I still don´t get the benefits of the 32-bit floating point. I´ve read that the project won´t clip, but I´ve tried it and it does clip. I have recorded my project in 24-bit, 44.1kHz wav, so all the audio files are in that format.
So are there any benefits to switching my project now to 32-bit and taking the audio mixdowns in 32-bit, 44.1kHz and bring the audio files to mastering engineer in 32-bit?
Where did you clip? and how did you do it?
I still don´t get it, but kinda do… 
Found this from the second thread:
Cubase customer support:
“We would recommend using 32 bit floating point for the file types in the project so that you take maximum advantage of Cubase’s processing and mixing capabilities. This allows the any changes in the gain, plugins, mixing, panning, etc to be processed in their highest bit depth fidelity. I hope this information helps.”
So…
GOOD / BAD ? Benefits?
GOOD / BAD ? Benefits?
If I change the project now at this point (middle of mixing in 24-bit) to 32-bit, will all the fx plugins (EQs, reverbs, compressors…) process the audio better in realtime and can anyone hear the difference?
I wish I would have recorded and set the project to 32-bit from the start. Then I wouldn´t have to think about this and ask these questions. 
"We would recommend using 32 bit floating point for the file types in the project so that you take maximum advantage of Cubase’s processing and mixing capabilities. This allows the any changes in the gain, plugins, mixing, panning, etc to be processed in their highest bit depth fidelity.
This is nonsense. Cubase’s internal calculations take place in 32 bit float, which is very good, but a 24-bit resolution in your source files is more than enough. As stated elsewhere, this represents a 144 dB dynamic range.
No need and totally superfluous. Leave as is.
Yes, here I can see benefits, since this way you won’t need to use dithering.
Remember that Cubase may save intermediate files after applying some functions or effects, it’s not only about source files. If the intermediate files are in 32 bit FP format they will be of the same quality as if they were processed internally, otherwise you’ll lose some integrity.
So, using 32 bit floating point files makes all sense.
While this may be true in the strictest technical sense, the fact is that 99.99% of the time you will NEVER hear a difference using 32 float files, as 24 bit has more than enough dynamic range to work with. It is only if you intentionally force errors and do crazy things that you would be able to hear a difference.
There is a penalty for using 32 float files as well - your files are larger. I stick with 24 bit and the smaller files sizes as I track quite a bit and use a good deal of disk space.
If you are sending your files off for mastering then the recognised format would be 24 bit at the sample rate of your project.
Of course if you ask whoever is getting the files if they would prefer 32 float then that’s fine too
There might be a tiny benefit to recording to 32-bit if you’re tracking through software plugs, which operate in 32-bit. Whether most people will actually hear any difference is something else. 
Internal processing (obviously) happens at 32-bit float so bouncing instruments that way won’t hurt… but again, I doubt if anyone could hear the difference. The other advantage (obviously) to bouncing stems or tracks to 32-bit float is not having to worry about clipping the file export, which will happen with 24-bit files if they’re exceeding 24-bit 0 for some reason.
I mean, you can pretty much ignore 24-bit zero on individual tracks but if you happen to stem something like that to 24-bit it will clip in the file.
Only times you’ll have any real advantage of using 32-bit files are: