Reply to thread

Message

<blockquote data-quote="John Roberts" data-source="post: 68223" data-attributes="member: 126">Re: How to do an analog insert?I don&#039;t know what an osc meter parameter is. Latency is not a huge issue since gain moves are generally smoothed, but the first MSecs of a new talker could be chopped off if latency is significant, clicks if you try to open up the channel too fast after that lag. This could be mitigated by limiting the channel attenuation to only as much as needed for effective feedback control instead of full cut reflecting actual input levels. If the channel is only dimmed a few dB, start up is only a little soft, not completely trashed. Again the general flaw with working from already rectified signal levels is lack of coherency information. AM math requires a dummy sum of the raw audio signals to use to compare to level of individual stems for gain apportionment. Pretty simple, but a powerful technique when many open mics are involved. Built into a digital mixer, the AM gain could be layered on top of normal fader moves. Sorry for the veer... not a huge application for musical sound reinforcement.&nbsp; If you can pull actual channel audio into a remote PC and sum them there (or set up a dummy bus in the X32). You could then perform the Dugan gain sharing based on that AM sum. You could calculate fader moves to send back. Dedicating a dummy bus for the AM sum inside the X32 corrects for the coherent-incoherent issue, leaving only the attack time as a potential execution issue working from meter levels. Drilling down into the raw data eliminates even that compromise.&nbsp; [/veer]JR</blockquote>

[QUOTE="John Roberts, post: 68223, member: 126"] Re: How to do an analog insert? I don't know what an osc meter parameter is. Latency is not a huge issue since gain moves are generally smoothed, but the first MSecs of a new talker could be chopped off if latency is significant, clicks if you try to open up the channel too fast after that lag. This could be mitigated by limiting the channel attenuation to only as much as needed for effective feedback control instead of full cut reflecting actual input levels. If the channel is only dimmed a few dB, start up is only a little soft, not completely trashed. Again the general flaw with working from already rectified signal levels is lack of coherency information. AM math requires a dummy sum of the raw audio signals to use to compare to level of individual stems for gain apportionment. Pretty simple, but a powerful technique when many open mics are involved. Built into a digital mixer, the AM gain could be layered on top of normal fader moves. Sorry for the veer... not a huge application for musical sound reinforcement. If you can pull actual channel audio into a remote PC and sum them there (or set up a dummy bus in the X32). You could then perform the Dugan gain sharing based on that AM sum. You could calculate fader moves to send back. Dedicating a dummy bus for the AM sum inside the X32 corrects for the coherent-incoherent issue, leaving only the attack time as a potential execution issue working from meter levels. Drilling down into the raw data eliminates even that compromise. [/veer] JR [/QUOTE]

Verification