This is something I've been wrestling with for awhile. Ideally, the ultimate way to record a solo acoustic guitar is with the right mics in the right position, in a great-sounding room, through the right preamp, and straight to tape -- no EQ, reverb, compression, limiting or any of that stuff that you hear on 99% of the commercial recordings released these days. We're used to hearing these kinds of processors and they sound "right," but they degrade the recorded signal very noticeably. If you want "real," it's best to leave this stuff out.

However, this purist approach assumes that the people who *listen* to the recording will have a great playback system that can handle the wide dynamic range of an unprocessed guitar, and many people don't. I'm working on a CD of original guitar music, and I need to bite the bullet and use some of this stuff (I've got some Waves software and some good hardware too). I haven't spent a lot of time playing my stuff over cheap speakers to get a sense of exactly how much EQ, compression and/or limiting I need to use so that it doesn't overload these kinds of systems.

So, elambo, rockerbob, and anyone else out there who does a lot of this -- how much compression do you use on acoustic guitar? What ratios, what threshold, etc? Do you use multiband compresion? Do you use limiting? How can I compress the guitar enough that little speakers can handle it, but without losing high frequencies and the natural sound of the guitar? Do I need to use EQ to put those frequencies back after I compress? I'd appreciate any thoughts about this.

phoenix - After you took the time to send me that HUGE file (thank you again) I'll be sure to help with any of these questions.

But first I have a couple questions for you:

-You said you have Waves, which plugins and on what type of software are they being used? Waves makes some great stuff.

-Also, you mentioned hardware - what hardware do you have available to you? Don't be afraid to be specific with model numbers, etc. If it's really good stuff it shouldn't degrade the audio quality much at all as long as it's set right. And you're right, you may need to eq after you compress to get some of the frequencies back. This depends on the compressor and how it's set. That's pretty common, though. Sometimes eq before, sometimes after, and in some cases before AND after. You might eq first to 'repair' (if necessary) then compress, then eq again to 'enhance'. Ideally you wouldn't eq at all, let the positioning of the mics be your eq. From the MP3s I've heard you record I'd say you're doing good.

-Will this be guitar only? How you mix the guitar will depend heavily on how much of the overall track it's able to consume.

I'd be happy to be more specific when I know more about what you're working with.

(Have you decided on which guitar you'll be recording?)

As something to think about - for a minimalist approach to recording and mixing acoustic music, listen to 'Joshua Judges Ruth' by Lyle Lovett. It was mixed by George Massenburg using all of his own gear. His eq'ing and compressing is very sparse and it's one of the best sounding records I've heard. It's pretty dynamic, but not so much so that it wouldn't translate to smaller speakers. I think I remember him saying that Lyle wasn't using a Collings at that point (probably a Martin then), but the guitar sound on that album is first rate. Listen to 'North Dakota' (I think that's what it's called) - it's simple, mainly percussion, piano and guitar at first, then it builds a little, but listen to how the guitar sounds. When I listen at home and close my eyes I would swear that Lyle was sitting on a chair a couple feet in front of me.

I've been doing this home recording thing for about 20 years, since the first 4 track cassette machine, Tascam 144 PortaStudio. I still have the light bulb going on as I discover something I should have been doing all along.

My goal has been to create what I call the "Magic Mix". The mix that sounds good everywhere, and sounds outstanding on a good system. A portable mix. This is far from trivial. Also, most of my stuff is pop ensemble type stuff. Drums, bass, guitars, vocals. The Magic Mix is elusive. I guess the reason for this is that no matter how good it sounds here in my studio, I want other people to hear it. Most won’t come to my studio. That means giving them a copy that I no longer have control over. Who knows what the heck it will be played back on?

