<A bit OT> Sonar and hardware summing

Author
Razorwit
Max Output Level: -66 dBFS
  • Total Posts : 1235
  • Joined: 2003/11/05 18:39:32
  • Location: SLC, UT
  • Status: offline
2011/04/08 17:49:20 (permalink)

<A bit OT> Sonar and hardware summing

Hi folks,
I know there have been some posts in the past about summing mixers/boxes and Sonar and I've been putting in some time trying to discover if there was really any "there" there (so to speak). I've gotten some interesting results and thought I'd share with the board what I've done, how I set it up and what the end result is.
 
Without going into too much detail, summing mixers are hardware devices that have no or very minimal controls.  The theory is that the process of taking the analog channels of a mixer and combining them into a sinlge stereo track (summing) has a particular, and in many cases desirable, effect. That theory led to the prodution of a bunch of summing devices like the dangerous music d-box.  After some reading I wanted to find out what this was all about.  So, that said, here's what I've done:
 
I have a pair of channels coming out of my Lynx Aurora interface that go to two places (via a patchbay).  One is my mains, the other is a pair of input channels on an SSL X-Rack 4 channel input module.  That input module takes the signal and passes it to the master module which then goes out an insert into a SSL X-Rack bus compressor and an SSL X-Rack Stereo EQ before coming back in the master module and finally out the record bus.  The record bus then goes into two input channels on the Lynx.  For this purpose the EQ and compressor have been disengaged, that is, the "in" button is not depressed so that they are not affecting the sound other than by having it pass through.  So the signal chain looks like this:
 
Lynx out -> SSL input -> SSL Master -> SSL Compressor (disengaged) -> SSL EQ (disengaged) -> SSL Master -> Lynx In
 
I wondered if this setup affected the sound and if so, in what way.  To test this I took a short drum phrase played with Jet City Drums and recorded it out of Sonar, through the SSL box and then back into Sonar. I then bounced it to a track without passing through the SSL box and then compared by ear and with an analyzer.  The results are interesting.  You can grab both wav files here:
http://www.xmission.com/~razorwit/SSL/SSL.zip
You may want to have a listen before reading any further so that my opinions don't affect what you hear.
What I heard was a change in the bass frequencies when passing through the summing box.  It seemed there was a tonal shift and some volume change.  It also seemed like the bass was a bit smoother (yeah, I hate that term too but it's the only descriptor I can think of) in the SSL version. To test this I isolated the kick and did the same thing, only this time I used the analyzer to compare.  Have a look at the frequecy plots below:
 

 
Notice that in the SSL plot (the top one) the very low freqencies (about 35 hz) are increased and then there is a dip in the 40 to 80 range. Also, the SSL box seems to cut a bit of the mid and high frequencies (above about 640).  I was sort of both right and wrong in what I was hearing.  I think the tonal shift was correct, but what I heard as smoother and louder bass was actually a dip in the 40 to 80 region and that lower bump along with the lower frequencies being emphasized when compared to the higher frequencies.
So, my conclusions, at least from this test: 
Is there a difference in sound when passing through a summing box, at least for this case?  Yes. 
Was it a big change?  Not really, but certainly not negligible.
Is it a good or bad change?  That's waaaay to subjective and really up for opinions, but it is interesting. 
 
What I'm really curious about next is what that tonal effect sounds like when it is additive.  This was a very simple test that only measured what signal passing through two channels of my SSL box does.  The real question is what happens when I have 12 channels passing through it and being summed instead of just 2.  Sadly, that will have to wait until I can afford an 8 channel input module, and right now is not that time....though if anyone wants to send one my way let me know! :)
 
Anyway, there it is.  Hope you folks find this as interesting as I did.
 
Dean

Intel Core i7; 32GB RAM; Win10 Pro x64;RME HDSPe MADI FX; Orion 32 and Lynx Aurora 16; Mics and other stuff...
#1

7 Replies Related Threads

    Jeff Evans
    Max Output Level: -24 dBFS
    • Total Posts : 5139
    • Joined: 2009/04/13 18:20:16
    • Location: Ballarat, Australia
    • Status: offline
    Re:<A bit OT> Sonar and hardware summing 2011/04/08 19:20:22 (permalink)
    Hi Razorwit firstly this is a Techniques thing. It is not just aligned to Sonar X1. Any software can do analog summing.

    You need to read (all of) this thread here. It has been done to death as they say:

    http://forum.cakewalk.com/tm.aspx?m=2134073

    There are some problems with your experiment. Firstly you have not used a true summing device so it is a bit invalid. You have used a two channel analog device and the signal is coming out at the same level it is going in. Also are there any transformers inside your SSL device. Because if so, they colour the sound, simple as that.  Also the device in question may in fact be designed to colour the sound in some way. (even if there are no transformers)

    True passive analog summing looses a lot of gain and makeup gain is required. It is the quality of the makeup gain device that has a lot of bearing on what you hear.

