COMPLETED: The Analog Summing Experiment

I'd say you're probably not interested any more in ITB summing....or summin like that ;)
 
 

This is interesting:
 
https://www.recordingrevolution.com/in-the-box-mixer-ken-andrews-wins-blind-shootout-over-analog-console-mixes/
 
This is a boring subject actually and also has been done to death. I get the fact you may use the mixer live and that is great but if you think it is somehow going to transform your mix you may be disappointed.
 
Many great engineers are ITB now that started out mixing on a console. That fact alone says it all. If the music ideas are great and the mix is great then it won't matter so much.  If the music and the mix are average then nothing will improve it either.
 
The fact now we have some great console emulations going on has made things even more interesting. 
 
Be careful how you do any tests.  Many AB tests are flawed and I could probably shoot a giant hole in your method.  If you are going to do this then you would only be able to use the analog mixer with no EQ or no processing applied anywhere. (and even then how would you know for sure that even with all your EQ's set for no eq is it truely flat or not. Can the EQ be switched out?)
 
Panning would have to be carefully set. Pan laws taken into account. Exact fader positions on both mixes ITB and outside etc.  You have to go to a lot of trouble to remove any possible variables. 
 
There was a great article in Audio Technology magazine a few years ago and I wish I could find it but cannot unfortunately. They did this exact thing and displayed the mix as a very complex three dimensional plot and both were identical and I mean identical. 
 
The fact that in that link above a room full of great ears all chose the ITB mix over all the others that were console mixes certainly shows something don't you think. The convenience of ITB mixing alone puts it miles ahead.

I think you should do some experimenting and I would be interested hearing the results too.  Pan laws come into play if you pan a mono source centre on the analog mixer.  You need to match that in the ITB case too.  But if as you say you are going to use stereo sources then that is a great way to get around it.
 
A null test would be interesting to do too as long as you can replicate everything perfectly on the analog mixer and the ITB situation. You have to be careful where you feed your input signals in as well. I have found from experience that input Mic Pre gain settings are not all equal between channels. Even if they are set to full anti clockwise etc. 
 
You would have to start by bypassing any input Mic Pres and feed signals into direct tape returns etc..try to avoid any variables there.
 
After years of using consoles and then going over to a digital mixer I have really discovered that every analog mixer actually imparts a sound onto your mix.  Not the summing engine so much but everything else.  Sometimes it may not be that great.  Unless you are using state of the art analog mixers which most of are not.  Most mixers are in the medium range category.  
 
I mixed an album on large format Tascam mixer a few years ago and it sounds pretty nice too but then after going all digital I started hear the sound of that mixer.  When I remixed one of those album tracks digitally in the end I preferred the digital mixer.  The analog sound started to become more apparent to me.  This is just me but once I got used to the pristine transparent sound that an all digital world offers I started to hear the analog sound more and it colours the sound slightly.  Everybody thinks analog is superior but it is actually inferior.  If there are any transformers involved you are getting distortion.
 
What I do love now though and this is something no analog mixer can do is the fact we can pass parts of our mix through analog sounding plugins but leave other parts alone.  When you do a mix on an all analog desk it is applying its sound to every source.  There is no way around it.
 

Cool! I'm interested in following your progress here. I just invested in Waves NLS and will try that on a couple of mixes tonight to see what it adds. I work primarily in Logic on a personal basis so no more PC console emulator for me :(
 
However, I have used hardware summers on several occasions before. In my case(s) it's been just summing boxes and not a console like yours. So far, the box with the most to offer - to my ears - has been the Thermionic Culture Fat Bustard. That made 5 mixes for an EP sound like night and day. The software alternatives I have tried didn't offer that much and some other summing boxes couldn't quite live up to that one either. But in every case, there's been a difference. I've been wanting to get a Fat Bustard but for some reason, I haven't. I hope NLS can offer back some of what I miss from Cake's PC module. But I still believe that a good hardware summing device adds something special.
 
