You boil them in acidic solution and strain the water. OK.
You then filter the water. OK.
At this point, you should have harmala acetates in solution, no?
Then you basify, which precipitates harmalas as freebase. OK.
Then you redissolve in hot acidified solution... so that redissolves harmalas in water, but doesn't it also redissolve anything that needed to be basified to precipitate in what otherwise seemed to be a clear acidified solution with nothing falling out of it for days?

Could one not proceed straight to crystallizing harmalas with excess salt?

But of course you can, it'll work just as well. You'll need to redo the salt crystallizations a few times before you end up with shining yellow crystal shards. The initial crystallization usually yields a brownish mud.

In the initial extraction I always go straight to salt. Its easier that way, in my experience.
You can make a perfectly usable product by only doing a series of salt precips. If you want a pretty pharmaceutical-grade-looking product with which to impress friends and seduce women, then you can so a basing step on your redissolved and filtered mansked harmala alks.
Manske gets accessory alkaloids and most plant gunk out. Basing can help get the rest of the plant gunk out.

One tip: I've found that in the initial manske, after you salt the alkaloids out of the seed decoction if you reheat the liquid, give it a good stirring, and then cool it again and keep it cold overnight you may isolate a little bit more harmala hydrochlorides. Thats often yielded several extra doses for me. I think in the first crystallization some ultra-fine crystals get missed and pass through into the filtered liquid. Also, this lets you get away with decantation instead of filtration if your seed decoction liquid is too thick with plant gunk to conveniently filter.

its critical to remove impurities.