    From the thread link above it can be deduced that the analog version of the waves was infact inferior to the digital version. We were all seduced by the initial bottom end and top end of the analog version but after careful listening, other problems emerged from the analog version. You are not listening for the right things. You are being seduced by frequency changes, but many other things come into play, transients, firmness of stereo image etc.. Skullsession used two (probably unmatched)  preamps with transformers in them for the makeup gain hence the reason the digital version actually sounds better. Some of us got carried away in that thread myself included and I apologise for that but I think some interesting facts also came out of it too.

    I have read a few articles on this and in very controlled experiments it was the opposite conclusion that was arrived at, and that was the differences between ITB and OTB summing were very hard to pick even by skilled listeners.




    post edited by Jeff Evans - 2011/04/08 19:23:02

    Specs i5-2500K 3.5 Ghz - 8 Gb RAM - Win 7 64 bit - ATI Radeon HD6900 Series - RME PCI HDSP9632 - Steinberg Midex 8 Midi interface - Faderport 8- Studio One V4 - iMac 2.5Ghz Core i5 - Sierra 10.12.6 - Focusrite Clarett thunderbolt interface 
     
    Poor minds talk about people, average minds talk about events, great minds talk about ideas -Eleanor Roosevelt
    #2
    John
    Forum Host
    • Total Posts : 30467
    • Joined: 2003/11/06 11:53:17
    • Status: offline
    Re:<A bit OT> Sonar and hardware summing 2011/04/08 20:17:57 (permalink)
    Jeff the only problem I have with your post is the repeated misspelling of Color. LOL

    Best
    John
    #3
    bitflipper
    01100010 01101001 01110100 01100110 01101100 01101
    • Total Posts : 26036
    • Joined: 2006/09/17 11:23:23
    • Location: Everett, WA USA
    • Status: offline
    Re:<A bit OT> Sonar and hardware summing 2011/04/08 21:15:55 (permalink)
    A passive summing device shouldn't color sound. It's just a network of resistors.

    Resistors have very small inductance and capacitance and would normally only effect very high (way beyond audio) frequencies. A summing device would avoid transformers or active circuitry of any kind so as to not color the sound. They are necessarily simple by design. (So why are such simple devices priced the way they are? Beats the heck outa me.)

    The purported benefit of analog summing has to do with adding channels together in the analog domain and avoiding distortion caused by digital summing and its inherent approximations and rounding. Or something like that. It's all hokum as far as I can tell, but irrelevant to your experience.

    More likely, something in your path is introducing some spectral nonlinearity along the way. Something as benign as a cable can do that, although I doubt a cable would affect frequencies as low as those in your experiment. Even bypassed devices can have an effect, so if you want to be certain the compressor and EQ aren't playing a part you'll have to physically patch around them.

    If you can get your hands on an analog spectrum analyzer you could probably determine where in the chain the nonlinearity is being introduced. My guess is that it has little or nothing to do with going out to analog domain and coming back in again.

    Thanks for posting, Dean. We need more of this kind of stuff in the forum and less...of the other stuff.


    All else is in doubt, so this is the truth I cling to. 

    My Stuff
    #4
    Razorwit
    Max Output Level: -66 dBFS
    • Total Posts : 1235
    • Joined: 2003/11/05 18:39:32
    • Location: SLC, UT
    • Status: offline
    Re:<A bit OT> Sonar and hardware summing 2011/04/08 21:29:56 (permalink)
    Hi Jeff,
    thanks for the comments. In retrospect, this probably should have gone in techniques.  My apologies.
     
    A couple quick notes.  I'm trying to be fairly careful about staying away from subjective analysis like "superior" and "inferior", but if I've managed to do it poorly (which is entirely possible), I do want to say that I'm really not making any statements about better or worse here, only different.  In fact, I'm still not particularly sold on the usefulness of the sound imparted by the SSL, but it's an interesting potential tool. 
     
    Beyond that, just a couple other notes.  Regarding transformers, I suspect you're correct that at least part (and probably closer to all) all of the audio difference is due to transformers.  It certainly has what I think of as that "transformer sound" (though I don't know the specs of the SSL modules to say for certain).  However, the presence of transformers certainly does not mean that it's not a summing box.  SSL is quite clear that the intent of their 4 and 8 input modules is to build a summing box, and other companies have also made summing boxes with transformers...the Neve 8816 comes to mind here. I think there may be a good distinction between this and passive summing though. Regarding only two channels, I deliberately did that in order to reduce to basics.  I don't doubt that the results may be different if I was summing more channels, but for this I was interested primarily in the sound of the hardware. I'm also curious what it will sound like with 12, but that will have to wait.
     