I hope your experiment here will shed some light on this. Best of luck and have fun!

Soundwise
 
Keep in mind that cables and patch cords impact the sound significantly. That is, the impact is both measurable and audible.

Excellent point! I will try to make the comparison as clean as possible, but in the end I am really looking for whether I like it "better" with or without going out to the console. I try not to buy junk cables, but I'm also not where I was when I was into a guitar rack where I was buying custom Lava cable and only using solder with a certain % of silver... :)

OK, here's what I'm thinking of for testing methodology:
 
Disclaimer, I know it best would be to mix into the changes but that's not feasible so will just be looking for differences on an existing mix. 
 
1. Output mix as developed with no outside summing
2. Create 16 channels in DAW to send to 16 tracks on Console and route all DAW tracks to one or a pair of these 16. That way, everything goes to the Midas.
3. Bring the summed signal back into the Daw prior to any mix bus plugins - so it's the same behavior as having all tracks ITB go to the mix bus.
4. Create files for the following scenarios:
   a. Using the Inserts of the console, bypassing both the preamps and the EQ. I will send a test tone and attempt to level each channels' fader so everything is at unity.

• Using the Inserts of the console to bypass the Pres, but putting the EQ in the chain with EQ zeroed out.
• Go line in to the console through the pres at unity, EQ bypassed.
• Go line in to the console through the pres at unity and EQ in path, but zeroed out.
• Go line in to the console through the pres with the pres slightly above unity to get some of the preamp sound, EQ bypassed.

 
That should give plenty of interesting reference tracks to listen to.
 
Am I missing something? Any other ideas?
 
I have a Stam SA4000 on order and want to play with that as a bus comp, but that's the next phase. I have actually set up the board to use an Aux to feed a sidechain of the bus comp (so I can adjust which tracks feed it and EQ it); I am looking forward to that experiment! :)

ha ha didn't look close at the two cables I bought from GC today. Thought both had male TRS, but wouldn't ya know one of them is a girl!
 
Also found I didn't have patch cables to return the sum to the Apollo.
 
Ha HA hardware rulz! :)
 
James

I dreamed of cables and faders all night long. LOL
 
I'm back at it though - and I did make a lot of progress!

From Sonoris website:
 
Calibration is simple, first select what channel to use for the calibrated pink noise (bandlimited 20Hz-20KHz or 500Hz-2KHz), L, R or both. You can calibrate your monitors at this level to show 83dB on a slow, C-weighted SPL meter using your monitor control. In stereo mode add 3dB. Mark the monitor control levels for each scale and you are done!
 
The 500Hz-2KHz noise signal can give a better and more stable reading on the SPL meter under some circumstances. Please try both options and see what works best for you.

Hi Jeff,
 
I just plugged the TRS connections from my DB25 snake into the inserts and they all work perfectly! I get what you are describing, but they are all working in my case...
 
The UAD let's be control the the reference level of all pairs on the Apollos by choosing either +4 or -10. Right now they are all defaulted at +4 dBu. 
 
The signal is good and strong coming into the board and in fact now that I've "roughed in the levels" it's grooving and I'm resisting the urge to create a copy of the project and wander off with it. Something sounds great, but at this point I may just have the drums and bass too loud or something; so I'll stay focused on calibrating.
 
I had an interesting experience. In Ableton all tracks are stereo tracks. So ran the bass out in stereo to a pair of stereo ins and it all sounded fine. For giggles, I thought I'd try the bass in mono so I put a mono plugin on the track that goes out to the Midas, switched to send on only one out and not a stereo pair, and chose a different single insert on the board that was a stereo track.
 
