I know SFP is 32 bit integer.
But I also know that there are SDK modules that are related to floating point. I also know that the guitar emulators in SFP use 32bit floating somewhere in their paths.
So has anyone developed a 32 bit floating mixer ?
Regards,
Dimitrios
Is there any 32 bit floating point mixer developed in SFP ?
-
chriskorff
- Posts: 371
- Joined: Sun Nov 26, 2006 4:09 am
Re: Is there any 32 bit floating point mixer developed in SFP ?
Apparently not!
Though I would also be interested in any responses from developers as to the feasibility of a floating-point mixer...
Personally I like using the fixed-point ones supplied, because (a) you know where you are with them - if something clips, it's for a good reason, and you need to do something about it. Also (b) they encourage good mixing practice. It's much too easy, in native DAWs, to get into the habit of additive mixing, and then having the master fader right near the bottom of its travel to compensate for all the channels that are peaking in +dBFS values. "Something too low in the mix? Just turn it up further!" IMO fixed-point mixers, where individual channels can clip, have no real disadvantage - unless you consider the need to have a good grasp of gain staging a disadvantage!
Additionally, even if you're using a native DAW with a floating-point mixing engine, there's still the chance that a plug-in clips internally (this is common with DSP-processed plug-ins, I think, but not exclusive to them). If that's the case, you might have to re-visit your entire mix before you discover the reason for the clipping, as it may not show up in your channel meters (eg. if it's an insert plug-in that's clipping, but you have the fader down low enough that the channel itself isn't exceeding 0dBFS). For this reason, I think it shouldn't make a blinding bit of difference whether you use fixed- or floating-point mixers, *IF* you bear good gain-stage practice in mind (a bit like you might need to if you were using an analogue mixer
)
Nonetheless, if anyone has a good reason for why one doesn't exist, or would be difficult/impossible to implement in Scope, I'd like to hear it.
Cheers!
Chris
Though I would also be interested in any responses from developers as to the feasibility of a floating-point mixer...
Personally I like using the fixed-point ones supplied, because (a) you know where you are with them - if something clips, it's for a good reason, and you need to do something about it. Also (b) they encourage good mixing practice. It's much too easy, in native DAWs, to get into the habit of additive mixing, and then having the master fader right near the bottom of its travel to compensate for all the channels that are peaking in +dBFS values. "Something too low in the mix? Just turn it up further!" IMO fixed-point mixers, where individual channels can clip, have no real disadvantage - unless you consider the need to have a good grasp of gain staging a disadvantage!
Additionally, even if you're using a native DAW with a floating-point mixing engine, there's still the chance that a plug-in clips internally (this is common with DSP-processed plug-ins, I think, but not exclusive to them). If that's the case, you might have to re-visit your entire mix before you discover the reason for the clipping, as it may not show up in your channel meters (eg. if it's an insert plug-in that's clipping, but you have the fader down low enough that the channel itself isn't exceeding 0dBFS). For this reason, I think it shouldn't make a blinding bit of difference whether you use fixed- or floating-point mixers, *IF* you bear good gain-stage practice in mind (a bit like you might need to if you were using an analogue mixer
)Nonetheless, if anyone has a good reason for why one doesn't exist, or would be difficult/impossible to implement in Scope, I'd like to hear it.
Cheers!
Chris