Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[John Roberts {JR}]

It does make sense.

Your first post made me think that you looked at the code I posted over on the Reaper forums and noticed something that could be improved.

JSFX is just Reaper's C-style scripting language. It is, of course, higher level than machine code but lower than what I'm used to working in. My only experience with machine code was in school but I understand what you describe.

Thanks for taking the time to respond. Your insight into the lack of need for a highly accurate RMS reading is valuable.

OK I went back and I think I follow what you are doing (a little).

I may need to clear something up, the "sum of all" level that gets compared to each channel to apportion gain, is not a mathematical sum of the different channel levels but the dry channel signals summed into their own "sum of all" bus and then the level of that bus signal is what gets compared to each channel.

I have seen AM applications that use channel meter data to compute a sum of all, and it can work after a fashion, but will not accurately discriminate between summed coherent signals (+6 dB) and summed incoherent signals (only +3dB). It's probably a fair ASSumption that all your signals are incoherent  (you probably won't have two mics picking up the same talker).   

have fun...

JR

[Corey Scogin]

I may need to clear something up, the "sum of all" level that gets compared to each channel to apportion gain, is not a mathematical sum of the different channel levels but the dry channel signals summed into their own "sum of all" bus and then the level of that bus signal is what gets compared to each channel.

I have seen AM applications that use channel meter data to compute a sum of all, and it can work after a fashion, but will not accurately discriminate between summed coherent signals (+6 dB) and summed incoherent signals (only +3dB). It's probably a fair ASSumption that all your signals are incoherent  (you probably won't have two mics picking up the same talker).   

You are correct. I'm taking the RMS value of each signal across the set window and adding those values up to create the "sum of all".  It seems to work reasonably well now but I don't have another AutoMixer to compare it with.

Thanks for bringing that to my attention. I'll look into summing every input channel signal (or sample rather) first to see how that works out.

That brings up a question I hadn't considered before. I've just assumed so far that calculating the RMS value of the signals across a time window before comparing was the way to go. Is an AutoMixer designed to function instantaneously? Could I do the comparison and adjust every sample?

Thanks again for the insight. I definitely have some things to think through and try out. You are much more knowledgeable than I'll ever be in this area and your advice is greatly appreciated.

[John Roberts {JR}]

You are correct. I'm taking the RMS value of each signal across the set window and adding those values up to create the "sum of all".  It seems to work reasonably well now but I don't have another AutoMixer to compare it with.

Proper AM that use the "Dugan" algorithm work like I described.

Using samples of level data depends on the time between samples to prevent choppiness. You can smooth the rate of gain change/fader moves, I guess. 

Thanks for bringing that to my attention. I'll look into summing every input channel signal (or sample rather) first to see how that works out.

Unless you are summing the actual bipolar audio signals, you will not parse out phase information, or coherence between sources.

That brings up a question I hadn't considered before. I've just assumed so far that calculating the RMS value of the signals across a time window before comparing was the way to go. Is an AutoMixer designed to function instantaneously? Could I do the comparison and adjust every sample?

AM technology is quite mature and indeed analog AM work in real time. There are attack and release time constants associated each stem to keep the gain changes smooth and transparent. Inside my analog AM designs I actually performed a log conversion, after rectification and smoothing to get voltages in the dB realm to compare and apply to VCA control ports.

Thanks again for the insight. I definitely have some things to think through and try out. You are much more knowledgeable than I'll ever be in this area and your advice is greatly appreciated.

Old 80-20 rule... you can probably get most of the AM benefit working just from levels, as long as the samples are frequent enough and gain changes smooth. Not being 100% mimic of Dugan, should not be expected.

To get the full Dugan, perhaps dedicate a pre-fader aux send for the Sum of All bus and just sample the level on it to compare to the channels. That would properly handle coherent and incoherent sources wrt gain sharing, while this is not a huge deal in practice, it kind of separates the real deal from also ran.

JR

   

[David Sturzenbecher]

The Waves Dugan plugin is like $500+ and I think it is meant pretty much just for live use. I've never heard of anyone using it in  Protools or other DAW session.

Mixing by hand, with a control surface, using the write automation function and a couple of passes through the session would probably be easiest.

This plugin is currently on sale for $300

http://audiodeluxe.com/products/waves-dugan-automixer

[Justice C. Bigler]

This plugin is currently on sale for $300

http://audiodeluxe.com/products/waves-dugan-automixer

I just came here to post this. Looks like it will be on sale through the month of March. Have to see if I can scrape together the $300 for it.


There's too many good deals on plugins these days and they are keeping me from buying a new smoker.   

[John Roberts {JR}]

I just came here to post this. Looks like it will be on sale through the month of March. Have to see if I can scrape together the $300 for it.


There's too many good deals on plugins these days and they are keeping me from buying a new smoker.   

Buy the smoker, it won't be obsoleted by free-ware.

JR

[Justice C. Bigler]

Buy the smoker, it won't be obsoleted by free-ware.

JR

It's easier to sneak plugins into the house than it is a 5 foot long 600lb smoker.    

[Timo_Liski]

And then a few years later, finally there is a Automixer plug-in the can be inserted as a VST3, AU and AAX plug-in:
https://www.wtautomixer.com by Wavemark ltd, where I work.

[Mac Kerr]

And then a few years later, finally there is a Automixer plug-in the can be inserted as a VST3, AU and AAX plug-in:
https://www.wtautomixer.com

Timo, please go to your profile and change your name to your full name as required by the posting rules. You also need to indicate your connection to this product.

Mac

[ProSoundWeb Community]