I use them both. I take crude spice and re-dissolve it in acidic water, then re-freebase it and re-crystallize it.
Recrystallizing only aims to change the form of what you have dissolved in the non-polar solution. It can also minorly clean up an extract. For example if I have dirty powder spice, and recrystallize, I might get the same powder form, or crystals, depending how the crystallization is carried out. But it'll still have the same contaminants, unless those contams fail to re-dissolve in the non-polar solvent. It can also dilute the contaminants - like if you had enough contaminant in solution to freeze precipitate some of that contamination out with the DMT, recrystallizing would further dilute the concentration of the contamination, so maybe there wouldn't be enough for it to crash out on the second recrystallization. Then recrystallization would effectively separate that contamination from the DMT when you decant the contaminated solvent away from the re-crystallized DMT.
Adding a mini-AB step takes it one step further. In theory, if you have non-polar contaminants, they won't be able to follow the DMT into the acidic, polar, aqueous solution. And you can wash the acidic solution with non-polar solvent to further strip away any residual non-polar compounds that haven't back-salted into aqueous solution. Or for example if you had some lye or alkaline contamination in your crude spice, this would neutralize that and give you a second chance to pull the DMT cleanly, without the rest of the bark soup causing potential problems. The acid phase is also said to reverse polymerization, but idk.
But without advanced testing equipment, it's hard to say exactly why you're doing it or what you're potentially eliminating. DMT is generally 95% pure from MHRB off the initial pull, and some of the more common contaminants like NMT or DMT-N-Oxide aren't necessarily removed with these steps. But they wouldn't noticeably affect the experience anyway.
Maybe someone with more understanding will chime in.
