I pretty much got the same results.
I couldn't align them perfectly with 1ms grid snap, and Sum2 visually looks a smidge ahead of Sum1.
Maybe if I can turn snap off I can slide it a little more. I was zoomed in as far as I could go but I may be able to get in closer.
I don't know if I can level match that accurately either...I can do 1db increments but maybe there is a way to determine the gain of Sum1 and then change Sum2 in .5db increments.
Maybe you can run them again and see if you can match them?

I have a feeling if we performed this over and over they would be forced to evenually match.

What does this tell us all?

That digital is digital? LOL

Ooooh...1000 posts...yeah for me!

Congrats!

Knowing this result, I would have two answers now. Prior to the second generation NULL test, SUM 2 is more acurate to my results and the best sounding. After the Second generation NULL test they both sound close enough that it does not matter. Digital has an effect.

This is very interesting don't you think?

I'm not sure....
What effect do you think it is imparting on the signals.
I need to understand what you did...

1.) You said you recorded two mics using your RME converters which created two samples into Sequoia. 24/44.1

2.) You took those two tracks now in Sequoia (ITB) and bused them to a stereo output right?
And then you bounced that and exported that to a stereo wav file....

3.) You took the same two tracks from Sequoia and output them on 2 separate tracks into an analog summing unit and then converted that again with the RME back into Sequoia and then did a bounce/export to a second stereo wav file.

Is that correct.?

yes

And also NULL them both prior to bouncing them to a wave file. This is what you hear online now, which is the first generation wave. However, after downloading them and doing a second generation NULL test, I cannot tell the two apart.

djmukilteo, post: 386119 wrote: I'm not sure....
What effect do you think it is imparting on the signals.
I need to understand what you did...

1.) You said you recorded two mics using your RME converters which created two samples into Sequoia. 24/44.1

2.) You took those two tracks now in Sequoia (ITB) and bused them to a stereo output right?
And then you bounced that and exported that to a stereo wav file....

3.) You took the same two tracks from Sequoia and output them on 2 separate tracks into an analog summing unit and then converted that again with the RME back into Sequoia and then did a bounce/export to a second stereo wav file.

Is that correct.?

Explained exactly:)

The idea with analog summing is bypass your master bus until all the tracks have been summed OTB. You use a high headroom analog system to juice up your digital mixes with quality hardware that focuses on the groups transient characters etc. (Example, an LA-2A inserted into the Vocal group.)

In this case we didn't do anything other than record a piano in stereo and called that track finished ( plain and simple). Then took that stereo track and bused it out to the summing system then re-recorded it onto a second stereo track. So now I have 2 stereo tracks of the same performance, one that went through an analog summing amp. Follow?

Then I inverted one of the 2 stereo tracks and got them to NULL best I could, thus only adjusting the analog track to the original so it is as close to level and timing as the original. Through this entire process I did not touch the original track. Once they NULLed as best I could do, I then bounced them both to a wave file each and uploaded them to Dropbox.

The rest is where we are now.

Ok..then that would suggest the analog summing unit isn't doing anything really different than Sequoia?
But that might be because you didn't impart anything with your summing unit and that the summing portion ITB is as good as the analog by not affecting the "dry" signals....it becomes more passive and gets down to the mathematical summing engine in Sequoia. Exactly in and exactly out...and both Sequoia and your summing unit are effectively the same.....or they're not!
Which it seems is hard to tell...there IS some difference between the two and you had 1 additional conversion pass with the summing unit didn't you?
In my mind the digital conversion process shouldn't do anything to the signals that appears at their inputs, it will be exactly as Nyquist dictates and that's the digital part common to both setups.
The only real detectable change would be if you manipulate the signals going through each platform before a significant difference will start to appear in the converted files...
I'm probably wrong here but dry "should" be dry and "sauced up" is a whole other thing...

I chose Sum1... to me it sounded a bit brighter and clearer, which I generally find in opinions is 'not warm'. I do like warm, but I love to hear that clarity too.

It's pretty crazy though, to me it seems that the change added by different gear is SO SUBTLE. But then you stack that subtlety track over track over track and then it makes a big difference!

If I wasn't listening for it, I wouldn't have been able to tell a difference. Actually, I'm still not sure I could lol

djmukilteo, post: 386103 wrote: Went back and did it again and I "increased" Sum2 wav 2db to match Sum1
LOL...sorry about that Bob good call!