I have always done my recording on my own with a budget studio. I have only been in a commercial recording studio a few times, and always as a guitar player, ever engineer or producer. Almost nothing I use here at home would be considered absolute top shelf except for my guitars and my computer. There are plenty who would argue that those aren't top shelf either. But because of the low budget I have had to find ways to fake things to get the sounds I wanted, and I was never sure what sound I wanted. Today, after many years of upgrades, I still work in what would be considered a high end home studio, still miles from a truly pro high end studio. I use a pair of the cheapest large diaghram condenser mics that would be considered pro. AKG C414 ULS. Under $1000 street price and pretty darn good. Not in the same class as $4000 mics, but pretty darn good. Sometime I feel like I need to EQ though. I always feel like I need to remove some low frequencies. I use the bass roll-off on the mic at 150Hz and another on the Mackie board I use for a pre-amp, also at 150Hz. There is no music coming from the guitar below 150 anyway and it's going to turn into the "whomp" sound you talk about when played back on some big home system with a subwoofer made for DVD movies. Most of the time I also run a high pass filter from a software parametric EQ and remove even more lows, maybe up to 200 or 300Hz, depending on what else is in the mix. I can only hear the first low cut very well in the studio. The others are very subtle on the monitors because mostly I'm just lower the same frequencies I have already lowered. Where I hear this help is on the big monster playback system tuned to shake the house when something blows up on screen. These days I remove all the lows from everything except bass and kick drum. Even the few acoustic recordings I've done recently. I might leave a little more, but not much or it gets to "whompy" on the big systems.

Of course this also help keep the high end sparkling. I don't have to do anything to make the high end crystal clear. With the low end whomp gone the highs sound great without doing anything. I have found that messing with high end EQ screws up the sound more noticeably than EQing the low end. Usually the only EQing I do at all is cutting the lows.

BTW, I’m recording my tracks into Digital Performer 3 running on a Mac with dual 1Ghz G4 processors. I’m using an M-Audio Delta 66 for my digital I/O. I have it connected with balanced lines in and out of a Mackie 8*24 eight buss mixer. I usually record 24bit 48000. I choose 48000 because in many cases when I mix tracks I go to analog to run through the mixer where I add any MIDI tracks live right to the mixdown. This mix then goes to a Tascam DAT recorder and them would be digitally transferred back as a mix for final EQ/compress, usually with a multiband compressor. With the new Mac I can handle more tracks, so I started recording the MIDI tracks, usually drums, into the computer. Then I can mix the whole thing in the computer without ever going back to analog. If I stick with this I will probably track at 44100 so I don’t have to sample rate convert.

I think the best things I've done to improve my final output are the monitors and multiband compression. A few years back I got some Mackie HR824 monitors to replace some mediocre JBLs I had. I really like these monitors and I think my mixes/masters started sounding better right away. Then just a little more recently, as I moved more and more to computer based recording, I started to use a multiband compressor on the final mix, as sort of a mastering step. I'm using software plugins for pretty much all my processing. For the Mastering multiband compressor I've been using PROSONIQ Dynasone. There might be better ones out there now, but I have this and I like what it does. It's kind of like the poor man's Finalizer. I also like all of the Waves stuff, in particular the renaissance compressor for use on individual tracks like guitar and vocal. I almost always compress acoustic guitar and vocal. Somewhere in the 2.5-1 up to 4-1 ratio with medium-fast attack, 15ms maybe, and a pretty quick release, <50. I set the threshold to get maybe 6Db of gain reduction on the peaks, more sometimes, less sometimes. For guitar I find the attack time to be important to the perceived high end. If the attack is too fast you will won't get the crisp sound of the initial attack of the string. I do it by ear. While listening I'll play with all the parameters, but a lot with attack time until I get a sound I like.

I don’t know anything really. One of my revelation years ago was that it should matter if it sounds real or sounds like some instrument in particular at all. My goal is to have it sound like music. I think that’s a tough enough goal. You, my friend, are already there.

OK, so I’m rambling. It makes sense to me.

Hey, thanks a lot for the replies.

Elambo -- I've got a Waves Native Power Pack II from a couple years ago. It's got the Renaissance Compressor and Renaissance EQ. Direct X for PC. This stuff sounds good to me. I use it in Wavelab and (occasionally) Cubase. I've also got an Avalon 747 compressor/program EQ. The compressor has a side chain with adjustable frequencies (as you probably know).
I also just upgraded to the new Wavelab and it comes with a multiband compressor. But I haven't used that at all yet.

