Proper stereo panning in Reason: RE or VST?

[chaosroyale]

OK! I finally remembered Thor and made a combinator that does this using only stock devices!

-No need for any 3rd party rack extensions or CV utilities.
-It also works as a very nifty autopanner with a couple of cool features.

Tomorrow I will make a nice combinator skin for it and release it as a free Chaos FX.

[integerpoet]

Its been a long while since I even saw an old SSL but if I recall they had mono inputs only, and you had to use 2 channels for stereo sources.

Let's assume your memory is perfect until someone says the actual desk in their studio proves otherwise. So what the original post describes as what one must do to obtain a "proper" pan sounds to me like… if not actual emulation then some historical consistency. In other words, one must provide a truly mono signal (or close the width to effect a mono mix-down). I mean, obviously today a lot of sources are stereo, which means there are extra steps compared to mono sources if what one wants is a "proper" pan. But it seems as if the complaint here is ultimately that RS took an approach to adding stereo input support to the SSL model which some people don't like.

I should add: the Reason mixer looks like an SSL (and has some of the limitations and disadvantages of an SSL) but it is not an SSL, and is not a direct "emulation" of any kind.

Well, yeah, if that were the case I would expect to have seen a lot more marketing about faithful simulation of analog circuits complete with components with unique manufacturing imperfections and different behavior as they heat up, blah blah blah. What I was really trying to suss out is whether a design intended to make people with SSL experience feel at home had implications which annoy people who want what they want regardless of history. It now seems to me likely that is not the case because (per your memory) a real SSL desk didn't have stereo inputs in the first place, so we can't lay the problem at the feet of oldsters.

[guitfnky]

Its been a long while since I even saw an old SSL but if I recall they had mono inputs only, and you had to use 2 channels for stereo sources.

Let's assume your memory is perfect until someone says the actual desk in their studio proves otherwise. So what the original post describes as what one must do to obtain a "proper" pan sounds to me like… if not actual emulation then some historical consistency. In other words, one must provide a truly mono signal (or close the width to effect a mono mix-down). I mean, obviously today a lot of sources are stereo, which means there are extra steps compared to mono sources if what one wants is a "proper" pan. But it seems as if the complaint here is ultimately that RS took an approach to adding stereo input support to the SSL model which some people don't like.

I should add: the Reason mixer looks like an SSL (and has some of the limitations and disadvantages of an SSL) but it is not an SSL, and is not a direct "emulation" of any kind.

Well, yeah, if that were the case I would expect to have seen a lot more marketing about faithful simulation of analog circuits complete with components with unique manufacturing imperfections and different behavior as they heat up, blah blah blah. What I was really trying to suss out is whether a design intended to make people with SSL experience feel at home had implications which annoy people who want what they want regardless of history. It now seems to me likely that is not the case because (per your memory) a real SSL desk didn't have stereo inputs in the first place.

it’s been a long time since I’ve been inside a pro studio, but this was how most analog consoles worked, as I recall. perhaps SSLs are built differently, but when I interned at a studio (they had a Sony desk), there wasn’t really such a thing as stereo tracks. every track handled an individual signal, and if you wanted a stereo spread, you panned one one way, and the other the opposite way.

[chaosroyale]

Here's the free combinator: "Pandemon"
Uses stock devices only (requires Pulsar LFO, so Reason 9 and up, I guess?) no extra utilities or CV stuff required!

What does it do?
Insert in the rack for easy one-knob true stereo panning. Try it before and after reverbs and delays.
Includes bass mono-izer and master stereo width, so you can adjust that directly in the rack.
Has a tempo-syncable autopanner with a variety of LFO shapes, which autopans around wherever you set the pan, with adjustable depth.
Has a creative feature to pan / autopan only the mid & hi end, leaving the low mid & bass alone; try it on pads.

Free to use, not for resale etc etc.
I'll put this in the Refills/sounds board later, but seeing as this discussion started in this thread, here it is for anyone who wants it.