That's a little better since I thought Sum 1 might be a bit louder, but 2 dB is much more than I would have guessed. A 2 dB difference in RMS levels will skew just about any listening test.

You'd think, but after I summed the tracks in analog, they sounded better to me. But after I did a couple of generations of Null tests, both tracks are very hard to tell apart.

The good in this, my analog summing system did not change the tracks for the worse. It did not add any discernible noise. In fact, to my ears it opened the tracks up. I find this test very important because I have a system that will allow me to insert other analog hardware to sound design my mixes. I have a pretty equal summing system to ITB.

This all being said, re recording these tracks into a digital system starts to remove the things I find pleasant in the analog domain. Rather than the analog being the culprit, its the digital domain that is changing the analog back to not being able to tell which is which as we are hearing now.

Follow?

I did a +1db gain process, checked the track meters and zoomed in visually on the waveforms and then did another +1db process.
Of course I don't know how accurate that was with Cubase and the meters, but they matched bang on with peak and hold bars as an indicator after doing the gain and timing. That was how I discovered they were different to begin with...they bounced at different times and the peak and hold were at different levels....so they were never really that close enough to sum to zero.
Isn't there a plugin like SPAN or something that can be used to check this stuff?
I'm not a mastering guru at all but I thought there were tools that get into 1/10th db or something and you can see the exact differences.
There was a difference....no question especially if your talking about the files as downloaded...actual listening...very subtle...I guess the null test proved that...because without mucking around with Sum2 there was no nullin' going on and so the test may still be skewed by the two different files whatever the cause..
I'm starting to question if you can make any real conclusions....good or bad...

audiokid, post: 386127 wrote:
This all being said, re recording these tracks into a digital system starts to remove the things I find pleasant in the analog domain. Rather than the analog being the culprit, its the digital domain that is changing the analog back to not being able to tell which is which as we are hearing now.

Follow?

I think I have to disagree with that....when I did a mixdown using the ZEDR16 and bounced/printed that back into Cubase and then compared that to the same straight ITB summation I had originally everything that the analog had imparted through the ZED mix was there 100%.
It didn't revert my open hot warm mixdown back to the sterile bland ITB mix that was in there originally...
After that experience I will never leave a mix in the box.

Well, it's no surprise that they don't null exactly. There are two different signal chains after all. On the other hand, we aren't talking about 8 or 16 stems here. This should be the smallest difference we would expect between hybrid and itb. It's crucial to get the rms levels as close as possible to do a valid test.

Um, if the stereo recording stayed fully panned throughout the process, what summing actually occurred?

bouldersound, post: 386134 wrote: Um, if the stereo recording stayed fully panned throughout the process, what summing actually occurred?

audiokid said he took two mics X/Y and recorded those two mono signals to a stereo track, 1 mic L and 1 R.
I don't know if he said he panned them though....maybe
The wav files that were posted were a single waveform, not stereo....so I assumed that single waveform were the two tracks summed together.....but I don't know and I'm confused now...LOL

What says it isn't the second pass through the converters that's responsible for the difference?

djmukilteo, post: 386135 wrote: audiokid said he took two mics X/Y and recorded those two mono signals to a stereo track, 1 mic L and 1 R.
I don't know if he said he panned them though....maybe
The wav files that were posted were a single waveform, not stereo....so I assumed that single waveform were the two tracks summed together.....but I don't know and I'm confused now...LOL

Soundcloud always shows a single waveform. The [[url=http://[/URL]="http://bouldersound…"]waves[/]="http://bouldersound…"]waves[/] from Dropbox are stereo and there is an audible difference between left and right.

bouldersound, post: 386139 wrote: What says it isn't the second pass through the converters that's responsible for the difference?

This was something I wondered as well, and according to audiokid that was the procedure..
but we need more detail from him.

This is why you have to beware of "studies", because the people who run them are generally trying to prove something and unconsciously build their biases into the protocol.

Well this is "Track Talk" and the first installment of said segment so it wouldn't really be a good start without some controversy...LOL
These always seem to turn out this way!
I think it's because it's "audio" and even though there is a perfectly rational logical explanation for anything technical like this, we continue to allow human involvement to somehow form opinions or decide by some majority.
So it ineveitably ends up with a he said she said or he said he said or vice versa and no one actually just does the technical analysis and calls it good.
Science has rules for this...it just needs to be developed, tested and retested to discover the facts.

