causes any weird issues or is cheesy?
so i’m finishing up a demo of another band that i just recorded. i mixed everything straight up the middle, and the drums weren’t in stereo, so the mixes are effectively in mono. now i know that mono is perfectly fine, lots of good stuff has been done mono; but still i kind of want that largeness from a stereo spread.
so, i tried mda’s stereo simulator plugin on the master channel. on the default setting, it pretty much does what i’d expect – makes everything spread out a little more. it also makes it a little louder and maybe harsher (not sure on that one), but overall seems fine.
my question arises because i’m not too familiar with how stereo simulation works and i can imagine that i might have unexpected weird side effects (i don’t know what – weird phasing issues? sounds funny in cars? doesn’t like to be mono?). i don’t really have access to many other stereos to try it out in.
also i have a feeling that it might be cheesy, or the exact sort of thing an amateur does to make his recordings sound “better” and ends up making them sound worse.
so: anybody’s perspective on the idea? and any recommended resources on the topic?
thanks!
chuck
Good news first: Starting by making the best mono mix you can is an EXCELLENT way to do any mix. You see, it’s harder to get clarity and separation in mono – in stereo, you can use direction and imaging and all sorts of nifty stuff to help separate the instruments in the mind’s eye. That can make you get lazy.
(Note: all rules are made to be broken. Rather, the reasons for the rules should be understood, so you know when you care and when you don’t. So, keep that in mind as I make some generalizations.)
Mixes should sound good in mono. Mixes that only sound good in stereo have a number of problems in different listening situations. A mix that sounds good in mono will sound good anywhere. So, good for you!
Now, let’s move on to stereo.
I’ve used pitch shift doubling and similar techniques to create a stereo image in a mix that’s already mono and can’t be separated – i.e., was recorded on one channel in a bar, or back in 4-track days after a bounce. It’s a great tool to have in the toolbox. In this case, you don’t need it. If you like how it sounds, great! However, these effects usually have artifacts that will sound bad in some situations due to phase cancellation – you’ll hear a colored or phasey tone.
There are different ways to simulate stereo, but most of them involve delay or phase-related techniques, and depending on the method used, it’ll either sound funny when summed to mono or else when you’re too close to one speaker. A good mix sounds good no matter where you are (though it should sound a whole lot better in the sweet spot, of course).
Now, most of us record most of our tracks in mono, with the big exception of drums, which are usually miked with multiple mikes and with stereo in mind when positioning the mikes, so that at least two can be panned pretty wide (the overheads, usually). Another exception is keyboards: if the keyboards are stereo, we record 'em that way because they sound a lot better that way. Let’s skip those for now and focus on all those mono tracks. What to do to creat a stereo image? Here are a few common techniques.
1) Record the exact same part again and pan them apart. This works great with rhythm acoustic guitar and electric guitar. I mean, it’s really a good technique, especially if you can play repeatably and with good timing. I can’t and I still like this technique. Once you use it, you’ll find there are nice little variations to use in the two parts to add interest. Way cool. Few disadvantages, other than too much of it can clutter and if the timing isn’t good it gets worse with added tracks.
2) Panning to make a soundstage. This works best when there are enough tracks that you can do it without making the mix lopsided. The downside is that sometimes you end up with too many tracks, diluting the goodness. It’s a commonly over-used technique, mostly because it’s the simplest.
3) Cloning a track and making a small change in the copy – like dragging it in time a VERY tiny bit, and panning the two apart. This technique can be very good for things where coloration is OK, like electric guitar. IMHO, it’s usually a disaster on vocals, but some folks like it. It’s a matter of taste. When using this technique, you always have to double-check your mix in mono carefully, because it can cause problems (coloration) when summed to mono. I rarely use this techique, but it is popular.
4) Stereo reverb. I probably over-use this, because it’s so nifty, and also because of the kind of music I tend to make is natural sounding, with a fair amount of acoustic instruments. Pay attention:
– a) In Track Properties, check the “expand mono track to stereo”. Note that this does NOT cause a stereo effect. All it does is make two channels internally so that plugins now can do stereo stuff. If you don’t check this box, all plugin effects on that track will be mono! This does NOT apply to FX on aux buses or groups. Sorry, too much detail for now probably!
– b) Plug in “n-Track Reverb” and select “Ambience 2” preset. While playing, dial up some reverb, with the track panned center or mostly center. Note that as you get reverb, it also starts to “place” the track in a virtual room that you can almost imagine. Here’s the deal: there are controls at the bottom that you use to control whether you’re adjusting the right side, left side, or both at once. Fuss with that while modifying something like “early reflections” or “room size” and see how it sounds when you make the left and right side different. TIP: You can play this trick with any FX that supports stereo, and almost all do!
