Well first you have the source. We’ll call them “lossless masters”.
So now let’s assume the DSD and PCM (and PCM can refer to MP3, M4A, AAC, Opus, FLAC, ALAC, raw LPCM, etc…) that the artist or studio releases are both directly converted from the master. Why would you grab the DSD and then convert it down to PCM when you can acquire the PCM version more directly? More conversion steps can only harm quality. I can only see a point if you don’t trust the competence of the artist or studio when they made their MP3/M4As or something.
I can see a point if you had a huge library of DSD music and there is no other formats offered by the author/studio or you simply don’t want to re-buy the track in it’s PCM form, and you unfortunately have no way of handling DSD. That case meets the “technological limitations” I mentioned.
So the sense behind down-converting falls down to availability and saving money.
There’s zero sense whatsoever in up-converting though, you gain nothing by doing it. I’ve seen people who think they’re doing a good job converting MP3s to FLACs and I’m just like “Bro you’re just making it take-up way more space than it needs to, this is pointless”.
The vast majority of music is only really available in MP3-320 or M4A-256 (I personally prefer M4A). Those are by no means bad quality (in fact in blind testing the vast majority of people cannot discern an MP3-320 from a lossless), but I can imagine it being rather sort-of disappointing to audio enthusiasts. I swear that FLAC/ALAC sounds better in some way but others would disagree, so maybe it’s placebo.
If I could, I’d probaby build a library of nothing but lossless PCM and DSD files, but that’s just unrealistic given the way music is actually distributed for common consumer playback, and I can understand the advantages of sacrificing a miniscule amount of quality to make the filesize a tenth of it’s original size. (FLAC compression is actually quite awesome middleground – compression without loss, basically the audio version of PNG image compression, the only thing sacrificed being the time it takes to process the compression which is inconsequential. “Compressed Lossless” is a thing and there’s no sonic reason to prefer the ALAC approach which is both lossless AND uncompressed and therefore massive).
I’ll also say that out of lossy formats, I kinda hope that Opus really takes off. High-quality Opus is better than MP3/M4A but not much larger (if at all, depending on the source). It’s just newer and hasn’t really taken off yet, I suppose the launching of the AV1 format (royalty-free competitor for H.265) may change that though as it primarily uses Opus for audio.
So if it were up to me… I’d probably have a library of compressed FLAC for lossless PCM, High-Quality Opus for lossy, and DSD of whatever happens to be available in DSD.
But, things just aren’t made available that way.
Agree on MQA being music DRM – it’s an excuse to lock down the platform. I only want the MQA version of the D90 because I want a do-it-all machine if I’m paying as much as I am for the DAC.