Stinking In Stereo

M

Matthew Walsh

New member
Dear Recording Guys,

You've helped me out of a few audio scrapes in the past, but this one is pretty major, and may involve me having to re-record a good chunk of the music I've already laid down on my 14-song album project.

What I did was, I sequenced all my songs in MIDI, the guitar parts, the drums, the bass, the "standard" keyboards. I then recorded them into Vision DSP on my Mac using stereo tracks. By this, I mean, I'd record the bass, let's say, from my PROteus MPS keyboard, using its regular right and left outputs. They run into my mixer. When I record, I turn the right pan knob on my mixer all the way right, turn the left on all the way left, and record the information into my computer via its stereo input onto one stereo audio track on Vision. I do the same for rhythm guitar, for the kick drum, the snare, etc. -- but I always assumed you recorded a stereo instrument, particularly anything coming out of a stereo keyboard, IN STEREO, with the pan knobs all the way out to capture the full stereo. Right? The idea is to treat all the MIDI instruments like "real" instruments and record and effect them just as if I had a real drum kit, a real bass, etc.

Well, to further compound this, THEN, if something needs processing, I process it in stereo. Of course. Then, I record vocals -- in mono, of course, but I process them in mono-to-stereo reverb/echo, etc. so they sit with the other instruments. Right?

Imagine my surprise when I mix down the song, only to discover that the stereo spread is waaaaaaaay too much when you compare it to virtually anything else that has been professionally recorded in the last, say, eighty years. It sounds horrible and ameteurish and, furthermore, the levels are untenable; I never go over digital zero and yet I get distorted sounding levels, particurly the highs.

So what am I doing wrong? Assume you're speaking to a small, slow child, of which, in the home-recording game, I certainly am. I split the drum machine into Kick Track, Snare Track, Hats Track, Rides Track, Crashes and Toms. Everything else, I record straight from the keyboards, through my Mackie 1202 and into my computer, using the Mackie to boost or cut levels. I don't process anything during recording. I don't boost or cut EQ during recording.

To expediate recording and mixing, I usually mix all my instruments first on the computer. This, I mix down to two tracks (which I normalize in Vision to attenuate anything that might go over digital zero), then, I open a new file, which contains my bounced instruments on a stereo track and my vocals. I mix the vocals to the the stereo instrument tracks. Then, I mix down again to two tracks, which now contain vocals and instruments (and is normalized again to prevent clipping).

For a while, I was utilizing a third file, which contained my stereo mixdown. I would run the mixdown through a comp/limiter plugin to even out the mix and to make up the gain reduction the two normalizations had cost me. This was a mistake. Though I did not go over digital zero, the levels were too hot and distorted. Well, without this third step, nothing's distorted, and the peak is digital zero (somewhere in the song, anyhow!), but the general levels are too low. So I know I'm doing about eighty things wrong. Thanks in advance; any help is gonna be deeply appreciated ...
 
First one question: Why do you first mix down your instruments, and THEN add vocals and mix down again? To me the logical way is to record everything and mix down ONCE...

Second question: You say "I do the same for rhythm guitar, for the kick drum, the snare, etc." None of these are stereo sources. Recording them in stereo is overkill. Only record stereo sources in stereo.

You can get unnatural stereo spreads if you have phase problems between the left and right channels. Thats the only ideas I have...
 
Whoops! Sorry about not making that clear in my first post.

Why I do it in steps: I only have X-number of tracks available to record on before I run out of RAM, or before the CPU chokes and you can't keep playing back without crashing. I record sixteen or so instrument tracks, or whatever number, then mix and bounce them down to two tracks and record the vocals and mix those. Basically, it's like if you recorded on an analog eight-track or whatever -- I'm bouncing the instrument part of the mix to free up tracks. Ideally, sure, I'd like to do it with instruments and vocals simultaneously in the same mix. Or course, I can go back to the instrument mix, change levels, whatever and re-bounce them to two tracks, so I'm not stuck with that instrument mix, or that vocal mix. So I guess things could be worse.