I'm not aiming to cause controversy or disprove anything in particular, just giving the experiment opportunities to show itself as sound or not.

Don, the tracks are stereo, not mono. I don't know about Cubase but in Sequoia, you can record mono or stereo. These are stereo tracks. The wave files are not mono. You must know this if you imported them. They should import as stereo tracks, yes? One file is a stereo track.

To the skeptics :)

Whether this turns into a bunch of technical jibber, thats up to you. I didn't do this to convert anyone here or to try and educate anyone either. I did this for myself and the few that chimed in here answered it the way I heard it, Thank you! I needed conformation from other ears.

I am very happy with the results, which are, this analog summing system is cleaner than expected! thumb

Also, I have past a 32 track session through it and no more noise than we hear with this example occurred. The analog summing system is as clean as the DAW/ close enough indeed. This isn't like the old days. This is ultra clean analog but it ain't sterile like ITB.
I knew it was quiet but the big question was how it would sum up to ITB. Would we notice a loss after going in and out of an analog summing amp. This was what the poll was all about.

For those who think I'm trying to make myself like OTB, well you are sadly mistaken and its wishful thinking if you really believe that. I do not care one way or the other. In fact, I would rather sell my analog studio and go on a holiday but so far, when you start adding hardware into the mix, it sounds way better to my ears than ITB and most plug-ins.

I'm not like most people you'll meet. I don't care one way or the other about material things and can switch on a dime when I learn something new. But I research things to death because I don't like making foolish mistakes either. The moment it proves to be inferior to ITB, I will dump it and go fishing. My boat is waiting for me.
Plug-ins are where it all changes things from this point on.

Thanks for your support on this. Don, you were extra right on!

Cheers!

djmukilteo, post: 386135 wrote: audiokid said he took two mics X/Y and recorded those two mono signals to a stereo track, 1 mic L and 1 R.
I don't know if he said he panned them though....maybe
The wav files that were posted were a single waveform, not stereo....so I assumed that single waveform were the two tracks summed together.....but I don't know and I'm confused now...LOL

In Sequoia, two mics recorded into a single stereo track are automatically panned hard left and right. I never try and make either track perfect in volume either. I set each side close and leave it. I would be impossible to try and keep them exactly the same volume. Its a stereo recording, live. Thats what those track were, stereo and live. They were not mono tracks. Did you guys think these were mono tracks?

djmukilteo, post: 386130 wrote: I think I have to disagree with that....when I did a mixdown using the ZEDR16 and bounced/printed that back into Cubase and then compared that to the same straight ITB summation I had originally everything that the analog had imparted through the ZED mix was there 100%.
It didn't revert my open hot warm mixdown back to the sterile bland ITB mix that was in there originally...
After that experience I will never leave a mix in the box.

I disagree with this after I did this test. But you would have to do the null tests like I did a few times. And you would also have to not do any analog eqing etc. If you were in my studio and saw the process I did, and how it ended up after nulling the two stereo tracks a few times, you would see something did change. The openness of the analog lost something via the null process. It ended up sounding just like the original ITB version.

Try is for yourselves.

bouldersound, post: 386134 wrote: Um, if the stereo recording stayed fully panned throughout the process, what summing actually occurred?

Thats a good question.

I summed the stereo track of the piano in analog, then sent the analog 2-bus back to the DAW, recording it as a new stereo track. What do you call that?
What would you call 8 stereo tracks of piano summed in analog, 2-bused back to the DAW, recorded as a new stereo track? Same thing, only more tracks were summed in analog, yes?

This is what hybrid recording is all about. Only I didn't insert hardware on this one.

bouldersound, post: 386142 wrote: This is why you have to beware of "studies", because the people who run them are generally trying to prove something and unconsciously build their biases into the protocol.

Agree, and so far its all been healthy discussion but this is where I get off the bus. I've had fun and done my best answering all your question and learned from this as well. The two things I walk away with are:

1. its hard to tell the difference between an un processed track summed OTB vs ITB.
2. you can make an analog track sound like a digital track if you null it enough times.

So Tired, long day with our festival.

smoke

audiokid, post: 386150 wrote: Thats a good question.

