Summing vs. Summing

[mark4man]

Peoples...

I recently asked a question over in the WaveLab forum...(about mastering w/ Wavelab & UAD-1)...whereby I wanted to bounce instrument & vocal stems out of SONAR as 24-Bit files; & master those in WaveLab...hoping it had a better audio engine; & would thereby yield a better result.

One of the more knowledgeable members over there (who is also one of the more knowledgeable members over here), explained that most modern audio engines are about the same...meaning there would be no benefit.

Then he went on to explain that I would be better off staying in SONAR, as the cumulative affect of successive DSP summing would erode the signal quality anyway.

Being a little busy of late trying to finish my CD...I just now thought seriously about this, last nite. At first I came to the conclusion that this made perfect sense...which it sorta' does...

...but also...

What about the native scenario...where I stay in the original project, route an instrument or comp set to a sub-bus, apply the plug-in to that sub-bus, routing it then back out thru the hardware mains ???

Since it's all digital; & hence all numbers...summing also occurs in this scenario as well, correct? The DAW is playing the track & adding in the numbers from the API & spitting out the sum, right?

So...what's the difference between listening to the summed track & printing the summed track.

Why would there be erosion by means of the DAW actually writing the numbers in the form of a file ???

And what about the null test ?...where the original & the bounce cancel
out.

I'm not getting this erosion thing.

mark4man

 

[Ashermusic]

This is one of the moxt contraversial topics in today's audio world.

While obviously different software uses diffeernt algorhithms it is my firm belief that sonic differences using the same hardware, same plug-ins, and basic intelligent gain staging are marginal. Usually IMHO when large differences are perceived it is from misunderstanding the way headroom works in fixed point like PT and floating point like Logic. Others disagree.

As for routing out to hardware and back in, obviously any electronics the audio is run through colors the sound and that can be good or bad depending on the mixer/summiong box and on taste. My rule of thumb, since it is a PITA to do, unless it was for a to be commercially released recording and run through a REALLY fine mixer/summing box, it is simply not worth the trouble.

 

[Dan Duskin]

digital summing is 1+1=2

it all sounds the same unless their is an error in the code

the reason this myth got started years ago is because some moron mixed in the box, and then mixed on a neve desk and found that it sounded better mixed on the desk. GEE... I WONDER WHY! What A Friggin Idiot!!!

now companys are making \"analog summing boxes\" and stealing your money. people say it sounds better running through them. GEE... I WONDER WHY! Idiots!!! When you go out of your D/A you are going through an anti-aliasing filter which rounds the high-end, then you are going through analog cabling and circutry which is adding analog harmonic distortion (NO MATTER HOW PURE THE SIGNAL PATH IS)... put these two things together and it will make it sound different. DUH! Why do they make 8 channel summing busses for this purpose instead of 2 channel (stereo)? To steal your money!!!

look at all this crap... hell, if i could work a soldering iron i might be making them and selling them for god awfull high prices too... it's amazing how dumb people are...

http://www.rollmusic.com/systems/folcrom.shtml

http://www.dangerousmusic.com/2bus.html

http://www.innertubeaudio.com/sumthing.html

http://www.nautilusmastertechnology.com/commander.htm

http://www.spl-usa.com/mixdream/2384_inshort_E.html

http://www.tube-tech.com/?show=product& ... oductId=17

Notice that the only company above that knows ANYTHING about digital audio is SPL, and they don't advertise it's use for \"fixing digital summing errors\" and crap like that... but the companies that no nothing about digital audio and just work with soldering irons advertise them as \"fixing digital summing errors\". Either they are clueless, or they are cheating people out of their money!



If you wanna do it the RIGHT way (without getting swindled)... run your whole mix (in stereo) out into one of these puppies! ...

http://www.uaudio.com/products/digital/2192/index.html

http://www.empiricallabs.com/fatdes.html

http://www.manleylabs.com/containerpages/mu99.html

http://www.manleylabs.com/containerpages/elop99.html

http://www.manleylabs.com/containerpage ... ering.html

 

[mark4man]

Jay...

You're a hell of a songwriter, man !!!

