| Originally Posted by u b k Okay, who knows the answer to a question that's bugged me forever: When a plugin does its thing, does it pass the result to the next process as a 32-bit FP word, or does it produce a 24-bit fixed word, which is then converted back to float for the next plug/gain/sum? Can anyone recommend a good primer on floating point arithmetic? Not too involved, but with some actual substance? I figure I'll be getting into the plugin game soon enough and would like to be able to converse intelligently with my programmers. Thanks! Gregory Scott - ubk . |
This depends entirely on your platform and configuration..
If you are running entirely host (PT LE, VST, AU etc.) the plug will output floating point to the next process at whatever precision the host uses.
If you use PT TDM, an RTAS plug-in will pass floating point to the next RTAS plug-in - but it will be changed to 24bit fixed if the next plug-in is TDM, because the TDM buss is 24bit fixed point.
If you use other internal expansion processing, depending on the external processing type, the signal may get converted to fixed point 24bit to drive their 3rd party system.
IF you break out of the the Floating point host to any digital outboard gear it will be converted to 24bit fixed point to drive the cable interface formats - and may or may not get processed in fixed point within the external box (depending on the system) - and be fed back into your system as fixed point 24bits because of the interface.
Basically it's a bit of a zoo..