I summed the stereo track of the piano in analog, then sent the analog 2-bus back to the DAW, recording it as a new stereo track. What do you call that?
What would you call 8 stereo tracks of piano summed in analog, 2-bused back to the DAW, recorded as a new stereo track? Same thing, only more tracks were summed in analog, yes?

This is what hybrid recording is all about. Only I didn't insert hardware on this one.

I think what he is saying is that you didn't actually sum anything: one stereo stem came out of the DAW, went through your analog system, and then came back in as a stereo track. So it's a pure comparison of the two signal chains not the effects of summing multiple stem itb and otb.

I think the valid conclusion is that merely passing the signal through the analog chain and two conversions did not degrade the signal in any audible way and it may have made it sound better to some people. But if the two signals were kept hard panned there was no summing going on and the experiment doesn't address that subject at all. All in all it's been a useful experiment.

So if we took 3 or more tracks and repeated this all, it would be summing then? 2 track do not count? I'll have to try this, and I expect the same results. Which will be, the analog sum will sound different but end up being unable to tell the two apart after a few null processes.

Would anyone care to share some tracks or stems for this experiment?

I guess I was just confused with the graphic representation of the waveform in Cubase.
Yes I can setup either a mono or stereo track in Cubase.
Whenever I setup a stereo audio track and record something "stereo" I get two waveforms one on top of the other within the graphic (L/R).
When I imported your wav files they came in as one waveform, so I assumed it was a "summed" mono track of your two mics tracks....which in my mind would be "summing".

But that doesn't seem to be the case. There was no "summing" going on if you didn't "add" them together onto a single track. It sounds like it was two mono tracks panned stereo.
So I also struggle with the term "summing" as opposed to the term "mixing".
I don't think it matters though because you were testing looking for "noise" between the ITB/OTB and the OTB takes 2 conversions to do that..
Which if you want to determine that, you should do like 10 loopback conversions on a single mono track and then compare or null the first pass with each successive pass....any noise generated will show up by the 10th pass...but ....that could be your converters rather than the analog equipment adding noise each time through....SO...
Then you do the exact same thing ITB and check for noise and then compare the OTB to the ITB to rule out converter noise.
Just thinking out loud here....smoke

I have a couple of stereo tracks that I would like to see what they could sound like if you want me to dropbox them.
It would be interesting to see what your hybrid system can do with them...let me know.

Well that explains all the confusion having to repeat myself so many times! It could be that none of your DAW's are able to do this with the same accuracy as Sequoia, dunno.... Another wrench is added to this quiz now.

FYI, Sequoia will separate a stereo track into their own tracks and pan it like it was summed. I thought all DAW's did this. So you may not have done this null test accurately at all unless you hard panned them and were able to move them independently and leave SUM 1 untoutched?
NOTE: Left and right sides were not the same level between SUM 1 and 2. They were not supposed to be!
These changes occured during the pass through the analog summing system. In order for me to NULL them, I had to do a lot of shifting. But once I lined it all up, the analog sounded like the digital track once again.

I don't know how others are passing analog in and out of your rig but I only do it ONCE or maybe twice but that's another topic! There is no goofy conversion steps and/or unnecessary noise etc added. The MixDream way is superior over using your DAW AUX as inserts. I'm sure 10 loops ITB,OTB,ITB etc would prove to be horrible and absolutely insane, which serves no purpose with this type of system. My way, you experience zero latency and use your summing amp as a HUB and/router.

Okay, so we are getting technical on the definition of sum, 2 tracks sent OTB and back in is not actually summing, or enough to tell whats going on here. Fair enough. I'm game to actually SUM 3 or more tracks if someone provides them this time. But more like 8 or 16 would be fun. I'll see if I can make an analog sum sound like digital.

Back to you Don,

I'd love to experiment with your tracks but be aware, I'm not going to add any processing for now. That's another Poll.

Back to this experiment
I have know idea how this will play out. I will attempt to prove digital summing does what? Please someone finish what you think I am revealing here?

I'll mix X tracks and sum them ITB, bounce to a wave file
Ill stem out ther same mix and sum it analog and post the two tracks.

Then, I will do a few null tests on SUM 1 and Sum 2 and prove they will sound the same after the NULL tests.

EDIT ADDED ,

I'm attempting to prove that a few rounds of digital nulling turns any fine analog mix into a generic ITB sound which could partly be why music sounds like ass today. Just thinking out loud.

Hehe...I have no idea now either...but I'm still interested and game for whatever you want to do!

