24-bit files are sold on the promise of more detail. What extra bits actually buy is dynamic range — the gap between the quietest sound a format can carry and the loudest. Set your listening level and how quiet your room is, and see how much of that range you can physically use.
Each bit of depth carries about 6 dB of dynamic range. Two conventions are in circulation: 6.02 dB per bit gives 96 dB for 16-bit and 144 dB for 24-bit, while the full-scale-sine figure adds 1.76 dB and gives 98 and 146. The calculator above uses the second, which is why it reads 98 rather than 96. Either way, noise-shaped dither puts 16-bit beyond 100 dB of usable range where the ear is most sensitive. And 32-bit float carries a 24-bit mantissa — the exponent buys headroom, not resolution — so its usable range is the same as 24-bit, which is why selecting it changes nothing here. Your usable range is the gap between your own noise floor and the loudest level you listen at — anything the format offers below the noise in your room is buried before it reaches you.
None of this makes hi-res files useless. Higher bit depth is genuinely valuable during recording and mixing, where headroom protects against clipping across dozens of processing stages. The question this tool answers is narrower: whether the finished file you play at home benefits. Sample rate is a separate argument again, and a more contested one — but it is not what the extra bits are selling.
If a hi-res release sounds better to you, that is usually worth taking seriously, because remasters are frequently mastered with less compression than the CD version. That is a different and often audible improvement — it is just not the bit depth doing the work.