The main downside to this is the mix can get too reverby. Since it’s the early reflections and not the tails that give the image, I crank up the damping a bit to deaden the tails. Other reverbs are good for this too, so now you get to spend hours upon hours sifting through all the neat free and expensive reverbs to find the ones you like best for this purpose. The other downside to this isn’t really a downside, but the effect is rather subtle instead of in-your-face, and sometimes we want the in-your-face effect.
5) Comb Filter. Download the effects from mda-vst.com and try the “stereo” plugin. (As before, check that expand box!) This little gem uses another technique to create a fake stereo image, that’s nice on some things, like cymbals, acoustic and electric guitars. One great thing about it is it sums to mono with the effect cancelling completely. The problem with it is it can throw the bass balance off, so you usually want to use it on something where you’re going to cut the lows anyway. I remember seeing a free comb filter plugin that had a crossover so that the bass bypassed the stereo effect, but I don’t remember where.
6) Pitch Shift doubling. This is the trickiest one to apply well. It was generally better back in the days of hardware, strangely (since it was done using digital delays, which are far, far better these days in software, right? right.) It’s great for taking a complex mix of sounds and putting each sound in its own place in the soundstage: and not just left and right but back and forward and even up and down. It works like a good stereo chorus does, except that (a) the image stays still rather than twirling, and (b) shouldn’t cause obvious coloration that we normally get (and want) with chorus.
Which leads me to …
7) Stereo chorus. You probably know just what this is and what it’s good for and what it’s not.
8) Stuff I don’t know about or understand. There are stereo widening effects that don’t say how they work, and I can’t quite tell by listening how they work (although I usually suspect mid-side pitch shift doubling, but it could be other stuff …)
That pretty much covers the tip of the iceberg. Welcome to the fascinating subject of mixing, soundstage creation, and stereo effects! Don’t worry, there’s hope: someday you just might get your life back.
BTW, It’s fine to combine a couple of these effects (using them on different tracks). If you use too many in the same mix it can lead to mud, though, so just pay attention.
Cheers
Jeff
Man, Jeff! Are you a professor? Cause you sure know how to profess!
That’s quite a lot of useful info, there, and thanks for posting it. For the sake of my future time on this forum, though, I should probably expound a little more on this situation.
This was actually a four-track recording of a band with bass, drums, guitar, and vocals. Having only one of each instrument means that panning would, like you said, make the mix feel lopsided (and I’ve heard Jimi Hendrix and Beatles songs with drums panned to one side and all that, but it still feels funny to me) – that’s why the mix is mono. That’s also why the drums are mono – I did use several mics on the set, but had to mix them down to one channel on the four-track. (The band is from out-of-town, and we met halfway between where they live and where I live, so I couldn’t bring my computer and do numerous overdubs.)
The “artifacts that will sound bad in some situations” was exactly what I was afraid of – thanks for confirming that. And you also reminded me to try out stereo reverb, which is what I thought of first and then apparently forgot 
And btw, the mda Stereo plugin is exactly what I was using 
I didn’t notice it throwing off the bass balance! I’ll check that out now. The only problem I’d noticed so far is that it makes the guitar kind of huge and overwhelming – almost TOO rocking! I know, I know, is that even possible?
Lastly, how does “pitch shift doubling” work? Is this just adding an octave somewhere to thicken it up? I can’t imagine how to use that to achieve an artificial stereo image.
Thanks much for the tips! I’ll surely come back to that list of things repeatedly.
chuck
Some thoughts…
Actually this guitar-bass-drums-vocals band situation isn’t unusual. What is unfortunate is mono drums.
I’ve tried a trick I heard on some early Van Halen stuff… Pan the guitar a little to one side and use a short echo panned to the other. Instant “stereo” guitar. Careful though or it gets to “'verby” as Jeff has mentioned.
Bass and lead vox are OK in the middle. Perhaps some stereo delay or reverb on the ld vox as Jeff has suggested.
Drums… Hmmm. re-record in stereo? (only kidding). How about this for an idea… I’ve not tried this so YMMV…
Clone the track twice so you have 3 drum tracks in total. Pan one left, one right and one centre. Then use EQ on each track to expose bits of the kit. Be brutal. Focus on the hihat (for example) and try to bring it out so you can pan it around. Perhaps gating would help too. Which brings me to another idea…
Try EQing and gating and KTDrumtrigger to trigger some drum samples via midi. You can then use stereo samples or pan to your heart’s content. I’ve had good success doing this on kick and snare. Again YMMV.
HTH
Mark
Good ideas, Mark! I don’t know if I can really use 'em this time, though; I’ll have to try them next time.
The idea of stereo-delaying the guitar sounds neat, but I don’t know if it’s possible at the moment (not to mention I’ve already mixed down and don’t really want to do it again! 