Me personally...I have no argument your analog gear is better than ITB....I'm pretty sure I believe that.
Now can I can tell the difference....hmmmm.
I suppose because I picked the right sample and you claimed I had good ears....apparently I can.
But I think I missed you stating SUM1 was ITB and SUM2 was OTB from the Dropbox files....correct?
And....just to be clear you switched those around from the Soundcloud versions....correct?
The Souncloud SUM1 was OTB and SUM2 was ITB....correct?

yes

I have 2 tracks that were recorded "stereo" but they are split into 4 separate tracks....so they are each different.
You could pan and sum them anyway you want to a stereo file.
I can export all 4 to wav files and put them in DropBox (if I can figure out Dropbox).
If you think that is enough to "sum", I'll do that...if not no problem.
I would like to hear what your equipment can do....but like you said that's outside this thread!

The only finishing your thoughts would be to define what your looking for first.
Noise is a pretty difficult thing to test because there are all sorts of places that it can be introduced along the line.
So you need to be pretty methodical and rule out each stage of the process to isolate that one object that is or is not generating it.
You should define a step by step procedure that we can all agree on that performs the test without any quirks or ambiguity.
The protocol of the experiment has to be solidly defined and flushed out to remove anything that could skew the results. Someone could object to a certain step or procedure...
It has to be done multiple times (repeatable) to eliminate any random flaw or error to be valid.
The samples have to be presented in a side by side blind fashion with multiple listening sets and statistical analysis of the results need to be done if your going to use human hearing instead of test equipment to arrive at a concrete conclusion.

You will need to open an account with dropbox. If you don't want to do that ( I think it costs something), you can send them to me. I'll email you so you have my addy and we can play around with this outside of RO for fun. I've got a few days before things get really crazy here so its a good time right now.

For this new test we need more tracks from someone willing. NOTE: I'm not looking to identify noise. I did that already. The noise floor is acceptable for me and we all agree its pretty quite. And I will NOTE: 1 track or 100 grouped and stemmed into 16 DA makes no difference because all the tracks coming from the DAW will sound like the DAW in the MixDream up until I add hardware. The MixDream is extremely quiet. The analog chain is within my specs so I'm not selling it yet!

NOTE: Noise increases when I introduce hardware and start playing with tubes, gains etc. From that point its a trade off ( plug-ins vs hardware). I'm not convinced the trade-off for clean clear plug-ins is doing us any favours so I've been on a rampage buying a lot of expensive gear to find more out. So far I think a combination of the right hardware and plug-ins is choice. We don't need a lot of it either, just choice products that plug-ins don't do well at. And this is where hybrid starts to get interesting.

Back to our test and what I am discovering:

I don't think you are with me yet. I have a hell of a time explaining things, I know its me. Words are not my strong point.

From the time I started this thread up to now, I've noticed something common occurs when you NULL SUMMED ITB and OTB Files. Why do we do this NULL experiment in the first place? We know the loudest track always wins so we NULL them close as possible so we can make a judgment call on which sounds best.

Okay, soooooooo

Where it gets interesting now: I've said this a dozen ways already, its becomes near impossible to hear which one is which after the null process. Once the analog track has been SUMMED and re recorded back to the DAW and NULLED a few times to get it as close to SUM 1, something generic happens and both files sound the same. We of course would not do that to our analog mixes for real because it kills the analog flavor. We know this now, or at least I DO!
After doing these tests, I cannot tell the difference between the two. So wTF is digital processing doing to our music?

I'll PM you Don,

Why couldn't you do this with a simple sine wave or tone.
Also how are you going to get around having to convert the OTB file twice.
I would think that has to be taken into account.
And also at what point are you adjusting the OTB to null...are you doing that ITB or OTB?

We've established we need 3 or more mono tracks to do an official sum. It all plays out the same no matter how many tracks, IMO.

Tracks are uploaded.

I'm assuming we are talking about your tracks on your last post yes? I could upload/download and import the same way and repeat it all like I did for mine. But even one NULL process on yours just now was sufficiant. Its hard for me to tell the difference with yours now.

SUM 2 seems to fit a mix better. I mean, The mids are more pleasent. So if we were adding vocals for instant, they would sit in the analog mix with less eq needed (SUM 2). Same goes for my SUM 2 in the poll with the my piano track. This is where it starts making a diffference.