timidi


Awesome stuff...............

Here's a maybe client for ya:)
http://www.audiotalkback.com/forums/ubbthreads.php/topics/140398/A_summing_mixer#Post140398

Heh...no doubt.  I could make the box for less than $100 if I didn't have to supply cables.  And he's willing to pay up to $2,000.  But he wants EQ.  So I suppose I could only charge $1,000 so he'd still have cash left over for an EQ of his choosing.
 
win/win
 
 

Bristol_Jonesey


Great thread Skull.

I could almost smell the solder when going through the build pages.

Right, Jonesey.  I had a good time with building the box.  But it was 14" TRS cables at a qty of 16 + that got really old!
 
I did discover a new trick though, so it was completely worth it.  In fact, I should take photos and post it here so yall can see.  That Canare cable has an awesome woven shielding in it, but it's a real pain in the butt to unweave while you're building cables.  After about two of them, I got completely sick of how long it was taking...so I found a wire brush.  By using the wire brush and GENTLY brushing the woven wires, they'd come UNwoven perfectly without breaking or getting tangled up in the cotton insulator that's also in there.  It only took a few strokes with the brush and BOOM...there it was....unwoven and ready to work!
 
 

It might be interesting to hear the results of running both outputs through a decent preamp at the end of the chain. That is the normal approach with passive summing i.e. from the DAW to the summing device into a decent stereo preamp for additional color and potential image enhancement. That might yield some additional revelations. Just a suggestion.

I assume that was done on both files otherwise a true comparison is tough.

bmdaustin


First file sounds a little darker, warmer, more depth.

Second file sounds brighter, more brittle, edgier.

I have to agree with this analysis.

Of course it is a fair test... Skull did his best with either method... ergo we are testing which method works better in that context.

There is no direct parallel when comparing ITB to OTB... so there is little need to parse the methods in an effort to match the process step by step... you do your best with either approach and that's what you have when it's time to compare.

best regards,
mike

Mike no insults intended. I am merely pointing out that sometimes people get convinced that they are hearing the analog version of something and in the end they find out they are listening to the digital one. I hope that most are picking the digital version here as it can be argued that digital actually sounds better than analog but then most people are not prepared to admit that. They are stuck in analog and can't get out.

So if most are picking the digital file then it is because digital actually sounds better. If you pick the summing file then it could also be argued that you are hearing the preamps that were used for the makeup gain. And is that file better or just different. So the test is really all a bit inconclusive.

I am not going to bother because basically I am thrilled at how great my mixes now sound ITB and all digital, and really that is all you have to do. Once you achieve a great ITB mix then you can just let go of chasing the elusive analog summing sound etc.. The rest as we all know is in the song, delivery, performance etc..That is where it really counts.

I am not doubting that there is a difference between the files. I am bit busy right now and if I can I am happy to do the test and what if I hear the difference. My point is one version is digital and the other could be the result of the preamps used for the makeup gain so therefore one is not really exposing the actual summing part of the test. If Skull is using the Daking Mic Pres they have a Jensen input transformer. So is the transformer adding anything to the sound. Yes it is.

I originally had an idea here that you could attenuate the digital version and also take the summed low level version and feed both versions of the mix through the DAKING Pres. But now as this experiment is complete, what has eventuated is that the Mic Pre version signal path is worse than the digital mix so I would only suggest a better way would be to compare the ITB mix with the passive summing mix with a much higher precision preamp as the makeup gain device.

I am not lecturing, I am merely pointing out the possible inconsistencies in Skulls test. People rush off doing AB tests and they are still not eliminating all the possible variables between the A and the B signals. It is hard to do. You have to get it down to ONE variable only. The Skull  test has got two variables already. (the summing box and the preamps) So which is it you are hearing?

Rupert Neve makes a summing box. I have got a friend who has one in a top studio here in Melbourne and he says even if you feed one signal into it and out again it sounds better. Sure I don't doubt it. But as we know Rupert has got active electronics making up the gain. So he has probably got some great sounding class A amplifer and a transformer in the output stage. (Rupert just loves transformers!) That could be why it sounds so good.

I still think (and we know know) the preamp should have no transformers involved and be a true transistor pure class A design of some sort. Then you are taking out what the transformer is adding to the sound. It may make it harder to hear the difference because it could be adding something nice into both versions.

Mike you have already thanked me in the past for getting people to think about what they are talking about and doing. This is no different.

And editing your posts.

Skull I am sorry for the confusion. My friend says that perhaps an individual track being patched through the Neve summing amplifer might sound better than going direct to the stereo buss in say Pro Tools for example. And this could be for various reasons. I personally believe however that the sound of any great analog processor can be produced digitally and therefore you can arrive at a similar point. eg a nice warm sounding track or buss etc..Plugins are just amazing now.

So digital is better because it can sound like many different analog sources, but not so the other way around. It is smarter but we all know that.

OK I have done some listening tests and yes I have some observations. For me there are some issues with the mix. They make the evaluation harder than it should be. But these two corrections below improve the mix and in my opinion make it easier to hear the differences between the two tracks. Also you can play them louder which is nice too.

The snare sounds a little chocky at 225 Hz and on Span it pushes up higher than other frequencies around it.  It makes the lower mids a bit thicker than they should be. But it is easy to build up energy around that area and in mastering they often take it out a little.  (I do anyway) A highish Q type, dip at that frequency and only 3 or 4 db or so fixes it nice. The bottom end becomes clearer.