[selig]

For those with Selig Gain, here's a version I made to demonstrate how I feel panning should work in Reason (basically what we've been talking about here), utilizing Pan and Width and including alter pan law selection and L/R swap for giggles.
https://www.dropbox.com/s/11wltcjmnkqhc ... b.zip?dl=0

[huggermugger]

I have no problem with the method Reason uses to place a stereo image in the total stereo field. It's as precise as it needs to be, given that the acoustics anywhere except a perfectly tuned room will mess up the imaging beyond any ambiguities that could be introduced by manipulating Width (very useful) and the Center point of the pan dial. ProTools and Logic (DAWs that I'm familiar with) each handle this scenario in their own unique ways. All of them give me imaging control beyond a simple full wide stereo or mono centered sound. I don't need more than that. Not sure who would.

[kbard]

For those with Selig Gain, here's a version I made to demonstrate how I feel panning should work in Reason (basically what we've been talking about here), utilizing Pan and Width and including alter pan law selection and L/R swap for giggles.
https://www.dropbox.com/s/11wltcjmnkqhc ... b.zip?dl=0

FYI Width is working but Pan knob doesn't do anything. Moving left or right doesn't have any effect.

[selig]

For those with Selig Gain, here's a version I made to demonstrate how I feel panning should work in Reason (basically what we've been talking about here), utilizing Pan and Width and including alter pan law selection and L/R swap for giggles.
https://www.dropbox.com/s/amdztdtac1drr ... b.zip?dl=0

FYI Width is working but Pan knob doesn't do anything. Moving left or right doesn't have any effect.

Sorry about that, forgot to reconnect the output of the mergers to CV 1 and 2 on the Combinator.
Here's the link to the corrected version, or you can just connect each merger output to CV input 1 and 2:
https://www.dropbox.com/s/11wltcjmnkqhc ... b.zip?dl=0

Compare it to the SSL "panner" with a drum kit with the high hat obviously panned (or create a simple pattern in ReDrum that does the same). Also listen to how the level changes, or does not change, with each panner!

[kbard]

FYI Width is working but Pan knob doesn't do anything. Moving left or right doesn't have any effect.

Sorry about that, forgot to reconnect the output of the mergers to CV 1 and 2 on the Combinator.
Here's the link to the corrected version, or you can just connect each merger output to CV input 1 and 2:
https://www.dropbox.com/s/11wltcjmnkqhc ... b.zip?dl=0

Compare it to the SSL "panner" with a drum kit with the high hat obviously panned (or create a simple pattern in ReDrum that does the same). Also listen to how the level changes, or does not change, with each panner!

Yes your panner is EXACTLY how I would expect a panorama sweep to work and how levels should behave when you move sound in left or right. I used -4.5 pan LAW and it kinda put a smile on my face because that's how I remember hardware SSL 9K console panner works.

This is great. With your combi I really feel more confident about panorama. I wonder why they did Reason panner like it is now?

Anyway I am finding few interesting things with your panner combinator.

1. When PAN is sent to Center (combinator knob) CV 1 and CV 2 VCA modes aren't the same in levels. CV1 is -23.9db while CV 2 is -24.3db even though in the combinator programmer min and max values are the vary same. For some reason combinator refuses to align db to the same values??

This (or the pan law value ) is resulting in slightly different output and a pan shift when you run instrument through your combinator. Non issue but it was weird to me at the beginning because I was thinking I sm imagining it. I noticed it and I was thinking that I am crazy there but then I compared it to a phase reverted copy and it didn't null. Again a non issue (for me).

2. It would be absolutely awesome to have some sort of upgrade in your Gain RE sometimes in the future. Where with switch one could switch between dual panner (current) and a single panner which would output same behavior your combinator is doing now. Perhaps some width control as well which would also appear next to single PAN knob upon mentioned switch.

It's not a feature request I know and I am aware that audio developers are having a hard time these days, everyone request anything and people usually think it's a 1 hour job. I am aware it isn't. I am just contemplating this.

I will simply use your combinator for my production from now on, and eventually if something like this arrives I'll just be happy user.

Regards

[kbard]

Here's the free combinator: "Pandemon"
Uses stock devices only (requires Pulsar LFO, so Reason 9 and up, I guess?) no extra utilities or CV stuff required!