Do you have any thoughts about the new Universal Audio/Mackie powered plug-ins (comes with an LA-2A, 1176, Pultec and RealVerb). Is the RealVerb good? I've got a PCM-91. Is the RealVerb as good or better? Was wondering whether it would be worth getting this.

I track to a Tascam DA78. Mainly solo guitar, but some of the material will be multitracked using a 2nd guitar and bass. 1 track will have percussion and violin too.

I'm using MOTU stuff (1224 and 2408 Mk2) for the transfers in and out of my PC. This stuff is a little quirky with Word Clock. It sometimes loses sync for no apparent reason and adds little pops to my audio (which totally drives me around the bend...). I've gotta fix that once and for all.

I'm probably going to use my 000-2H my OM-3 and my Merrill for different songs, depending on which works best for each.

I'll have to pick up that Lovett CD to use as a reference recording. I understand that Massenberg is a genius. Thanks!



Rockerbob,

Interesting stuff! I think your C414's sound great. I was just reading a feature article in Mix about recording this classical/bluegrass violinist (Mark something...). They use all of these ultra high-end DPA mics all over the hall. But for acoustic guitar they use a C414.

That's interesting about attack time. I just started to play with that a lot more. I can see how the longer attack would preserve some high frequencies. Do you find that you can lengthen the attack time to 15 ms without overloading small speakers with those attack transients?

What frequencies do you dial in on the multiband compressor?

Thanks very much, guys.

I'm heading out in five minutes but I'll try to get a few comments in first...

Use the waves stuff when you can. I love their renaissance stuff, especially the compressor. The eq is also good but it might not be the best eq for this type of recording. For your guitar I'd look to the Avalon. It has fantastic eq, but I'm not sold on it's compression.

Reverb - the PCM91 should take you all the way. It will sound better than RealVerb by a long shot, but you may want the UA stuff for everything else. I use the LA-2A and Pultec about every day. The Pultec is limited in function but it's very musical for vocals and guitars...

Oops, gotta go. I'll get back later tonight...

OK, like I was saying, the Universal Audio/Bomb Factory stuff is very, very good. But if you already have a PCM91 then you should make it your primary reverb.

If you're getting pops and clicks it could be a number of things, but most likely it's because of a Word Clock cable, as you have said. Make sure the cable you're using to pass clock is 75ohm, not the 40 or 45ohm type, and not too long - keep it under 6 feet and you're OK. The shorter the better. The MOTU stuff seems to be very picky about clock. If you're not using a dedicated clock cable, but are instead passing it through the audio cables, be careful about lenght again. And try not to let these clock-carrying cables run along power cables. The AC cords can garble the word clock embedded inside the audio cable.

My initial thoughts about your MP3's are that the perspective is very wide and shifted a little towards the left side. I can't tell for sure but it sounds like each Brauner is panned equally to both sides, but the left-hand side which is closest to the soundhole and therefore louder by nature makes the listener feel like the whole track is panned heavier to the left. There's just less energy making it's way to the mic at the fretboard. I would recommend bringing in both sides just a little then bring up the right side in volume a little so that the perspective comes back towards the center. It will still sound wide-stereo, but it won't sound shifted. And bringing in the panning a little will let the reverb take over that space on the extreme outsides, which the ear will perceive as more natural.

Be careful with multiband compressors because they are quick to get ugly. They divide the audio spectrum into 2 or 3 or 4 or even 5 different frequency 'slices' and compress each of them then recombine them all at the end. This seperation and revival can be dangerous if the unit isn't accurate at splitting up the frequencies. And if the compression settings for each 'band' are very different from each other you start to really distort the original source. This is actually a bit like eq, but instead you're compressing each freq. range which is more harmful to the source. It can definitely help, but it's easy to abuse. Use to taste.

