joel77
Thanks for starting a great, informative thread, Ron!

 Garsh.

Sanderxpander
It's important to realize that many hardware synths output a lot of "fake stereo". Like if you record some brass patch out of a synthesizer, 99 percent chance that they didn't actually sample the trumpet in stereo (why would they) but just slapped on some chorus and output it as a stereo signal. Same goes for most analog synths and their VST versions - often it's the effects that make them stereo. There's nothing inherently stereo about a sawtooth wave and apart from the MiniMoog Voyager (stereo filter section only) I don't know of any analog synth that has a stereo signal path (meaning separate oscillators, amps, filters for the left and right sides), though you could pan the oscillators on an OBX I believe.

In these cases, adding those as a stereo signal to your mix will usually only make it harder to mix properly (especially with the fake chorus stereo of ROMpler synths) and it can be helpful to record them in mono. So I agree - think about your source, is it really stereo, in effect does it make sense to have a different left and right side for that sound, and then go stereo. Otherwise stick to mono where possible. You can always reintroduce fake chorus stereo through a VST effect if you feel a particular sound needs to be bigger/wider.

Yes and No. Some synths such as my Kurzweil sample heaps of stuff in stereo and the stereo effects are breathtaking and can do stuff no external effects can as well. (VAST) Especially organ Leslie effects and stuff. It is all integrated and very stereo.
 
This is bad advice. Simply record everything from any hardware synth in stereo (as long as it has two outputs that is!) and use Channel Tools to bring it down to mono if you need to. Also sometimes even the chorus inside something like the Juno 106 is very unique and you will simply never create it anywhere else.
 
Also some analog synths (VST's) have got amazing stereo effects. My Oberheim OPX Pro Mk II which is a copy of the fabulous OBX has panning for every oscillator. Try doing that outside in your DAW. Cannot be done. You have to hear this to believe it.  Also EMU Emulators have the amazing ability to modulate the panning of every single sample within a patch. Yes every voice, its own LFO speed, centre position, depth of panning etc you name it. You also have to hear that to believe it too.  (dense complex ambient texture for example) It is quite something and something you could NEVER create in any DAW after the fact. You would be silly to record all that in mono and just miss out on the effect.
 
Korg Wavestation is also ridiculous in stereo and you could never create the effects that are going on inside that machine as well. I keep thinking of others too. JD990 comes to mind. All stereo samples, 4 tones per note running through stereo filters and effects. Stuff that is just so complex stereo wise it is incredible.
 
The thing is that yes some synths are like what Sander has described and others are well and truly not. So unless you are a full synth expert and know the internal architecture incredibly well of every synth and I imagine most would not then best to record in stereo and deal with it after the fact.
 
I teach sound engineering as well and my students are always forgetting to record hardware synths in stereo. Then during the mix we have this total bland boring mono sound. And then they try to recreate the stereo effects and it just never works. I am always telling them  "hey I know that instrument and it sounds killer in stereo and you should have used two DI's on that instead of just the one".
 
There are some techniques for blending a whole lot of very wide stereo synths though and having them all feed hard L and hard R is not the way to go but that is a different thing. You still have to have the stereo tracks to start with.

So, to sum up:
1) when recording using a single mike, set the track's input to either Left or Right to record in mono,
2) when recording a DI guitar or bass, same thing: set the track's input to either Left or Right to record in mono, and
3) when recording any type of synthesizer, especially one that outputs stereo effects (such as the Leslie effect Jeff mentioned) set the input to stereo.
 
If there are other situations I'm neglected to mention, please let me know.

Jeff Evans

Sanderxpander
It's important to realize that many hardware synths output a lot of "fake stereo". Like if you record some brass patch out of a synthesizer, 99 percent chance that they didn't actually sample the trumpet in stereo (why would they) but just slapped on some chorus and output it as a stereo signal. Same goes for most analog synths and their VST versions - often it's the effects that make them stereo. There's nothing inherently stereo about a sawtooth wave and apart from the MiniMoog Voyager (stereo filter section only) I don't know of any analog synth that has a stereo signal path (meaning separate oscillators, amps, filters for the left and right sides), though you could pan the oscillators on an OBX I believe.