What does it do?
Insert in the rack for easy one-knob true stereo panning. Try it before and after reverbs and delays.
Includes bass mono-izer and master stereo width, so you can adjust that directly in the rack.
Has a tempo-syncable autopanner with a variety of LFO shapes, which autopans around wherever you set the pan, with adjustable depth.
Has a creative feature to pan / autopan only the mid & hi end, leaving the low mid & bass alone; try it on pads.

Free to use, not for resale etc etc.
I'll put this in the Refills/sounds board later, but seeing as this discussion started in this thread, here it is for anyone who wants it.

I tried your panner. It's great but as reported with Selig panner (which is corrected now), not sure did anyone noticed it but tempo-syncable autopanner does not work for me at all. Everything else works fine.

I really liked your background

[chaosroyale]

did you see the label on the mod wheel?

I tried your panner. It's great but as reported with Selig panner (which is corrected now), not sure did anyone noticed it but tempo-syncable autopanner does not work for me at all. Everything else works fine.

I really liked your background

[kbard]

did you see the label on the mod wheel?

Sweet lord no I never expect this control there in FX unit. Sorry about that. Perfect!!!

[huggermugger]

It's not that easy, due to the limitations of the combinator - you can't set a knob to only affect a parameter from halfway through its rotation onwards. I've tried to design elaborate workarounds in the past with CV devices but it doesn't really work.

You can define the range of Rotaries, Dials and even Buttons in the Programmer.

[selig]

Yes your panner is EXACTLY how I would expect a panorama sweep to work and how levels should behave when you move sound in left or right. I used -4.5 pan LAW and it kinda put a smile on my face because that's how I remember hardware SSL 9K console panner works.

This is great. With your combi I really feel more confident about panorama. I wonder why they did Reason panner like it is now?

Anyway I am finding few interesting things with your panner combinator.

1. When PAN is sent to Center (combinator knob) CV 1 and CV 2 VCA modes aren't the same in levels. CV1 is -23.9db while CV 2 is -24.3db even though in the combinator programmer min and max values are the vary same. For some reason combinator refuses to align db to the same values??

This (or the pan law value ) is resulting in slightly different output and a pan shift when you run instrument through your combinator. Non issue but it was weird to me at the beginning because I was thinking I sm imagining it. I noticed it and I was thinking that I am crazy there but then I compared it to a phase reverted copy and it didn't null. Again a non issue (for me).

2. It would be absolutely awesome to have some sort of upgrade in your Gain RE sometimes in the future. Where with switch one could switch between dual panner (current) and a single panner which would output same behavior your combinator is doing now. Perhaps some width control as well which would also appear next to single PAN knob upon mentioned switch.

It's not a feature request I know and I am aware that audio developers are having a hard time these days, everyone request anything and people usually think it's a 1 hour job. I am aware it isn't. I am just contemplating this.

I will simply use your combinator for my production from now on, and eventually if something like this arrives I'll just be happy user.

Regards

#1: welcome to "how to find the center of a knob with an even number of values". Is the center of 0-127 63? or is it 64? Neither! It's 63.5, which you can't "hit" on a knob that only supports integer steps! Thus, using the technique I choose imposed this limitation on the outcome which is fine for a "proof of concept" device but obviously not for a final release device! The Thor solution to keeping a control from moving through half of a knob's range may have been the better, although far more complex, solution.
#2: I will add this to my "to do" list for any future updates, excellent idea!

I'll give the Thor angle another look in the mean time…

[selig]

It's not that easy, due to the limitations of the combinator - you can't set a knob to only affect a parameter from halfway through its rotation onwards. I've tried to design elaborate workarounds in the past with CV devices but it doesn't really work.

You can define the range of Rotaries, Dials and even Buttons in the Programmer.

Yes, this range is "matched" by the range of the front panel knob. That means 100% of the panel knob will be assigned to control less than 100% of a destination knob/slider - BUT, unfortunately not the other way around which is what we want here.

To clarify, we want the front panel knob to NOT control the destination through 100% of its travel. We want the front panel knob to only affect ONE of two parameters at a time, one for the left 50% of its travel, and another for the other 50% of its travel. This is the "tricky" bit… Make sense?

