Analog summing vs. Digital Summing

F

fenix

New member
Analong summing vs. Digital summing is a constant heated debate on these boards. I always see people claiming analog summing is better with no hard evidence to support their claims. There have been great commercial albums mixed in the box just as there have been great commercial albums mixed with an analog board.

I did an experiment with a good friend of mine who owns a studio. I mixed both mixes and tried to get them as sonically similar as possible. All processing is done in the box on both mixes, but one mix is summed in the box and one mix is summed on an analog board. Can you tell which is which? Hopefully this thread will put an end to the analog summing vs. digital summing debate or at least settle on the fact that one is not necessarily better than the other.

Both files are encoded at 192.

Please, no guessing. Use your ears.

http://www.nowhereradio.com/artists/album.php?aid=1410&alid=-1
 
That is a stupid debate. Digital summing can be done without including any new distortion or any such problems. Analog summing can not. However, analog summing can be done well enough for any artifacts to be inaudiable, which for all practical purposes are the same as being done perfectly.

Therefore, since both digital and analog summing can be made perfectly, neither is better than the other. If mixing through an analog board sounds better, it is NOT because of the summing.

With these example files, there are no significant difference to tell which is mixed analog and which mixed digital.
 
There is a clear difference in sound between the 2 tracks... I'd say the first one is mixed analog, the second digital....

And the first one sounds better to me than the second.........
 
Blue Bear Sound said:
There is a clear difference in sound between the 2 tracks... I'd say the first one is mixed analog, the second digital....

And the first one sounds better to me than the second.........
Click to expand...

Is it fair to take it from this that you consider analog summing to sound better than digital ? I don't have speakers on my internet pc, so I can't listen. But I have done some blind summing tests using CEP 2.0 and Samplitude 6. Samplitude sounded alot better to me every time.

I personally believe you need a pretty high end console for analog summing to sound better than digital done properly.

Nathan
 
WTF? said:
Is it fair to take it from this that you consider analog summing to sound better than digital ?
Click to expand...
Not necessarily.... I think whichever gives you the sound you want to hear and you're most comfortable working with is what you should use.

And I don't think it's quite so cut 'n dry about needing a $250,000 analog console to outperform digital summing. That's like generalizing that a Neumann U67 will always be a better choice than an SM57...... and of course, that isn't the case at all..............
 
First - With all due respect to regbro, the difference in these files is huge. Enough that I believe the results can not be attributed to "analog or digital summing" alone - These almost sound like two entirely different mixes in their "core." I would assume that this is basically just the "imperfections" of the analog path vs. the repeatable numerical perfection of digital. If the test were done with a Dangerous Two Buss, with it's limited and unobstructed signal path (I've done that test, too) the results would probably be less overwhelming. Even when phase/nulled, the results were FAR more than "null."

Second - I agree with the Bear (as I find myself doing often) - Use what you like. If "chocolate" was suddenly decided the "best" flavor in the world, everything would taste like a Hershey bar.

On to the meat of things...

"X" has a center image with far more warmth and body than "Y"

"Y" has a spacious airy quality that "X" can't touch. And a panning issue...

If I were sent both of these files with a "master the one that you think would come out better" (which I get pretty frequently) I would want to work with "X" -

That being said, with a few quick tweaks, all "in the box," I got "Y" sounding almost exactly like "X" AND the other way around without much difficulty at all.

So the answer, of subpart A of subsection C, paragraph two, of the question, is:

4?

By the way - Nice tune.

John Scrip - www.massivemastering.com
 
WTF? said:
Is it fair to take it from this that you consider analog summing to sound better than digital ?
Click to expand...

No, it's fair to take it that he considers the sound of a analog mixer to sound better than that of a digital mixer.

A mixer is much more than just the summing.


And I didn't say that the mixes sounded the same. Just that there was no artifacts in either of the mixes to attribute either of them to either digital or analog. The mixes probably sound different for many reasons. One thing could be the simple fact that they are different mixes. :) A test like this will hardly reproduce the mixes exactly in both cases.
 
I don't know which is analogue but I prefer the first track better.
Althougth the second track seems to be a more intimate mix, the dynamics seem a little less controlled.

Which is which?

I have often tried mixing through a console as opposed to in the box, but although different, I could not say one was better. It was easier for me to mix in the box.
 
Hey fenix - I feel like y has a noticeable loss of headroom & depth and I'm guessing (not supposed to - I know) is most likely the digital summed version. It seems like a big difference to me.

The primary focal point is the quality of the tracking & recording, without a version x or y to compare I think either one would stand. I'd take x though now that I've heard it.