In these cases, adding those as a stereo signal to your mix will usually only make it harder to mix properly (especially with the fake chorus stereo of ROMpler synths) and it can be helpful to record them in mono. So I agree - think about your source, is it really stereo, in effect does it make sense to have a different left and right side for that sound, and then go stereo. Otherwise stick to mono where possible. You can always reintroduce fake chorus stereo through a VST effect if you feel a particular sound needs to be bigger/wider.

Yes and No. Some synths such as my Kurzweil sample heaps of stuff in stereo and the stereo effects are breathtaking and can do stuff no external effects can as well. (VAST) Especially organ Leslie effects and stuff. It is all integrated and very stereo.
 
This is bad advice. Simply record everything from any hardware synth in stereo (as long as it has two outputs that is!) and use Channel Tools to bring it down to mono if you need to. Also sometimes even the chorus inside something like the Juno 106 is very unique and you will simply never create it anywhere else.
 
Also some analog synths (VST's) have got amazing stereo effects. My Oberheim OPX Pro Mk II which is a copy of the fabulous OBX has panning for every oscillator. Try doing that outside in your DAW. Cannot be done. You have to hear this to believe it.  Also EMU Emulators have the amazing ability to modulate the panning of every single sample within a patch. Yes every voice, its own LFO speed, centre position, depth of panning etc you name it. You also have to hear that to believe it too.  (dense complex ambient texture for example) It is quite something and something you could NEVER create in any DAW after the fact. You would be silly to record all that in mono and just miss out on the effect.
 
Korg Wavestation is also ridiculous in stereo and you could never create the effects that are going on inside that machine as well. I keep thinking of others too. JD990 comes to mind. All stereo samples, 4 tones per note running through stereo filters and effects. Stuff that is just so complex stereo wise it is incredible.
 
The thing is that yes some synths are like what Sander has described and others are well and truly not. So unless you are a full synth expert and know the internal architecture incredibly well of every synth and I imagine most would not then best to record in stereo and deal with it after the fact.
 
I teach sound engineering as well and my students are always forgetting to record hardware synths in stereo. Then during the mix we have this total bland boring mono sound. And then they try to recreate the stereo effects and it just never works. I am always telling them  "hey I know that instrument and it sounds killer in stereo and you should have used two DI's on that instead of just the one".
 
There are some techniques for blending a whole lot of very wide stereo synths though and having them all feed hard L and hard R is not the way to go but that is a different thing. You still have to have the stereo tracks to start with.

I actually do consider myself a synth expert, though there's always more to learn. That's where my advice comes from. I specifically mentioned the OBX stereo voice mode and the fx on some synths making them stereo (e.g. the chorus on my JX10 and Alpha Juno 2). I actually own OPX Pro II and Emulator X3 as well. While they CAN do stereo specific stuff, much of the time, they don't. I've never used the stereo filter on my Moog Voyager usefully so far. Mainly the filter spread knob is where I expect the Resonance/Emphasis to be and I grab the wrong one when in a hurry.
Obviously, if there is anything interesting going on with the stereo image, you'll probably want to capture it. But a whole bunch of layered fake stereo sounds just muddle your mix. The fact that a some trumpet sample may sound nice in stereo out of the synth because it has some chorus on it doesn't mean that it will sit well in your mix.
 
I will stick with knowing my source, but perhaps you're right and this doesn't work as general advice because many people don't realize what makes a sound stereo, especially in some random hardware synth. Sorry if I come off a little defensive. My ego felt a little stepped upon because I feel I know what I'm doing.

Actually something worth mentioning is that some synths can sound better with their effects turned off too. The Yamaha SY77 is one such machine. Yamaha drenched every patch in chorus and reverb and it sounds like crap. But bypass all the effects (one button on the front panel called FX Bypass!) and this wonderful beautiful interesting and very stereo sound just moves forward and sounds amazing.
 