Now the bass was too low. I think maybe the bass was mono to begin with and when I put the mono plugin in it I halved the volume. I need to test. Anyway, I put a gain switch on at +6db and it sounded good. I cranked it to +9 db and it sounded GREAT! Not objective by any means, but still a shload of fun! :)

OK, here's a first set of files. :)
 
1. BelieveITB.wav is the In the Box version. This is how the song was originally mixed and rendered.
 
https://www.dropbox.com/s/9xkvbs6kp1chhrc/BelieveITB.wav?dl=0
 
2. Believe direct no eq.wav. This is the first attempt at summing. All tracks where tested for unity with a 1k sine wave. VUMT wasn't playing nice so I used Dorrough. That's not quite exact though, so I may want to do it with VUMT again. At any rate, they are close!
 
https://www.dropbox.com/s/50s24hrolsdpxxo/Believe%20direct%20no%20eq.wav?dl=0
 
There is no limiter on these, so feel free to use your own.
 
I added this to a project, solod one and played the other then mapped the solo button for each track to the same midi key on my Novation LaucnControl. This allows me to toggle between then with one touch of a button.
 
I absolutely hear a difference! I also don't ALWAYS hear the difference! :)
 
Tempo is set to 88, so use that for a reference with my notes:
 
PS: I believe they are level matched but if you find they aren't, please let me know.
 
1.   The sound effects at the very beginning sound a bit more distant in the outboard mix. The kit almost seems a bit lower.
2. But at measure 5 (10s) the kit sounds pretty much identical.
 
3. At measure 19 the ITB mix guitars just disappear. I can't find any differences in the mix though to explain this. :/
 
4. The Aux buses with the reverbs are much more pronounced in the outboard summing mix. They would need to be brought done. You can hear them when the vocals start at 29. 
 
5. The build up/rise that starts about 42 sounds "smoother" to me on the outboard summing. 
 
6. The bass line that starts around 51 (one of my favorite parts, thanks Bapu!) seems more focused in the out of the box mix and seems to sit better to me.  The low end in general is much better and very noticeable here!
 
7. The synth bass line that starts about measure 64 again seems more focused and better blended in the out of the box summing mix.
 
8. At measure 70, the background "ooohs" seem fuller in the out of the box mix. 
 
Let me know what you hear! Also, let me know if you think I messed up the methodology at all...
 
This is actually challenging, but a whole lot of fun! I've got some other outboard gear on order and will compare methodologically as well when it gets here. :)

OK listened to these.  Firstly they don't null in any form.  Not even close.  The two waveforms are very different.  So that  means the test is not accurate.  When Audio Technology did this they got very close to a perfect match of waveforms.  Something that is not happening here.
 
I think the mixes between the two are too different.  I am hearing more kick in one etc..
 
Also has the ITB mix got no eq etc everywhere the same as the OTB mix?
 
My impressions are there is nothing I could not do in the ITB mix that is happening in the OTB mix.  i.e. not worth the hassle of sending everything out to mixer etc...
 
If you read that link I posted right at the start it is saying someone who does a better ITB mix will sound better than someone doing an ordinary mix OTB. I agree with that.

Yes it is tricky getting an identical mix with ITB and OTB setups. I think you need totally mono tracks. They should only be panned L C and R. The pan law for your mixer needs to match that of the ITB version. You cannot change the pan laws for the mixer so you need to find out what it is.  (i.e. send a tone from a channel to one side of the Midas mixer e.g. L. Set the output meter to read 0 dB VU on the left channel. Then without changing anything, pan centre and see how far it drops down. It will either be 0,-3, -4.5 or -6 db. Then you need to set the ITB DAW for the same pan law.)
 
I think the best way is to set the mix via clip gains only and leave all your ITB faders at unity. This is hard in Sonar because you cannot do it easily.  In Studio One you can by just grabbing the gain handle of the audio event on each track and pulling it up or down. The level and the waveform height change. 
 
You also send all the tracks out to the mixer and set all their gains at unity too. (which you have calibrated of course to all be identical) 
 
No plugins anywhere ITB of course and all EQ's bypassed on the mixer. No dynamics or reverbs anywhere.
 
Theoretically the same mix should exist ITB and at the stereo outs of the mixer which of course you record back into your DAW.
 
I bet if you did this, things would be much closer.