).
I’m working with pretty old equipment here. The input to my computer is via the line-in on my SoundBlaster16 (!). It’s actually not too shabby, but since I only have a stereo in, that means I can’t send all four tracks in separately. I’ve tried two-at-a-time, but this 4-track (Yamaha MT100 II) doesn’t play the exact same speed every time and so they don’t line up. And I’m a little intimidated by the fine time-stretching required to make that work! So, the computer is my mix-down deck in this situation. (And that unfortunately kills the possibility of working individually with the drum track. I could re-mix, sending the drums to one channel and everything else to another, but I still don’t really want to remix.)
For mixing I’ve been experimenting and having fun. The four-track has an effects send, so I used my delay pedal to simulate reverb (short delay with a lot of feedback). If I had another delay pedal I might could do the stereo guitar trick, but, alas.
Also, the vocalist has kind of thin, high-pitched vocals, so to thicken them up a little I inserted my octave pedal! You can hear the effect, but in the midst of all the noisy cymbals and delay on the vocals, it doesn’t really sound weird. Or maybe that’s just me being optimistic
Anyway, thanks for the ideas! This li’l project has been a fertile ground, as is this forum. Definitely looking forward to participating more in the future.
chuck
<!–QuoteBegin>
| Quote |
| I’ve tried two-at-a-time, but this 4-track (Yamaha MT100 II) doesn’t play the exact same speed every time and so they don’t line up. And I’m a little intimidated by the fine time-stretching required to make that work! |
Don’t be put off by this technique. It’s much easier to do than to read - trust me, I’ve done it! Then you’ll have so much more freedom for mixing in the digital domain.
Well, Mark, I’d think it would depend on how much crosstalk (bleed between tracks) there is. If there’s a lot (e.g., recorded live in studio) that would probably get pretty messy. If it’s a one-track-at-a-time recording then chances are good. It also depends on how stable the tape drive speed is, of course. But if Mark says it’s worth a try then it’s worth a try!
The MDA stereo effect is one that would have the least artifacts – actually, other than that bass shift I mentioned (which may or may not happen due to things like, IIRC, what key the song is in (!)), it’s the safest and least likely to have phase cancellation problems causing coloration. If it sounds good, then chances are it IS good. Be sure to listen to the mix on a few different stereos. Or just leave it mono, which isn’t a sin! If you do create a stereo image, don’t overdo it.
In the case of building a stereo image for a mixed track, that’s exactly what I used pitch shift doubling (PSD) for back in the days of 4-track recording, on mixdown after bouncing 3 rhythm tracks to a single track and then recording 3 more tracks. I’d use PSD on the mono rhythm track submix. It was a hardware device built back in the 80’s, and despite the rather poor A-D converters in those days, worked a lot better than any digital approaches I’ve tried recently.
To do PSD, you pitch shift by a tiny amount. You can send the dry signal to one side and the shifted one to the other. Alternatively, for an effect that completely cancels out in mono (which is good, since it’s only purpose in our case is for stereo imaging) is “mid-side” (MS) , where you send the dry signal to the center and with the shifted signal, add it to left and subtract it from right (or vice versa). The results sound almost the same, suprisingly. MDA has a plugin called “image” that’s a great little mid-side encoder-decoder, which I use for purposes like this.
When using non MS stereo techniques, you have to test the mix in mono mode. When using MS techniques, you can test it in mono but if you’re doing it right, the effect should cancel out completely to give your original dry track. But there’s another cross-check that’s important: listen to just the left or just the right side. When using delay-based techniques in MS, instead of getting coloration in mono, you get it when you only hear one side or the other.
If you’re not sure how a plugin works, it’s best to check both mono mode and L and R separately. (In n-Track, leave the master channel in mono mode, and just pan the affected track to each side and listen. If it sounds OK while panning the whole time from side to side, you’re golden. If there are some minor issues you can usually ignore them. You’re looking for noticeably unpleasant coloration here, not subtleties.
OK, now I’m going on memory rather than knowledge. But IIRC, while MDA stereo (comb filter) sounds good, it doesn’t create a “partitioned” image – all the sound is spread, or perhaps smeared – across the soundstage. Yes, it does still help the brain to tell the instruments apart, but not nearly as well as a good application of PSD. With PSD, when applied well, it’s almost astounding how well the brain can separate the different instruments.
With PSD, the location in the image is related to the sound color and pitches. Since the pitch shift is small, bass stays nicely put in the center. Each drum goes to its own location in the image, though that location has little to do with the location in a drum kit – just the pitch and color of the drum. Piano has a nice spread in its own region, and ditto electric guitar (though I haven’t tried this on a heavy metal sound, more like blues guitar tones).