In the past folks have used Thor and a Combinator to do this, I tried to use the CV feature on Selig Gain to overcome this limitation but with some inaccuracies (and to easily allow pan law to be included in the Combinator). I'll try a version with Thor just to see if I can clean things up a bit…

[kbard]

#2: I will add this to my "to do" list for any future updates, excellent idea!

Someone here in my room is dancing

work-it.gif (247.43 KiB) Viewed 1259 times

[selig]

https://www.dropbox.com/s/lbke0y5qozqc0 ... b.zip?dl=0

In the mean time, Thor to the rescue once again. Here's an improved Combinator made by going back to my Thor CV library and finding the work I did on building a Threshold Switch, which led to building a knob that moves one control during it's first half, and a different control during it's second half of travel. Applications? First was a LP and HP filter on one knob - to the left you LP gradually reducing the high frequencies, center is flat, and to the right you HP gradually reducing the low frequencies (very natural sounding when you move the knob across the full range). Second application was this panner. Such a control is also useful for stuff like attenuverters, alternating sends (send one to the left, send two to the right, off in center), bass/treble (boost bass to the left, treble to the right), etc.

[Note: The same issue from the first Combinator is still essentially still haunting us even with this approach, it's just that it is easier to "fix" it because of Thor. I had to create a small signal to push the control all the way to the side, but that signal only needed to be that strong at the extreme - too strong overall and you mess up the center position when fixing the edge position. The solution was to create an exponential curve so it only reach it's higher levels at the extreme end which is exactly where we needed it. See Slot 3 of the Mod Matrix for the offset tweak: the exponential curve is created by scaling (multiplying) the control by itself. The end result is the new "curve" of the parameter is now lower in the center but still starts/ends at the same min/max level, and since the center is lower it no longer has enough level to affect the center pan. If you're curios what that "fixes", set the amount to zero and the main Pan control in the Combinator to the far right/max. Now look inside the Combinator at the "Panning" Selig Gain's left Pan knob and notice it's at 94 now and not 100. Notice that in Slot 9 Amount 2 you see "99" and not 100. Setting it to 99 moves our "problem" from the center as it was in the previous Combinator to the side which is easier to fix. Set this to 100 and put the Pan knob in the center and you'll see the familiar problem from before (Left Pan -98, Right Pan 100). Hope this is at all interesting to someone… ]

[kbard]

This is some really good stuff

[dannyF]

This is interesting. But what is the actual difference between real "panning" vs "balance" surely for all intents and purposes the same thing is accomplished in the final audio ?

If not, what IS the actual difference in the final audio ?

[Ottostrom]

This is interesting. But what is the actual difference between real "panning" vs "balance" surely for all intents and purposes the same thing is accomplished in the final audio ?

If not, what IS the actual difference in the final audio ?

The essential difference is that with balance you are turning down the volume of either the left or the right channel. Which means that if your stereo track has information only present on the left channel for example and you turn the balance knob all the way right, that information (or audio) is not shifted to the right, it is only muted.

With panning it actually moves the signal so it is present on the right side.

[dannyF]

The essential difference is that with balance you are turning down the volume of either the left or the right channel. Which means that if your stereo track has information only present on the left channel for example and you turn the balance knob all the way right, that information (or audio) is not shifted to the right, it is only muted.

With panning it actually moves the signal so it is present on the right side.

Thanks. I mostly understand this.

Aside from the technical production differences what difference is there in the actual listening experience?

[jam-s]

Aside from the technical production differences what difference is there in the actual listening experience?

If you do a proper mix (taking into account pan/balance): None.

[selig]

This is interesting. But what is the actual difference between real "panning" vs "balance" surely for all intents and purposes the same thing is accomplished in the final audio ?

If not, what IS the actual difference in the final audio ?

With Reason (balance control) panning a stereo signal to the left just turns down the right channel. Hard panned signals to the left are JUST the left channel with the right channel totally silent (and vice versa).
Panning as has been built by myself and others here is different - panning a stereo signal instead actually PANS each individual channel, so panning to the left pans the right signal to the left, while hard panned signals contain both left and right signals and are mono by definition.

As for accomplishing the same thing, you can pan using two faders or you can pan with a pan knob - sometimes it's about how quickly and easily you can do things. With a balance control you can still get both signals panned hard to one side, you just have to use two controls to do so: panning, and width.

