I suspect there may be little practical benefit to exporting audio at a higher bit depth than 24-bit, but this is unfamiliar territory for me, so I experimented with Dorico’s 64-bit float export option.
I think, although I may simply be hearing what I expect to hear, that the 64-bit export sounds slightly more refined. However, when I import the resulting WAV into Logic, Logic reports it as 32-bit in the Project window. I get the same result when checking the file in Fission: the WAV exported from Dorico using the 64-bit float option is identified as 32-bit float.
This leaves me wondering:
Can Dorico actually export audio at 64-bit float, given that the option is provided?
If so, why are Logic and Fission identifying the resulting file as 32-bit float? (Perhaps that’s a question for the respective forums.)
Is there any practical advantage to exporting from Dorico at 64-bit float rather than 24-bit or 32-bit float?
I assume there must be a reason for providing the 64-bit option, but I’m not familiar enough with the technical side of digital audio to understand what benefit, if any, it provides in practice.
I’d be interested to hear from anyone who can shed some light on this.
None of the samples that you’re using have anywhere near that precision, so it’s like describing integers to 10 decimal places. I doubt there’s any meaningful difference, unless it’s an error.
It’s MASSIVELY greater precision than your ears can physically hear, and only of any use if you’re planning to process the music in ways that work with that level of precision.
Why is it there? Well, probably because it’s easy for the team to include it, using the third-party library that offers all the rendering options. And because audiophiles like bigger numbers.
Thanks, that’s really helpful, particularly the point about Logic only supporting 32-bit audio. That would certainly explain why Logic is reporting the file as 32-bit.
I think I was probably approaching the question from the wrong end of the problem. I wasn’t really expecting 64-bit to produce some magically higher-resolution sound; I was more curious about why Dorico offers the option and whether there is a particular stage in the rendering or subsequent processing where it might actually be useful.
Your point about the samples themselves is well taken. I hadn’t really considered the issue in terms of the precision of the source material versus the precision of the rendered file. And I certainly take the point that any audible difference I’m perceiving is much more likely to be expectation (or simply a difference in level?) than some meaningful improvement resulting from 64-bit precision.
So, thank you, that’s clarified quite a lot for me.
I suppose the remaining curiosity is simply why Dorico exposes 64-bit as an option, but perhaps the answer really is as mundane as you suggest: the rendering library supports it, so there’s little reason not to expose it. @Ulf may have a view to share…
Thanks again for taking the time to explain everything - much appreciated, as ever.
The reason is: Because we can. Dorico’s audio engine is directly derived from Cubase, and Cubase can do, so we offer that in Dorico as well. Practically it makes no sense, because Dorico only works with VSTInstruments and sample libraries that don’t have such high precision, but if someone thinks it sounds better with 64 bit…well, here you are