Cool shootout :cool:
kylen
 
oops - I played this for the wife (she's a vocalist) and she had the opposite reaction to me.

She felt that the vocals were tighter and were more in the front of the mix in version y. She said things to the effect that version x had too much space and the acoustic guitar was too loud and it lacked cohesion.

Did I mention she's a vocalist - he he so she was harmonizing the entire time and digging it. I guess that's what a musical presentation is supposed to do - move you.

kylens' wife :D
 
Blue Bear Sound said:
And the first one sounds better to me than the second.........
Click to expand...

Wow. I would say just the opposite.

But for the record, I am putting absolutely ZERO stock in this untill I know the methodology used for comparison.

The major beef I have with just about every comparison anyone seems to ever do on this board is complete lack of sound methodology. The testing is usually just plain flawed, because no one seems to take the time to eliminate the important variables.

The only way for this to have any meaning, for me, would be if the following were to apply:

All mixing decisions and processes, including panning and relative volume, to each individual track were fully "applied" and mixed down to a complete stereo track with mix bus at unity gain or 0.

Then each stereo track, complete with relative volume, panning decisions, effects, Eq, etc. were combined in a brand new session . . . each track including 2-track mix bus set to "0" with no panning, effects, etc. whatsoever. Mix this session down to a stereo mix and call it "X."

Then, take the very same stereo tracks and run each out to a mixer with ever channel at unity gain, no panning, eq, etc. Then simply combine / sum everything and then run the final 2-track mix back to the computer at the highest possible resolution using only the highest quality converters and call it "Y."

Then normalize each file -- relatively -- and do not apply anything to one without also applying it to the other. Dither to 16-bit and then compare.
 
Yeah, but then you are applying the leveling and panning digitally, but that is a part of the summing process. :)

And you need to find an insert point in the mixer that is post-EQ, pre-fader.

See, told ya, it's almost impossible to do this kinda test, because it's hard to rip out the summing process as a separate process.
 
Ah . . . but at least you're removing the variables. Enough to where you can attest that any differences between the two can be attributed to the summing of several audio sources to two tracks. Yes, the panning and levels are determined digitally, but as long as it's applied to both sources, you're at least removing them as a possible variable so you're not comparing two different versions of a mix.

I realize it's not going to test everything you want it to test . . . but at least you're breaking it down to one very important one: that of one thing combining/summing several audio sources to two tracks versus the other thing combining/summing more than one audio source to two tracks. I'd still find that relevant.
 
Chessrock


That all sounds about right to me. Wouldn't you want the channels all set to 0 when doing the analog part? I assumed you would want the track levels determined by the mixer and not in the box. Or are we strictly speaking the summing of tracks?
 
Yea, just the summing part. It's really the only thing that I can imagine would be simple and easy enough and still provide something relevent.

At least from there we might be able to start narrowing down where the differences lie. If someone carries out such a test as outlined, then at least you can conclude for yourself that the summing bus makes a difference to you or it doesn't. And if it doesn't, then you can at least get that out of the way and start concentrating on the faders and panning as being the main source of the differences and develop other tests from there.

And hopefully, you'll come to the conclusion, at some point, that you're just splitting hairs and what matters is making music. :D Na.
 
chessrock said:
I realize it's not going to test everything you want it to test . . . but at least you're breaking it down to one very important one: that of one thing combining/summing several audio sources to two tracks versus the other thing combining/summing more than one audio source to two tracks. I'd still find that relevant.
Click to expand...

That can be replaced by a passive resistor network, and it completely trivial to do. ;) OK, you need a good opamp to buffer the output too. Yes, it's gonna sound fine. But most of the audio-differences you'll hear will be from the DA-AD conversion. That test will test the converters, not the summing.

It wouldn't surprise me if a passive resistor network os exactly what they have in the so called "Dangerous 2 track". It would cost $200 to build, and they sell it for $3000. :)
 
regebro said:
But most of the audio-differences you'll hear will be from the DA-AD conversion. That test will test the converters, not the summing.
Click to expand...

Very true, and I hadn't thought of that one. Shit. Alright, then what if you were to take both samples and put them through the same conversion process . . . only one goes to the mixer where it's summed and the other goes straight back in to the computer. :D I realize that's a lot of work to have to do to every track, but I suppose it would eliminate that variable, right?

Then you could do yet another control sample that stays in the box . . . and then compare the two in order to hear to what extent the added conversions affect / degrade the audio, and what role it plays.
 
Yup, that would eliminate that variable.

You end up with comparing the digital adding of numbers, to the analog summing of a resistor network.

Without having made this test my self, I'll bet $100 that you won't be able to tell the difference. :)
 
You must log in or register to reply here.
Back
Top