"Do you know what all those knobs do?"
If you stand near the mixing console at enough live sound events, you will hear that question sooner or later. It's usually asked with an undertone of, "Wow - that looks complicated." Now, what if you are at a public policy forum with your own Mixer on your laptop, but instead of controlling the sound, you are modeling the policy using parameters you choose? Say, more bike-lanes and fewer buses, or more buses and fewer bike lanes?
Operating a mixer is less complicated and more difficult than it looks. Less complicated than it looks because there's a lot of repetition of the controls. Once you figure out how to work channel one, you know how to work channels two and up. More difficult than it looks because mixing the input to each channel has its own requirements. A channel for the kick drum needs to be handled differently than a channel for the guitar amp.
This diary is a pre-Request For Comments (RFC) about expanding the concept of the audio mixer to cover concepts themselves: Preferences, policies, ideas, etc. Tax policy or search criteria, to give two examples, treated like the sound for a rock band or a play.
Call it - what? - a Meta-Mixer, IdeaMixer, ConceptMixer. Doesn't matter a lot, but a good name would be useful. It will be Meta-Mixer for now, but suggestions are welcome too.
Who owns the Meta-Mixer idea? Nobody - I think. I'm saying it first, and as far as I can tell it's original, so if you make a small fortune with your own use or variation of it, you'd be welcome to take me to dinner or something. But you wouldn't be obligated to do so. I'm pretty sure that with me publishing it openly, you can use my (John Abney's) broad idea and never even mention me. Or you, if some of your ideas are used.
How would a Meta-Mixer work? A relatively simple example (at least to use) might involve diary-reading preferences on a site such as dailykos. The "Input" would be a user's preferences. "Channel one" might be 'total comments.' Channel two could be 'unique commenters.' Channel three could be 'Trusted-User comments.' Channel four could be 'user-ID of commenters.' Channel five might represent the 'longevity of commenting.' Each channel would have a default setting of "0" which would be analogous to "0 dB" on an audio mixer.
Perhaps unintuitive at first glance, on an audio mixer 0 dB can be ripping-loud. But pushing the fader up can make it even louder (or more distorted). Pulling the fader down can make it softer. Complete attentuation - in theory - is available at the bottom of the fader's travel. This is often represented on a mixer's control surface with the symbol for infinity: ∞
So let's say Ursula User decides that she doesn't have the desire to get involved in diaries filled with, "Is," "Is not," "Is so" ping pong comments. She boosts Channel two: "Unique" (commenters) to +64 (all the way up) and leaves all the other faders at 0. The site's database then lists those diaries with only unique commenters at the top of her 'I may want to read this' list. She finds the results sparse at the top of her list so she drags the fader down to +10. She also wants a little bias toward more currently followed diaries, so she pulls the fader for Channel four: "Longevity" down to -10.
Depending on the database administrator's preferences, Ursula's moving of each fader might update in 'real' time or she might have to push a discrete 'Sort' or 'Run' button.
At 0 dB Digital Audio Mixers see each input signal as a series of varying numbers. If the user pushes a fader above 0 - say to +3 dB, it adds 3 dB worth of volume to each number flowing through the channel. The number it adds for +3 dB is unlikely to be 3, because the dB (which is short for decibel - named in the somewhat terse honor of Alexander Graham Bell) is logarithmic. Fortunately, the user does not need to do the math.
Whether a Meta-Mixer would use logarithmic representation or not would depend on several factors. One of the most important factors is the ease of adapting existing Digital Audio Mixers to handle other than audio signals. If it should be that logarithmic and non-logarithmic channels can be combined in the same Mix (that's what using a Mixer produces) then the Channel's Creator would choose whether a particular Channel is handled internally using logarithms or not.
Right now, I'm thinking of just a few levels of activity:
1. Meta-Mixer - the basic concept - not including the mix-engine itself. The Publication of this diary on dailykos and the creation of the Meta-Mixer group on dailykos should be considered as prior-art. The Meta-Mixer concept itself is not for sale.
2. Mix-Engine - There are a number of free and commercial Digital Audio Mixers available. They co-exist in the audio world, and should be able to co-exist in the Meta-Mixer world.
3. Channel - How should an intangible be represented numerically? How should (or should) a tariff rate be weighted? The Channel Creator decides those sorts of things. A Channel should be testable against one or more historical models to try to keep mischief at bay.
4. Output - If you raise the tariff on steel wool and lower the tax rate on paper-plate manufacturers, how will that be represented as an output? A color? A sound? A pie-chart? Like a Channel, an Output should be testable against one or more historical models.
5. Mix - A user may Build his or her own Mix or use a third-party Mix.
What isn't included yet is the database tie-in function for network use.
Think of the Meta-Mixer as a peoples version of Congress's (hopefully non-partisan) General Accounting Office.
The idea is now in the open. Interested?
best,
john
PS, Please forgive my over-simplification in the block-quoted sections. You are invited to make such corrections and additions as you deem necessary.