Here is a real world example
When I have two similar stereo sources, many times I like to pan one from Left to Center and the other from Center to Right. Let's just deal with one stereo channel, and you want to pan it from Left to Center.

With a "two panner" system like Selig Gain, that's super easy, just leave the left panner where it is and pan the right panner to the center. With the Combinator I built it is also super easy, just move pan 50% left as one would expect. How do you equal this panning in Reason? One would think you set width to 50% and pan 50% Left (or right). The graphics around the pan knob even suggest this should work - but not quite, it seems. Luckily it's easy to check because if done correctly then one channel will be perfectly in the center and the other will be hard to one side. But you can easily see that in Reason these settings do not quite get the right signal to the center, nor (more surprisingly) does it leave the left signal 100% in the left!

Try this if curious: use ReDrum (with any drum sounds) and pan kick center, one hat left, another hat (or snare) right. The point is we need three signals mixed to stereo; one center, one hard left, one hard right. Now we can test any panning system to see how they compare. When I compare Selig Gains two panner system to the Combinator I built, I can get these to null doing the above panning. But the SSL panner in Reason isn't as straight forward, and you can check this by playing only the drum sound panned hard left and hard right, one at a time, to see where they actually end up being panned.

When I play what should be the center signal (the signal that is originally sent to the right channel by ReDrum) I see it's not center, but moving the knobs around I can get it pretty close. You'd think that was the difficult channel, that whatever is panned hard to one side stays there when you pan FURTHER to that side - but that's not quite the case either. Play the sound panned hard left and try to adjust the width and pan so there is NO signal in the right channel. Mute the left channel to be sure it's gone - not so easy, right? With width at 50% and Pan at 50% left, the left channel should still be 100% in the left but it's not. I'm seeing the left signal still around -13 dBFS and the right channel is reading almost -29 dBFS where there should be total silence. I'm also seeing a 2 dB difference between the channels when it should be centered (and thus both channels SHOULD be equal).

Why is this happening, I though a balance control only turned levels up/down? Well, once you adjust the width control, Pan is no longer just a balance control. At full mono (width fully down/min), pan is indeed a true panner. And at 100% width it's a true balance control. But what's happening in between? I can't figure it out, and It just doesn't work quite like the panel graphics (LED ring around the pan knob) indicate.

This is but one example of how there are things you cannot do with the Reason approach to panning, but which you could easily have done if you choose another algorithm while still keeping the Pan/Width control set (OR using separate L/R pan controls). So it's not the Pan/Width approach that's causing the issues, it's the way it was chosen to operate. And I can say it was initially quite confusing when first working with the SSL pan and trying to do the things I was used to doing in every other mixer (including "real" SSL mixers).

Here's another example - what if I wanted to pan the left and right panner half way to center to basically collapse the stereo image by half? Simple to do on the SSL, right, just set width to 50% - right? Right???
Sorry, no. To get the equivalent setting on the SSL you have to set the width to 1/4 or 32 to equal individual panner half way between left/right and center. Odd, no? You'd think that half width would be 50% on the Width control, right? It's not pan law either, I'm using a 3 dB pan law in Selig Gain to match the 3 dB pan law in Reason, plus I have to turn down the Selig Gain version to match levels because Reason uses +3 dB center and Selig Gain uses the more traditional - 3 dB center - But even after that, you'd think things would be closer. Turns out when Width is set to 50% it is equal to Selig Gain Pan set to 80% to one side (not 50% as one would expect).

SO - the difference in the final audio is enough to be heard in most cases but obviously has not impacted the quality of work coming out of Reason - In either case you mix to what you hear (just like when using different pan laws). But still, there ARE some things impossible to replicate in Reason compared to traditional dual panner OR to modern "stereo" panner. Still not sure why they didn't just go with a true stereo panner/width control set instead of the balance control approach, but it's yet another Reason why I built Selig Gain for Reason…

[selig]

Aside from the technical production differences what difference is there in the actual listening experience?

If you do a proper mix (taking into account pan/balance): None.

Mostly…

[dannyF]

If you do a proper mix (taking into account pan/balance): None.

what does this mean?

oh nevermind I just saw Selig's post.