32 and/or 64 bit?
-
- Posts: 93
- Joined: Mon Jan 09, 2017 3:22 pm
32 and/or 64 bit?
There is once the possibility to set 64 bit, but in a completely different menu then also (additionally?) 32 bit.
Which setting is recommended?
Which setting is recommended?
Re: 32 and/or 64 bit?
You mean "Playback > 64-Bit Processing" and "Audio Hardware > 32-bit Output" ?
The first one chooses the data type for DSP ( digital signal processing, that is equalizer, crossfeed, etc ) as 32-bit float (off) or 64-bit float (on). Reduces calculation errors and its associated noise. But dubious as to whether anybody would hear the difference.
The second one chooses the data type for transfer of the processed data to the DAC ( digital analog converter ) between integer (off) and 32-bit float (on). If there were a DAC that supported it then switching it on would be preferable because it would avert conversion from float to integer, but afaik there isn't.
The first one chooses the data type for DSP ( digital signal processing, that is equalizer, crossfeed, etc ) as 32-bit float (off) or 64-bit float (on). Reduces calculation errors and its associated noise. But dubious as to whether anybody would hear the difference.
The second one chooses the data type for transfer of the processed data to the DAC ( digital analog converter ) between integer (off) and 32-bit float (on). If there were a DAC that supported it then switching it on would be preferable because it would avert conversion from float to integer, but afaik there isn't.
-
- Posts: 93
- Joined: Mon Jan 09, 2017 3:22 pm
Re: 32 and/or 64 bit?
Thank you for your answer.
So, when my hardware supports it, 64 bit float for DSP and 32 bit float for DAC would be preferred?
So, when my hardware supports it, 64 bit float for DSP and 32 bit float for DAC would be preferred?
Re: 32 and/or 64 bit?
Yes. But you won't notice if your hardware doesn't support 32 bit float for DAC, so you better don't use that.CarlFredricksen wrote: ↑Tue Mar 01, 2022 2:22 amThank you for your answer.
So, when my hardware supports it, 64 bit float for DSP and 32 bit float for DAC would be preferred?
-
- Posts: 6
- Joined: Sat Apr 09, 2022 5:24 pm
Re: 32 and/or 64 bit?
Better not? Does it have any negative effects like increased cpu load? Don't the processors have 32bit float computation units, or what they're called?blaubär wrote: ↑Tue Mar 01, 2022 5:21 amYes. But you won't notice if your hardware doesn't support 32 bit float for DAC, so you better don't use that.CarlFredricksen wrote: ↑Tue Mar 01, 2022 2:22 amThank you for your answer.
So, when my hardware supports it, 64 bit float for DSP and 32 bit float for DAC would be preferred?
Re: 32 and/or 64 bit?
It's not the processor but the DAC. The point is, if it is not supported then it's not defined what will happen. Somewhere some transformation from double to int will take place, but then there's no way to tell what quality that will have.
-
- Posts: 6
- Joined: Sat Apr 09, 2022 5:24 pm
Re: 32 and/or 64 bit?
Ah, ok. That clears things up. Thank you
-
- Posts: 6
- Joined: Sat Apr 09, 2022 5:24 pm
Re: 32 and/or 64 bit?
I don't know what changed, but now I can't use 32bit int with the Generic Driver anymore, I have to use 32bit float (32bit Output option) instead for music to work. It worked fine yesterday. Weird.
Re: 32 and/or 64 bit?
I suggest you hit the defaults-button in "Settings > Audio Hardware".itGoesBlargh wrote: ↑Wed Apr 13, 2022 5:04 pmI don't know what changed, but now I can't use 32bit int with the Generic Driver anymore, I have to use 32bit float (32bit Output option) instead for music to work. It worked fine yesterday. Weird.
Who is online
Users browsing this forum: Google [Bot] and 34 guests