    Oh, and thanks for the link to the other discussion.  I'll have a look a bit later tonight.
     
    Thanks again
    Dean

    Intel Core i7; 32GB RAM; Win10 Pro x64;RME HDSPe MADI FX; Orion 32 and Lynx Aurora 16; Mics and other stuff...
    #5
    AT
    Max Output Level: 0 dBFS
    • Total Posts : 10654
    • Joined: 2004/01/09 10:42:46
    • Location: TeXaS
    • Status: offline
    Re:<A bit OT> Sonar and hardware summing 2011/04/08 22:26:02 (permalink)
    Dean,

    cool.  But what you aren't testing (as has been pointed out above) is not summing but the effect of the analog mix bus (and associated electronics) on the signal.  A few years ago SOS did a test on one of the new (at that time) summing mixers and found just running the signal through it in stereo acheived most of the sound of the summing mixer.  There might be something to analog summing or might not - there is definately something (as you've proved) to going through a nice analog chain, even if some of it isn't engaged.

    All the old guys who made the transition from custom made recording gear to "produced"  gear - Neve, Putnam etc. were trying to overcome the nonlinearties - distortion, in the gear of the time.  They used their ears, as well as science, not just electronic tools to get the best design.  So, whether we expect these irregularities in recording or they really do make the sound more pleasing doesn't matter.  They did a good job of tuning the right kind of distortion.  Just running a signal through such well designed audio equipment will make it sound different - many will think better.

    The cleaner our recording ability has gotten, the more people enjoy adding some of the irregularities, saturation or pixie dust back into the media.

    https://soundcloud.com/a-pleasure-dome
    http://www.bnoir-film.com/  
     
    there came forth little children out of the city, and mocked him, and said unto him, Go up, thou bald head; go up, thou bald head.
    24 And he turned back, and looked on them, and cursed them in the name of the Lord. And there came forth two she bears out of the wood, and tare forty and two children of them.
    #6
    Razorwit
    Max Output Level: -66 dBFS
    • Total Posts : 1235
    • Joined: 2003/11/05 18:39:32
    • Location: SLC, UT
    • Status: offline
    Re:<A bit OT> Sonar and hardware summing 2011/04/09 02:18:14 (permalink)
    Hi AT and Bit,
    Thanks for the responses.  I think it's probably a point well taken to discern between the process of summing and a summing device.  If I read you correctly, the distinction here is that even though this was going through a summing device, there wasn't much actual summing (the process) going on, and thus what I was testing above is a summing device, but potentially not (much of) a summing process.  This is a good point and I'm actually curious to test it once I get a few more inputs to the SSL box.  My suspicion is that if there is a particular sound to the summing process (and that's a big "if"....I think I remember the SOS article you're talking about AT and I think I'll probably have the same result), it it's probably more apparent the more sources you sum, but having no direct evidence to back that up that's something that I'll have to try out...that's half the fun after all :)

    Oh, one more thing.  I need to go and correct my OP, but I found out that the compressor and EQ were not inline after all.  I was recording off the record bus, and the master module insert loop isn't inline with the record bus.  So my SSL chain was only:

    Lynx out -> SSL input -> SSL Master -> Lynx In

    I had thought that the effect modules were coloring the sound even disengaged, but with them out of the chain it means that only the input and master module were coloring.  Well...and the Lynx Aurora, but they're awfully clean...

    Anyway, thanks again for the comments.  I must say that I've enjoyed most of the other stuff that the participants on this thread have posted over the years, so it's nice to have comments from you guys on this.  I appreciate it.

    Dean

    Intel Core i7; 32GB RAM; Win10 Pro x64;RME HDSPe MADI FX; Orion 32 and Lynx Aurora 16; Mics and other stuff...
    #7
    AT
    Max Output Level: 0 dBFS
    • Total Posts : 10654
    • Joined: 2004/01/09 10:42:46
    • Location: TeXaS
    • Status: offline
    Re:<A bit OT> Sonar and hardware summing 2011/04/09 09:50:52 (permalink)
    Hey Dean,

    engage those suckers and see if you hear more of a change  I think analog summing really helps when you comp/EQ the stems going through.  Of course, then it ain't just summing.

    @

    https://soundcloud.com/a-pleasure-dome
    http://www.bnoir-film.com/  
     
    there came forth little children out of the city, and mocked him, and said unto him, Go up, thou bald head; go up, thou bald head.
    24 And he turned back, and looked on them, and cursed them in the name of the Lord. And there came forth two she bears out of the wood, and tare forty and two children of them.
    #8
    Jump to:
    © 2024 APG vNext Commercial Version 5.1