As for recording my instruments in stereo, that's more or less the basis of my post: I record EVERYTHING from keyboards in true DIY fashion, meaning, all my guitar parts, drums and bass are MIDI patches on the keyboards. They're ALL stereo sources, technically.

Logically, these patches are all stereo patches, and in the case of my drum machine, the entire kit is pre-panned (just as the keyboard patches are pre-panned) and output from its left-right output jacks. I just always assumed that's what you did when working with stereo keyboards -- pan all the way out and record, say, a stereo electric guitar patch in stereo, like it is on the keyboard. letting however the instruments were set up to pan in the factory decide how things sat in the stereo field. Or plug your drum machine's stereo outputs into your mixer and just record your kit, or indiviual drums, in the stereo field. What I was asking above is, is that right? Doing it the way I've just decribed seems to make everything sound too spread out and hokey. I notice that on a "real song", things sound much more centered, particularly drums. Should I do that? Or should I make all the instruments mono, recording them from their mono outputs only with the pan centered on my mixer, then create a stereo field in the mix by panning the individual instruments?
 
OK, I thought you recorded the each drum onto a separate track in stereo...

Thats omething TECHNICALLY is stereo doesn't mean you have to record it to stereo, because if you are recording a buzzing lead synth, even of it has stereo-outs, if there is no difference between left and right, doing a stereo recording is meaningless.

Otherwise you seem to be doing it correct, unless your PC software actually records things in stereo. I have no idea if Vision DSP does that. If it does, you might have two stereo recordings instead of one mono, which could fuck things up. Also, you might wanna try skiping the Mackie when recording the keyboards. Doesn't seem useful too me.
 
No -- that IS what I've been doing: recording the snare from my drum machine in stereo, through the mackie with the pan knobs turned extreme left and right, onto its own stereo instrument track in Vision, recording the kick in stereo, the bass from my MPS in stereo, the electric guitar patch, etc. Originally, I wasn't recording anything in mono at all.

Unfortunately, unlike a "real" tape machine, you can have up to 64 tracks of either mono OR stereo instrument tracks (you can set them to mono or stereo individually) in Vision DSP. The mono tracks you can pan, like you would on your analog 4 or eight track. The stereo tracks are in stereo, so you don't have the pan option. I've been recording all my MIDI instruments' audio in stereo tracks, panned all the way out. Traditionally, I guess you'd record the right stereo side of a stereo keyboard patch onto its own mono track, and the left one to ITS own mono track, then utilize the pan controls to shape its stereo width after recording. The way I do it, I have it panned out all the way when it's recorded, then it's on a stereo audio track and you can't pan the left and right channels indiviually; you're stuck with the stereo field you have recorded. Again, I just assumed that the stereo keyboards and drum machines were already panned the way they were supposed to, definitely the way they sounded when I composed and arranged in MIDI, so why screw with it? D'oh!

I SEPERATED the drums into kick, snare, hats & crash and a toms' track, and recorded them indiviually, but they're indivually recorded in stereo tracks from the drum machine's stereo outputs, not mono, which I assume would be the right way. And I don't record them dry, either -- I leave the drum machine's built-in verb on and augment it with computer stereo-to-stereo verb. Same with the other keyboard tracks. It's all messed up ...

Yeah, I thought about the Mackie, and running keyboards through any sort of mixer as being a contributor to the problem (not to mention added noise) ... I mainly do it that way to boost or cut levels being fed from the mixer into the computer's sound input.

By the way, thank you for your help so far!
 
I have to admit I went through youre post quickly and so I hope I didn't miss any info.

Regebro was right.

Check your Keyboard sounds by editing them and I'm sure you will find that
it doesn't output many of it's patches in a stereo mode.
Unless you are using a most interesting bass sound I highly doubt it is stereo.

In short: don't download your sounds stereo unless you are sure that they are.
When you pan them do you check the left and right and hear a diference ?.

