First, there are a series of caveats to contend with. Both tracks need to be aligned exactly to make this work, which usually requires extreme zooming and moving of the audio to get right. Secondly, audio formats like MP3 don't work well with this process (after all, they have already had some audio removed and that amount will vary between the two different versions).
Finally, tracks mastered to tape or taken from vinyl (or any other analogue medium) will exhibit small and varying timing discrepancies that make everything harder to line-up correctly for the duration of the track.
Step 1: Import the original and instrumental track versions to your DAW. You don't need to know the tempos of each but you will need to align them correctly before you start. Start by fractionally cutting each before the first down beat. The zoom tool is your friend here.
Step 2: Once aligned perfectly, invert the polarity of the instrumental version only. This should at least reduce the level of any non-vocal elements.You may need to experiment with the relative levels of both versions to achieve the greatest amount of reduction.
Step 3: Assuming all that went to plan, you will now have an acapella ready to export to a new file. As this is a true stereo track it can be manipulated using Mid Side processing or filtering techniques to get at an even narrower range of vocal elements.
Step 4: For example, through our friend MSED we can engage the Side Mute button. This will reduce ambient treatments such as reverb and delay, and bring the lead vocal to the front of the mix ready for you to apply your own processing.