I really like \"What's Become of Love\"...clever changes, phenomenal melody...wow.

But where were we...oh...yeah, the reason I attempted this in the first place was...our projects had originally been recorded as 16-Bit (don't ask...it was the lead guitarist's fault); & I had bounced everything to 24-Bit stems for a mastering session in NY where the M.E. opened those in PT & ran them thru a world-class analog signal chain anyway. I bounced from 16 to 24 to capture more stardust out of the plug-ins (to have the nicest sounding files I could for the session.)

That didn't pan out...I ran out of money; & now I'm mastering myself.

The only time I go out of & back into the box is thru a Lexicon FX processor; & only half the time for studio verb. The other times are to \"condition\" really harsh vocals with a room, chamber or ambience program.

But on bouncing in the box...I don't think I'm in the \"degradation\" camp. There's even been times when I bounced SONAR2 files out of SONAR4 & they sounded a hair livelier or had a tad more clarity.

Thanks for the input,

mark4man

 

[electro77]

Don't the analog channelstrips with no inserts apply some phase shift and light compression that pure digital channelstrips with no inserts don't? That was my only understanding of any real difference.

 

[mark4man]

Dan...

Interesting take...very informative.

(I feel bad now, tho...I had actually been eyeballing the SPL unit for serious consideration.)

So...keeping in mind what those manufacturers seek to exploit...

Tired of the flat, narrow sound of your Digital Audio Workstation? Do individual tracks sound full and deep while whole mixes sound lifeless and thin?

...& assuming that really happens when you bounce in the box...

...say you were to go out of the DAW via ADAT lightpipe optical...8 discreet channels; & each channel representing one stem (i.e., all guitar tracks/comps assigned to Ch. 1, all drum tracks/comps assigned to Ch. 2, etc., etc.)...& into an ADAT capable digital multitracker.

Now you have 8 tracks on an outside HD system transferred there digitally...no D/A conversion in the chain & hence no rounding. You've also divided your project into 8 sets of numbers, as opposed to 2...more accuracy for each stem; & less of the \"lifeless & thin\" stereo summing thing quoted above.

Now you take each track from the multitracker & run it back into the DAW thru a world-class analog hardware piece, like the Manley Slam you mentioned (or in my case...the most world-class I can muster up at this time would be a UA 2-610 pre...which imparts a real nice analog character w/o sounding tube dark.)

Then you place the Precision Limiter across the 2 bus to catch transients & bounce those 8 tracks to a stereo file.

Would that work? You would still be summing to stereo inside the box...but you'd be utilizing analog energized stems to do so.

Any benefit here (IYO)?

I'll try it.

mark4man


Dell Dimension 8250 / Windows XP
Intel 850E Motherboard/Chipset
Intel P4 2.53GHz CPU (512 KB L2 Cache, 533 MHz FSB)
1024 MB PC1066 RDRAM
Ultra 60GB Primary HD / Single Volume (OS, Apps, Files/Folders.)
Maxtor DiamondMax 9+ 80GB Secondary HD / Single Volume (Audio Data only)
nVidia 64MB GEFORCE4 MX420 AGP
Echo Audio Layla 24/96 PCI Audio Interface
Universal Audio UAD-1 DSP Powered Plug-In System
KRK Rockit RP-8 Studio Reference Monitors
SONAR XL 2.2
SONAR4
WaveLab 5

 

[Ashermusic]

Jay...

You're a hell of a songwriter, man !!!

I really like "What's Become of Love"...clever changes, phenomenal melody...wow.

But where were we...oh...yeah, the reason I attempted this in the first place was...our projects had originally been recorded as 16-Bit (don't ask...it was the lead guitarist's fault); & I had bounced everything to 24-Bit stems for a mastering session in NY where the M.E. opened those in PT & ran them thru a world-class analog signal chain anyway. I bounced from 16 to 24 to capture more stardust out of the plug-ins (to have the nicest sounding files I could for the session.)

That didn't pan out...I ran out of money; & now I'm mastering myself.

The only time I go out of & back into the box is thru a Lexicon FX processor; & only half the time for studio verb. The other times are to "condition" really harsh vocals with a room, chamber or ambience program.