To start a good mix you must NOT record them wet. Disable your effects UNLESS it's a part of the sound and you want that specific effect. Best to keep them dry until mixdown.

The drums I would not record Stereo unless you are recording the whole kit at once to a stereo track. Then it would be a good idea to use the stereo outputs.
Look at your keyboard and drum machine and look for a mono output as part of the stereo output pair. One of them should say mono and use that.

Don't record any sound panned during the process unless you are sure that it's a stereo patch and you want that sound panned that way other wise record the 2 tracks (left - right) as if it was a double mono track. Then you will have the freedom to pan as you wish.
Not every stereo sound should be panned hard left and right.
 
Well, now, I did just try recording the drums from the left (mono) output -- at least the ones you're traditionally SUPPOSED to, such as kick, snare, hi-hat, tambourine, etc. I guess I would still record toms and the cymbal section in stereo -- but as you said, record left onto its own mono channel and right on ITS own mono channel. Means more channels, but now I can control the left-right spread, right? I mean, correct?

I think you and regebro outlayed two distinct problems that might make all the difference here, and I'm gonna try them both. A) I shouldn't input thru the mixer for the keyboards and drum machine; I should go directly from their out(s) into the computer's audio input, be it the stereo pair or the left (mono) output. I tried recording a discreet left and right signal from the keyboards, through the mixer, and recording the left and right channels onto two mono tracks, left being its own and right being ITS own ... with the pan controls on the mixer set dead center. Well, you know the results. All that did was pan two signals into the same path and it might as well have been mono. The way regebro described -- bypassing the mixer and inputting direct from the keyboards into the computer, would keep the left and right channels on their own channels, and this way I could record a true mono left channel and a true right mono channel for stereo instruments, like strings or piano -- then, I could pan the left and right channels at will.

B) not recording them wet -- also a great idea. Again, I assumed you just trusted in the factory presets and let their reverb do the work rather than add it yourself. Which, by the way, isn't true; I've often had to augment the input reverb with extra reverb at the mixing stage, anyhow. And I'm assuming you lose volume that way if you use the drum machine's reverb; you get a cleaner sound?

Okay, damn good advice guys, which unfortunately only leads me to question #2, and one I've not seen covered much in any forum or (forgive me, Dragon!) online column: what's supposed to be stereo and what's not? I think I've got the gist down: kicks, snares, hi-hats, tambourines, maybe your rides are panned mono. Your bass is mono. I'm assuming that, since "real" toms and crashes are usually recorded in a pair of overhead mics, they're recorded in stereo. Some of the other stuff, like pianos, strings, chorus, French horn, orchestra sounds are traditionally recorded in stereo. I read somewhere that acoustic guitars are usually recorded stereo. But what about electric guitars (I keep hearing songs where it sounds like the rhythm guitars are in stereo)? And some other questionable instruments? Leads? I warned you guys; I'm a child when it comes to this stuff ...

Thanks again for the advice so far -- this is exactly what I need: definite answers!
 
I can relate to your recording, as I do a lot of electronic, *direct in* style recording also. My POD is also stereo out..same for just about everything else. I think the trick is to select a few key instruments that are the most *important* and keep them stereo. You might pan some of them at 12 o'clock and 3 versus 9 and 3....not as wide a pan, and off to one side. I like drums (especially toms) in stereo, not to mention the nice keyboard pads and such. Solos are nice in stereo, not to mention MONO vocals with stereo reverb. Try not to overlap you sources, but position them in slightly differnt left/right perspectives. You will still get a nice spread, but not over do it if done carefully...and remember...always check that mix in mono (for phase cancellation).
 
The rule of thumb is that single sound sources should be mono and others stereo. Now you can often record things like guitars in stereo, anyway, to get a fat sound with a wide stereo spread, but unless you are out for that effect, any single sound source is mono.

So, a trombone should be mono, but a horn section stereo, basically. (And as always, breaking rules is a Good Thing). :)
 
You must log in or register to reply here.
Back
Top