The Korg M1 is another machine that often had overdone effects. You can still leave them on but it is good to go into and edit the effects and really bring back the wet/dry mixes to be much more subtle. Tighten up the reverbs and reduce the amount of them. Turn the choruses off and voila you will end up with a patch that sounds way better more often than not. It is not the machine that dates a synth. It is often the use of effects that date it.
 
So it is not a bad idea to audition things too before you record because in some cases some quick editing to the effects will make a big difference. Not so common now though. Most effects are much nicer and more subtle these days I have found in new Kurzweils etc...
 
 

In my opinion that goes for most digital synths. Certainly I'd never record reverbs or EQs from a synth like a Kurz K2600/PC3K, Korg Triton/M3/Kronos, Roland Fantom/JP80, Yamaha Motif etc.
While it's true the fx have gotten better over the years, and I'm actually for the first time happy with my live reverb (O-Verb from the Kronos) they're are definitely still worse than what the VST world has to offer. So unless an effect is an integral part of the sound (like for instance the Polysix chorus on the Kronos or a wah on the Motif), turn it off and be happy with the improvement you can make in the DAW.

As a sidenote, does an SY77 actually do anything stereo on the oscillator or filter level? Maybe a stereo sampled piano or some panning effect? For many synths, turning off the fx makes them effectively dual mono for the largest portion of sounds.

I used to think that way too and not record reverbs and things from the K2000 for example and add it later but over time I have changed that view and do the opposite. The argument for not using an external reverb on something is that in a funny way you are loosing something from the internal effects being off and replacing the reverb from the outside. And yes that does work at times too for sure, these days I am into using and recording internal effects more so. I find when you do this a synth part can sound a little more distinctive and stand out a little more and have a slightly interesting character. It just means you have to go in and edit the FX a little more so before tracking to make sure they are not too over present etc...
 
It is well known that lush expensive reverbs actaully don't sound ideal on everything and sometimes a courser or rougher cheaper sounding reverb can be just the ticket to make a synth stand out a little and sound a little more interesting as a result. You can only have so many lush expensive sounding reverbs in a mix, one usually!
 
BTW modern effects such as inside the new Kurzweils are absolutely stellar and first class. Can easily rival anything outside. Hardware or software. People who don't own them don't know this that is all. There are 16 KDFX processors inside the PC3K now, yes 16! The Yamaha A5000 has got a killer effects unit inside too. So has the top of line EMU samplers such as the E4XT or E5000.
 
I think the SY77 even with its effects off is creating lovely stereo images so yes there must be things going on further back to create the stereo imaging. The SY77 uses AFM and AWM waves to make the patches eg 4 tones per note as well and there are a lot of programming options in terms of panning etc. I think some of the AWM waves may be stereo too.

I'm sorry but I've owned and play many modern current synths and I simply disagree about the fx quality. While I agree that you can sometimes need or want the specific sound of a synth's effect, they simply don't stack up against the cream of the crop VSTs. If you can find me a single piece of hardware that has a reverb as good as AltiVerb, an equalizer like Equilibrium, or a bunch of expertly modeled vintage compressors like Waves Classic Compressors I'll take it all back.
I don't hate all fx from hardware synths and agree that they can add a flavor. Certainly the E4XT fx expansion is pretty famous. But if you know what you're doing, it's usually a better idea to take them out when you have clearly superior options.

That doesn't include replacing fx on a 1 on 1 basis by the way. That way you'd end up using a whole bunch of different reverbs. Use a buss instead unless you need some extra room effect. Use EQ when needed. But don't leave the Korg EQ with it's 12dB of high boost on in the synth. Etcetera, etcetera. Yes, this requires a lot of knowledge about the synths. But I don't think that's unreasonable when you buy a piece of hardware like that and are proficient and technically savvy enough to do your own recording in a program like Sonar.