@knownothing2:
I just don't know where to begin...
Let me tell you a few things that hopefully make you understand why what you are proposing is not a good idea.
Do you know the history of stereophonic sound? A long time ago someone discovered that when humans hear a sound from 2 sound sources (in front of you, but no more than 60 degrees apart) then it seems to them that the sound comes from somewhere inbetween those 2 sound sources. And if the 2 sound sources are not equally loud then it seemed the perceived location of the sound was closer to the louder sound source.
If you place 2 loudspeakers in front of you, one 30 degrees to the left of straight in front, the other 30 degrees to the right of straight in front, and you play a stereo recording then you will experience a sound stage at front of you, between the loudspeakers. (And due to wall reflections and secondary clues in the mix - that in part may be related to reflections in the recording space(s) - you may feel there is some extra width and depth around it).
Now note that of course it is an illusion that the sound seems to come from somewhere inbetween the loudspeakers. In reality the sound is coming from the two loudspeakers. By the way, this stereophonic illusion is not experienced by all humans, a very small percentage of humans do not.
There are different types of stereo recordings. Some only "pan" different instruments somewhere inbetween the speakers by means of the left-right level ratio. Some use additional effects. Some of the instruments may have been recorded with microphones in a room. Some instruments in one room, and others in other rooms. Some instruments may be electronic and directly "panned" into the mix. Some may have artificial reverb added. A "trillion" other special effects may have been applied to some of them, and not to others.
A recording of a live orchestra often will be made using many different microphones, in many different positions.
The stereophonic illusion is not the same as localizing a sound in real life.
What do you want to test?
-How that M.I.T. alogorithm localizes sounds in an average stereo music recording? By letting the algorithm analyse the stereo signal coming from the DAC directly?
Well, hopefully you will understand by now that that is a terrible idea, because an average stereo recording does not have the sound locations coded in a similar way as it is coded into the sounds going into your ears when listening to a natural sound in your environment, not even close!
Have you heard of Impulcifer (free software by jaakkopasanen), or of the Smyth Realiser A16 (or A8)?
Using any of these you can get a very realistic binaural simulation of loudspeakers over headphones, based on personal measurements.
First with in-ear-microphones test signals are recorded that are played over real loudspeakers in a room. Next test signals are recorded played by your headphones on your head.
The first measurement is to determine which changes sound undergoes on it's way from each loudspeaker to both your ears. Including reflections and reverberations of the room. And including the colorations by bending around your head and into your ears (head related transfer function), applied to all components of the sound entering your ears (direct and indirect).
The second measurement is to measure the frequency response of your headphones on your head.
At playback time all the same changes that happen on the way to your head can be applied to any input signal. Next the inverse eq of your headphones frequency response is applied to compensate for the influence of your headphones.
The result is that it seems as if you are listening to loudspeakers. In particular that the sound seems to come from a distance. When listening to stereo over 2 virtual loudspeakers you will experience a soundstage at a distance in front of you, for example 3 meters (if you measured loudspeakers 3 meters in fron of you).
In fact you are listening now to a combination of two "spatial audio rendering concepts". The first is the stereophonic illusion of sound from 2 loudspeakers in a room. And the second the binaural simulation of those loudspeakers in a room through headphones. The second part is in fact very closely related to how humans localize natural sounds in their natural environment.
Why am I telling you all this?
Because hopefully it can give you an idea of the scale on which things happen in the signal path that relate to soundstage. The difference between the input signals for such a binauralising system and the output signals that are being send to the headphones is the result of extensive digital processing and are massive and easily measurable. In fact the weakest link in the concept is the placement of the in-ear-microphones. A 1 mm shift in position of the microphones will give clearly measurable differences in results, even if the illusion of virtual speakers afterwards will still possibly be equally good! The differences are not at the scale of differences between DACs or amps let alone total audiophoolery like special power strips etc. The influence of these things, in comparison to a small shift of the microphone, is "0.0000", absolutely meaningless.
Another important thing to note is the role of the personal hrtf of the user. Using one person's measurement may sound totally strange and inconvincing to another person. And these differences also are many orders of magnitude above the differences between DACs and amps.
Soundstage has nothing to do with "magical" "unknown" "now immeasurable" properties of DACs and amps or something like that. Soundstage is a result of what recording you are playing and how you deliver it to your ears. And in some situations your personal hrtf will play a role in that as well. Ultimately, it are your brains that create the soundstage. And that is a complicated process. But it has nothing, nothing at all, to do with DACs and amps, unless they are unbelievably bad.