But on bouncing in the box...I don't think I'm in the "degradation" camp. There's even been times when I bounced SONAR2 files out of SONAR4 & they sounded a hair livelier or had a tad more clarity.

Thanks for the input,

mark4man

Thanks for the kind words. I wrote that song with the late Paul Jabara ("Last Dance", It's Raining Men") as well as many others for years until he passed away.

I think you are over-thinking this stuff. All of the following are more important than what you are woorying about IMHO:
1. Good song.
2. Good performances.
3. Clean non-digitally clipped recordings using the best a/d you can afford.
4. Good mix using the best d/a and flat monitoring you can afford, leaving sufficient headroom for mastering.
5. A good M.E. or if yourself, the best you can do with the mastering plug-ins available to you.

 

[Eric Dahlberg]

If you wanna do it the RIGHT way (without getting swindled)... run your whole mix (in stereo) out into one of these puppies! ...

http://www.uaudio.com/products/digital/2192/index.html

etc...

Assuming this is correct, that any benefits are gained by running the signal through "good electronics" & has nothing to do with the actual summing happening outside of the box, would the cumulative effect of running the channels individually through multiple channels of these "good electronics" be better than running an entire mix stereo through a pair of them? Let's say we compare the UA 2192 against the 8110 (which is based on the amp stage of the 2192), would the 8110 sound better because more "good electronic" stages are being used or might you just as well pre-mix the channels & run them through a 2192?

 

[Dan Duskin]

if you ran 16 outputs into an analog summing box for \"summing\", and then compared it to running the mix out in stereo and using just two of the analog summing boxes channels it would sound exactly the same

the only way to get \"more warmth\" out of a summing box would be to open it up, grab a soldering iron, and run each channel chained into the next one, so that the stereo output goes through 8 stages instead of just one. splitting the channels and summing them back together is the same as going through just 1 stage.

however... 95% of what the users of these stupid boxes are hearing is the anti-aliasing filters on their D/A-A/D converters. they could have got the same sound by two two 1/4\" cables and connecting their D/A to their A/D.

 

[Dan Duskin]

Tired of the flat, narrow sound of your Digital Audio Workstation? Do individual tracks sound full and deep while whole mixes sound lifeless and thin?

it's called \"learning how to mix\"

that's why mixing is an art, you can't just bring the faders up and hope it sounds full.

so, yeah... what a great marketing ploy! of course... everyone would say they've noticed that mixes sound thin compared to solo'd tracks... that's normal, until it's mixed properly.

 

[Dan Duskin]

here... i will prove it!

- create a new project in your daw program
- insert some audio on a track
- copy that track 20 times (so you have 21 tracks)
- set the volume of the 1st track to 0dB
- set the volume of the other 20 tracks to -26dB
- mute the first track
- now hit play and listen
- while during playback, periodically hit the solo button of the 1st track

...you will hear no difference!

their is no \"flaw\" in digital summing.

the point is that it has NOTHING to do with summing! everyone needs to stop using this word!

ok... so mabye you think \"what does that have to do with benafits of analog summing?\" ...again, it has nothing to do with the summing! try this test...

using the project above:
- insert a compressor plugin on each channel (all 21 of them)
- make sure you use the same plugin
- make sure you use the same settings

now you might think listening to the 20 channels playing at the same time would sound different because you are running through 20 compressors. NOPE! same theory goes to running 16 analog outs into a 16 channel summing box... it's the same as going out stereo (2 channels) into the summing box!

it can't be argued... it's a simple math equation. i baffles my mind that people spend all thing money on 8 or 16 channel summing boxes when they could get the same result from a 2 channel line level box.

 

[taylor]

dan, following your explanations carefully... becuase i'm an ITB mixer who is caught up in the summing buzz.

so.. if you're correct that digital summing does not degrade the sound at all... don't you still feel there would be a benefit to running your signal through the nice sounding analog guts of an analog summing mixer?

or are you saying there isn't much difference between that and simply using a 2-bus device....

i'm actually looking into the Neve Portico tape emulator to do just that... run my 2-bus mix through it.

i have the amazing Color Tone Pro plug in.. and the Portico IR for that is like magic for my mixes... i could only imagine the real thing would be even sweeter...