I feel the vocal sound is harsh. If you are intending to get that gritty nasty upper highs thing going in the vocals then you have achieved it for sure. But I would not master it that way though. Once again on Span there is a nasty peak at 2.65 KHz every time the vocals are present and it is much higher than everything else too. This gets in the way of hearing the subtle differences between the two tracks around this area of the spectrum. I found the section without the solo very helpful too. You can hear kick standing out etc..and how the guitars really sound. I also found it useful to loop over 2 or 4 bars of the music and switch back and forth.

So a little notch at 2.65 Khz even 4 or 5 db down and not too wide either brings those vocals back into line and the mix  sounds better. I had both tracks parked on sep tracks and had this eq on the master buss so it applied to both. Use the mastering eq to ensure best sound etc.. A mastering engineer would do this for sure. Of course ideally, you could do it on the vocals and snare tracks. The music behind the vocals sounds great. If its a big fat natural vocal sound you were after you have not achieved it. Sounds like there are a lot processors on the vocal track. But these are very musical decisions as well. Your current vocal sounds good in the mastering stage if you correct around 2.65 K, the rest of the music does not seem to change much.

Now the music is great and I liked it and to be honest I liked the music the same either way. I can hear a difference. The kick sometimes gives it away too. And after a while the snare shows up the differences too. The eq in one mix is different to the other. I have my ideas as to which is which. I hear one of them as being a little more robust (slightly firmer sounding) and the other being maybe a tad different in the top end and maybe softer (transient wise not level) I want to do some more listening too before I commit to which I think is digital etc and also which one I like. I do like them both. There is not a huge difference between the two. For me anyway, not spending large sums of money on summing boxes. You did well to build it for the cost.

So which one is which?!?

I must say I like the mix Skull well done. The more I listen to it the more I like it. Nice top end especially after the 2.65KHz stuff is tamed a bit. For me repeated sampling has not gravitated me towards one version but in fact I have switched completely. I heard the differences early on.

Here is my take on it anyway.  Remember the differences are small and require repeated listenings to confirm. I think Wav 1 is the digital file. It is the better more robust mix. Nicer separation, acoustic guitars sound slightly different. Highs are nice in this. There is also a giveaway in that there is hum in Wav 2 (that is not present in Wav 1) right at the end if you really crank your monitors here you will hear it. Hum picked up in the analog signal chain. You might have picked up some tiny hum on the way. With rock music it would not be an issue but with very quiet things it might reveal itself. (Any power transformers or wall wart power supplies near your mic pres or those exposed cables, parts of your busses are exposed inside the box. Is the box grounded, shielded, steel. etc)

Wav 2 is the brighter one and you can be attracted to this. I put a very mild shelf EQ over Wav 2 to bring it into line with Wav 1. If you balance the top end you can start hearing tiny differences a bit more easily. I think the transformers are degrading the transient quality of the music to a very tiny extent. This is a good reason why you do not have to send a great digital mix out to some analog outboard thinking you are going to get something better in return. It might give you this but it might take that. Unless it is a GML equaliser or something!

I still believe the passive resistors are not damaging the transients. The preamps are. Maybe a transformerless class A military grade chip that Rupert uses a lot in his designs or a discrete components high quality class A amp. With a transformerless preamp the difference is only going to get harder to hear. And dont get me wrong about the preamp. A nice transformer coupled mic pre like yours connected to the right microphone and only passing one instrument is probably just the ticket. But not necessarily on a whole mix.

So to preserve any mix in the analog world requires a fair bit of expense applied to the external devices in order to maintain integrity. If you are passing your valuable mix through any analog external devices you need to be pretty confident it is not damaging or altering the integrity of the mix in any way.

One of the things I learned about AB testing is that different musical genres can make the distinctions much easier to hear. Rock music like this can sometimes make it a little harder to hear but there is always things to be heard with any music. Other things like very well recorded jazz (piano bass drums, listen to Brad Mehldau to see what I mean) or even world music with lots of softer percussive instruments being played live coupled up with electronic parts that are also percussive in nature. Also incredible orchestral playing can really give some circuits a hard time and the sound coming out the other end is just not the same as what is going in.

It is good to create these tracks at a K reference of K-20  and stay there with no further mastering. That will also bring out some differences more so as well. The Skull tracks are much louder than K -20. (Skull's tracks are 5 db louder than the K -14 ref level which means they are averaging K -9db which is pretty loud. Not fully mastered necessarily but not all that great for transient testing. It means there is only an average of 9 db for transients. 14 db transients sound a little better and 20 db transients better still. With taller transients the transformer signal path will be exposed a little more. But even with 9db of transients there is plenty of transient energy in the mixes especially in Wav 1.

I did all my testing at 85 db SPL but did do one session at over 105 db and came to the same conclusion.

This would make a good preamp for makeup gain.

http://www.mil-media.com/hv-3c.html

Transformerless and pure. I see Sweetwater has them for $1800. This is where the expense would be in building summing boxes.

Ok...which file was which?

SUM 1 - Summed via Sonar's engine.
SUM 2 - Summed through the analog chain with Daking as makeup gain.

The differences I hear in the two are very, very subtle, but I'll continue using the box since It's all wired in at this point.  And I'll put off my final decision after I get a few months of experience with it.  I've heard enough to give it a little more time to see if I really prefer one over the other.

Thanks Skull.

Very interesting stuff.

best regards,
mike