Here’s the problem. This effect works using phase information (= delay information). The old hardware device I used had dual bucket brigade delay lines, feeding both at the same rate. To produce the pitch shift, it would pull the data out of one delay line faster or slower. Eventually this becomes impossible, either because it runs out (pulling out faster) or full (pulling out slower), so at that time it switched to the other delay line, cross-fading between them during the switchover. This cross-fading causes noticeable phasey glitches in high clear notes, so I just didn’t put any of those in the rhythm track.
However, modern pitch shifters use FFT-based techniques, which do a nice job of pitch shifting, but completely mess up all the phase information, which gets lost in the process (if my hazy understanding of it is correct – but it’s what my ears tell me as well). Since phase info is critical to the effect, it doesn’t work nearly so well: it does make a stereo image, but the instruments aren’t nearly as well defined.
I’m still looking for a plugin that does pitch shifting using a technique similar to the bucket-brigade delay. Most chorus plugins use delay lines, but rather than taking out at a constant rate (faster or slower than real-time), they modulate faster and then slower, so there’s no need to cross-fade between two different delay lines.
However, this chorus technique sounds terrible when used at 44.1kHz sample rates. When using most chorus plugins, you need either record at a much higher rate or else upsample it (to 192k ideally), apply the effect, and then downsample back to match the rest of your tracks. If I found a good PSD that used dual delays, this would be necessary there too.
See what I mean that there are technical issues? I haven’t sorted out an ideal method yet, unfortunately.
PS: I’m not a professor but some day I’d like to be. But I suppose I’d have to go back to school and at least get my bachelor’s degree first! I wonder if credits from 1975-1978 are any good today?
<!–QuoteBegin>
| Quote |
| Well, Mark, I’d think it would depend on how much crosstalk (bleed between tracks) there is. If there’s a lot (e.g., recorded live in studio) that would probably get pretty messy. If it’s a one-track-at-a-time recording then chances are good. It also depends on how stable the tape drive speed is, of course. But if Mark says it’s worth a try then it’s worth a try! |
Good point Jeff. I did this with individually recorded tracks. I guess the worse that could happen would be some phase stuff between the “main” track and the bleed.
I used my old Yamaha MT150 (like aikan’s but a different model) and it worked well. I can’t remember exactly but there was little more that a millisecond or two that needed to be accounted for and witht he “stretch” thing described in the FAQ it was a breeze.
Of course I then dumped the tracks because they were rubbish compared to what I was managing to do in n-Track!
Mark
Very useful info, Jeff! Thanks again.
There’s actually not too much bleed between tracks, despite it being recorded live in a church. To block off the guitar amp from going into the drum mics, we had to put a table on its side in front of the amp. That was fun for the guitar player I’m sure 
It’s true, the mda Stero plugin just spreads everything out like butter, and seems to collapse to mono well. I totally forgot to check for bass cancellation, though. Oh well.
This has been a really interesting conversation. I went ahead and left it in mono because I just need to get it done, but I’ve definitely learned a lot from you two. If you want to check it out, I put the tracks up in mp3 for the band to listen to, they’re (temporarily) located right here:
tectophonic.com/tka/
(The band’s name is the Kennedy Assassination. Not sure why, but I’d guess it’s just to be shocking.) Now the recording itself could be better, but what the heck, I just feel like posting it since I can! 
You don’t have to listen to’em.
Thanks again!
Chuck
Wow, that’s nasty stuff. I mean that in a way that would be taken as a compliment in certain circles, hopefully the players involved …
I think it sounds pretty good in mono, gets that retro ear-bleeding thing going. But it might still be fun playing with master stereo imaging.
Next time, record directly into your computer! If you’re diligent, you can find an 8-channel card for under $100 (I did, for my son). Let me know if you want my ebay seach string.
Yeah, you know, I realized that this doesn’t even sound like a CD. Even on a CD, it sounds like a tape. These guys aren’t terribly punk-rock, but I think that’s what they’re going for in the music. So, the mono thing works.
And I would love that search string. I will probably need a new hard drive, too, while I’m at it
chuck
Not unless you need the space or plan on doing lots and lots of tracks. I did fine with the 4200 RPM drive on my laptop, frankly. Up to a dozen or so tracks, no problem. You only need a faster hard drive for handling lots of channels, especially when using lots of channels for both record and playback at once.
My ebay search string is:
(“delta 1010LT”, “delta 1010”, delta1010, dsp2000, layla20, “layla 20”, ews88mt, “ews 88mt”, terretec88, “terretec 88”, inca88, “inca 88”, aark24, “aark 24”)
sweet! thanks man!
(i’d need a new harddrive because mine’s only 12 gigs it works though!)