Technical: What is summing?

[The Chase]

What do people refer to exactly when they mention "summing"?

for instance, when someone says in a comparison between 2 hosts:

they are certainly not the same in how thier mixer and summing functions work. In other words, there is going to be a difference when you start talking about more than one sample or track and its going to be fairly distinct from my experience. I recently mixed down a tune for a fairly big artist that was done in FL, we then mixed a version in Pro Tools and the difference was pretty apparent (although plug ins and things like that are a big part of it). I think that almost all the DAWS sound a bit different when you start comparing mixdowns of whole tracks. Even Pro Tools LE and HD sound different

What does the term mean?

[Dozius]

I'm guessing it has to do with the way the host sums the seperate audio waveforms into one waveform. The mathmatical sum of the two or more waveforms resulting in a single new waveform. When talking about waveforms and summing (in electricity) that's usually what it means. but I might be way off when it comes to this and DAW's.

[Muff Wiggler]

the way multiple individual audio tracks are 'summed' into a single stereo track, i.e. mixdown

some swear digital summing lacks the 'air', 'space', 'width' and 'clarity' of analog summing, and have 'tests' to prove so.

some swear there's no difference between digital summing, it's all the same math, 'summing arguments' are snake oil, and they have 'tests' to prove that too.

some even sell equipment to bus your individual tracks out of your DAW, sum them in an analog domain, and return the resultant stereo misdown back into your digital domain.

stuff like this:

http://www.soniccircus.com/store/mercha ... creen=PROD

the debate rages on

[Muff Wiggler]

oh yeah some sell CDs comparing the summing differences 'if any' of the same files mixed-down in different environments

like this:

http://www.mercenary.com/3dauawdawsum.html

[The Chase]

oppinions?

[james0tucson]

the way multiple individual audio tracks are 'summed' into a single stereo track, i.e. mixdown

some swear digital summing lacks the 'air', 'space', 'width' and 'clarity' of analog summing, and have 'tests' to prove so.

If analog and digital summing sound different, it's because one method or the other introduces aberration.
If the aberration is desirable, it's "colour".

Some digital artifacts are desirable, and some analog coloration is desirable, but it's far more likely for a summing bus in the digital domain to be entirely precise, than it is for an analog summing bus to be entirely without measurable distortion.

Digital mixing and digital summing are different problems, and it's possible to introduce errors in either.

Any argument that a particular analog summing bus sounds better than the garden-variety summing in any DAW, should consider the possibility that the analog tool is introducing desirable aberration, and not simply "more transparent." It could be that "too transparent" as in the case of most things in the digital domain, is a bad thing.

[Muff Wiggler]

it's also worth mentioning that the serious proponents of analog summing will (perhaps to better sell their wares?) tell you that if your analog summing is being done on commerical or pro-sumer sort of gear with generic opamps in it (behringer, mackie, etc) you are better off to stay with the summing in your daw. you need the $3K Neve if you wanna feel the love bro

you can buy a $20K a/d converter too, it all gets down to priorities, and where in the signal chain you want to get nitpicky about (or nitpicky about next)

[Kingston]

It's subject that should be deal with absolute avoidance at all costs. If you don't know what's being discussed and are unclear about the facts. Do yourself a favour and leave now. forget. summing discussion became irrelevant more than 5 years ago.

It's all snake oil and placebo now. Hence the sudden relevance and constantly reopening topic.

[The Chase]

would the summing of 2 different digital hosts vary that much? at all?

[M'Snah]

No. But in theory it can differ, Katz has a nice example in his book (and it's part of numerical math - or however it translates - in a.o. computer science studies)...

If you have 3 signals a, b and c, and you sum those, you can do either (a + b + c)/3, or you can do (a/3) + (b/3) + (c/3)... If there's a difference between these two numbers, it's because a or b or c is a very small or big number, almost at the limits of what can be represented with the number of bits.
e.g. if I use 8 bit integers, the maximum number is 255... If I would add (200 + 200 + 200)/3 the result is 88/3 = 28... If I would use 200/3 + 200/3 + 200/3 = 66 + 66 + 66 = 198.
That's theory. With 32 or 64 bit floating point numbers this doesn't happen.