One of the best tools for mixing this kind of music is the mind. I often just close my eyes and listen, and if I don't believe what I'm hearing I make a change. All of our minds have been preconditioned to hear this type of music in a certain way, because we've all heard it live so many times. If you start to stray too far from this expectation the mind says 'No'. This isn't true for Top 40 and rap and a lot of other genres because most of that stuff is synthetic to begin with and has no real-world precedence. You can (almost) do whatever you want with them, but acoustic music has certain guidelines, if not rules.

The one thing that will be most difficult to talk about on this forum is dynamics. It's hard to say how much is approprite. Again, you need to be honest to the human ear - if you're lightly picking it needs to be quieter than the hard strumming or the ears will say 'huh, there's something strange about that'. But just how much quieter? This is why I recommended Lyle Lovett's CD, it's dynamic enough to sound very natural but it's still controlled. This is where you need to just listen. Don't compress so much that everything is the same volume all the time, that's just boring. Let it breathe, let it sound natural without forcing the listener to get up to turn down the loud parts and turn up the soft parts. On the other hand don't be affraid that the listener's speakers will distort because it's too dynamic. If that were to happen it would more likely be caused by an overload of certain frequencies (usually bass), not dynamics. And you should probably stay away from a limiter. It's simply a compressor with no tolerance. If it sees a level getting to loud it lays down a heavy hand to stop it in it's tracks. These aren't very natural at all, unless they're used very easily.

Tell me about your mic positioning. Where are you putting them, how far away and in what direction are they pointing. They might already be in the best position, I'm just asking out of curiousity.

Elambo,

I really appreciate all the excellent advice -- it really helped a lot.

Will hold on to the PCM 91 (good to know that it beats the RealVerb -- I should have guessed that, given the power of the Lexicon's DSP chips).

I'll have to check the impedance of my WC cables. They're very, very short Apogee WC cables. Thanks for the tip about power cables screwing up WC -- will watch for that. I may just send my stuff into the PC via the Apogee's AES/EBU out going into the 1224.

You know, I'm not really impressed with the Avalon's compressor either. It doesn't sound that natural to me -- I can easily hear it even at fairly low settings. The Waves Compressor sounds as good or better to me. I've been using the Renaissance EQ very conservatively to clean up "mud." I wonder if I'd be better off figuring out what EQ settings to use on the Avalon and then just tracking through that, so I don't do an unnecessary DA/AD conversion.

Re: mic positioning. I have to admit I wasn't quite as careful about mic positioning with the MP3's as I usually am. The VM1-U's were set to omni about 13" from the 14th fret and from the top's belly. Your observations about the left channel were right on -- that was the belly mic, and it picks up a lot more signal. For serious recordings, I usually equalize the level of the two mics, so they're balanced at the center. I decided not to do that with the recent A/B MP3 in order to keep it really raw and unprocessed.

The panning tip is great. You're right, I've got them panned hard L and R right now. I'll move them in a bit so the reverb can fill up the edges.

That's really good to know about cutting low end rather than compressing a lot. I haven't played my stuff much over cheap home systems, but when I have, it DID kind of sound like it was too much low end. I wasn't sure though. You saved me some trial and error there. It's good to know this can be fixed largely with EQ, because I HATE the sound of lots of compression. I don't even want to HEAR compression.

Re: levels. I'll keep that in mind. Instead of normalizing everything close to 0, I'll try and keep the levels realistic (soft for fingerpicking, loud for strumming, etc.)


Did some recording for my CD today. I got some cloths and magazines off the wood floor in my "tracking room" and that saved me a significant EQ "mud" cut. Amazing how cleaning up the room cleans up my sound! It sounded very noticeably brighter and clearer thanks to the floor reflections. I also moved the belly mic over to the area where the strings are near the bridge. That helped clarify things too.

Thanks again for the great advice.

You're D-2H sould be back in a few days, no? Looking forward to a report.

[This message was edited by Phoenix on July 15, 2002 at 10:21 PM.]

I haven't forgotten about the question on the multiband compressor, but I'm on the road. I want to look at what I have been using so I can tell you the bands I use. I do it by ear and don't always pay attention to the bands, but I can lookup what I have used in the past. I'll post again later this week.

Gotta love that humidity in Dallas.

