and how it works. I can see how it might be used in something like one of those laser listening spy devices. Shooting a laser onto a window and measuring the vibrations and converting them back into sound.
Keep us updated on your progress please.
