Hi Aharon, I certainly agree that at the transcription stage the style of the edition should be followed. I thought you were talking about a generic version of Sefer Tehillim itself.
In general, though, there needs to be some systematic thought about numerous parallel editions of the same text. When you dozens of editions that are extremely similar but not identical in terms of text and style, what is it that we actually want to do? Transcribe many dozens of siddurim (same for Tanakh and other books) individually? Create one giant coded text that incorporates the data from all of them in all of their myriad variations?
I personally lean towards something in the middle: Creation of some well-edited, documented and formatted basic text-types that stand on their own (e.g. Siddur Nosah Ashkenaz or Tanakh with Teamim) and document within themselves a limited amount of small variations. These can then be utilized in different directions: As the basis for working on the transcription of a very specific and important edition, or as independent components within a larger mega-text that can show and deal with major variations.