I got back in town and have a few minutes.

I took a look at what I ended up with on a couple of tunes. It looks like I'm spliting at 56, 404 and 6761. I have a high threshold set in the middle two bands and lower in the outside bands. Like elambo said, this can act like EQ by controlling the amount of compression and the output levels of each band. To some extent that is what I use it for, although I think the split compression helps smooth out those sounds that seem to leap out from time to time. Especially in the lows. I want the warmth, but it's easy for the lows to get out of hand just here and there. With the multiband I can compress those out of hand sounds in the bass without messing with the rest of the frequencies. The same for the highs, but I have less problems to deal with there usually. The two middle bands get the least compression, and that is where most of the music is anyway, so to avoid messing it up too much these bands don't hit the compressor much with the higher threshold.

Like elambo says, "use to taste." You can really screw up anything with over use of any of these tools. Believe me, I've done the research. :-) I have overproccessed and ruined many recordings. I do think the multiband compressor has been an improvement in the portability of my masters.

Phoenix, how's the project goin'. I can't wait to hear what you're doing and I'm sure there will be some great things coming out of your studio.

Compression - ya' know, I hate the stuff, too. Unless it's very high quality equip and sparsely used. I've gotten into the habit of drawing all the volume changes into ProTools rather than have a compressor do all the work while stealing the tone. It's a lot more work but it pays off because the original tone is preserved. Actually, I ALWAYS do both; compressor AND manual volume changes. This works well.

The wordclock issue - I would say that you've ruled out the cable issue if you're using the Apogee cable. It's not that, unless it is passing by a few AC cords, but doubtfull. Maybe it's the MOTU. I was using a MOTU Timepiece AV for a while but it's clock was so weak (generic term, but applicable here) that I had to stop using it for wordclock. I've never heard of this problem with the 1224 or 2408, but sometimes it's just a certain combination of devices that don't like each other. You said "I may just send my stuff into the PC via the Apogee's AES/EBU out going into the 1224." How where you doing it before if not via the Apogee and which Apogee are you using? Can you describe your recording chain, from recorder all the way to the computer, including types of cables and which connectors you're using on each device (AES or spdif or optical or firewire, etc.)? This might help...

My D2H - as you probably know from other threads it's still at least 2 weeks away.

elambo,

Thanks! I had some serious MOTU related problems in the process of finishing up a CD project I'm producing for someone. I had to uninstall and reinstall all the MOTU hardward and software and also my Cubase software. In the process, I think I found and solved the root of the clock problem. It kept me up all night and I did a lot of yelling, but now everything is working beautifully. What a pleasure to have signals running through 4 or 5 pieces of good digital gear, meters jumping, and everything running smooth as silk! I believe my MOTU problems are solved once and for all (fingers and toes crossed).

Yeah, compression can be a blessing and a curse. I used my Renaissance compressors a lot on this rock project I just finished and gained a lot of experience from it. Using compression is really an art-form. Everything has to be dialed in perfectly. When you do that, for example, you can make somewhat sloppy rhythm guitar tracks sound great -- really tight and smooth, with nice sustain. If you don't dial everything in exactly right, compression just sounds awful.

That's interesting that you redraw the waves as an alternative to compressing them a lot. I just have Wavelab, but it also allows you to adjust volume envelopes in a very precise, localized way. I use that a lot. I find it to be a good alternative to using lots of parametric EQ too, and the phase-related problems that that causes. If there's one hot frequency on the guitar that booms a bit, I can just draw in a volume cut.

I'm kind of surprised that the guitar repair is getting drawn out like this. I have a feeling, though, that Collings is doing everything to make that top look and sound great, to make the wait worth your while. Hang in there!


Rockerbob,

Thanks a lot for the info on your multiband settings (and sorry it took me so darn long to reply!). Question: do you find that using the multiband compressor sounds better than doing a precise EQ cut in the problem frequencies? Do you hear less phase-shift when using the multiband compressor than when using EQ? Are you using the multiband comp as a susbstitute for EQ, or just when there's a frequency range that is jumping around too much dynamically? Thanks!