Streaming services now offer lossless and high-resolution tiers as a matter of course, and the marketing implies the file format is what stands between you and the recording. In most listening situations it is not even close to the limiting factor.
Walk the signal path. A well-made lossy stream at a modern bitrate is already difficult to distinguish from lossless in a blind comparison. Above that, the file is decoded, converted by a DAC, amplified, and turned into air movement by a transducer — and then it interacts with a room, or with the seal of an eartip. Every one of those stages introduces more deviation than the difference between 16-bit and 24-bit ever will.
Two things genuinely do change what you hear. The first is the master: a hi-res release is often a different, less compressed master than the CD-era one, and that is a real audible difference wrongly credited to the sample rate. The second is the transducer and its fit — the largest single variable in personal audio, by a wide margin, and the cheapest to improve.
None of which argues against lossless. Storage and bandwidth are cheap, and an archival copy that is bit-exact is worth having on principle. It argues against spending money in the wrong order. Fix the transducer, the fit and the room first; the file format is the last few percent, not the first.