 

[Royal T]

obviously, there will be some summing errors, due to rounding errors, occuring with the current 32 bit software.

higher bits, less rounding errors.

 

[davidkakon]

I suggest that anyone who has summing envy purchase the 3daudio neve vs protools cd. It is a real ear opener. What I suggest as the best improvement on any mix is to invest in a great front end (room, instruments, amps, mics, preamps, converters) and hire a mix engineer that has great ears and knows how to use software that sounds good.

good luck
david

 

[Dan Duskin]

or are you saying there isn't much difference between that and simply using a 2-bus device....

exactly!

nothing wrong with going D/A to run through something that makes it sound better. it's just stupid to pay 4-8 times as much for 8-16 channels when you will get the EXACT same sound with 2 channels.

 

[Dan Duskin]

obviously, there will be some summing errors, due to rounding errors, occuring with the current 32 bit software.

higher bits, less rounding errors.

and you will get those rounding errors when you go D/A and D/A again.

 

[Dan Duskin]

I suggest that anyone who has summing envy purchase the 3daudio neve vs protools cd. It is a real ear opener. What I suggest as the best improvement on any mix is to invest in a great front end (room, instruments, amps, mics, preamps, converters) and hire a mix engineer that has great ears and knows how to use software that sounds good.

good luck
david

I agree with your suggestion regarding front-end gear. But in regards to the pro-tools versus neve thing, that has nothing to do with summing... it sounds different because a neve desk adds it's own color, and not do to summing.

I also want to point out that the back-end will also work well. i.e., like analog mastering!

 

[brian]

This whole summing thing is complete bullshit. I have mixed on a Neve VR, SSL 6000G+, SSL 9000J, ITB, and on some digital boards...and the best results I get are by doing the summing math manually. Go out and buy a TI-82/83/89/etc calculator. Add up all the tracks on the calc and plug the final value back into your DAW.

I prefer the sound of the TI-89 personally, but the original TI-82s worked great for summing. The later TI-83s had more rounded buttons and chassis, as did the re-released TI-82s, which made the final mix sound less precise to my ears.

When I want some more color/warmth on the final summed mix, I get nice and drunk and do all the track summing by hand with a pen and paper. This introduces some random \"analog\" style human-error that gives the track \"that\" sound, instead of the boring clinical digitally-summed flavor. This method will only cost you around $8 for a 6-pack of Sierra Nevada Celebration Ale, or you could buy into all the marketing BS and blow $1500 on some ripoff summing box like a sucker. Plus once you are done with the beer bottles, you can recover some of the investment at the recycling plant and buy more pens or a new notepad for the next mix!

 

[Royal T]

and you will get those rounding errors when you go D/A and D/A again.

doesn't necessarily sound the same, but I agree, it's sorta lame buying a box just for summing, when daw makers could improve the itb summing experience, using less erroneous math.

has anyone checked out sonar 5? it might have better sounding summing due to the higher bit resolution.

 

[Dan Duskin]

[quote="Dan Duskin":33b9apf7]

and you will get those rounding errors when you go D/A and D/A again.

doesn't necessarily sound the same, but I agree, it's sorta lame buying a box just for summing, when daw makers could improve the itb summing experience, using less erroneous math.

has anyone checked out sonar 5? it might have better sounding summing due to the higher bit resolution.[/quote:33b9apf7]


believe me... i am all for daw software that has a 64bit floating-point mixing engine. however, higher sample rates will give you similar results... because the rounding ends up fluctuating (more like an image dither) resulting in the sound of a higher dynamic, and inter-dynamic, resolution.

my point of reply to your comment was that the rounding ends up being the same... i.e., if it's by the A/D or in digital summing the rounding resolution is the same... if you combine everything in the analog world it makes no difference, because the summed numbers still end up rounding the same. again... summing outboard or in-the-box, it's the same. the difference in sound is the analog components, not the summing. i'm sick to death of hearing about "summing"