Trying Something New

August 11, 2009

In recording and pro-audio it’s all too easy to fall into a rut and never try new things. You have found a variety of mics that you like for certain instruments, particular pieces of outboard gear you always fall back on, and mixing techniques that you constantly use.

There are a number of reasons for that, of course. The biggest is probably time, which often equals money. In live situations, there usually simply isn’t time to experiment, unless you are fortunate enough to work with an act big enough where you can have technical rehearsals and full-blown dress rehearsals. In the studio, you are most likely working with a band who is under very tight budget constraints, and that is not conducive to being able to spend time trying different things.

When you have the time, though, it can be extremely rewarding, and can result in some stunning results.

There is also a trend started by some “how to” books to sort of ‘mix-by-numbers.’ They tell you that the kick drum should be at about -3VU on the stereo bus meters, and the lead vocals should be -5VU, or some such thing. Nonsense. Only your ears can tell you how loud something needs to be in the mix. There are some shortcuts you can sometimes take. For example, if you’ve mix a particular band often enough, you usually have a pretty good idea of where things are going to be panned, so it wouldn’t hurt to set the pan pots where you think they’ll end up. But, don’t be afraid to play around with those while you are mixing. You just might find a location for something in the stereo spread that works better for that particular mix. There are a lot of very stunning effects that can be accomplished with phase, delay, and panning, and you will never stumble across any of them if you don’t take the time to try something new.

Sometimes your mic selection and placement has gotten to be such a habit, that you don’t even consider something really different. The absolute best cranked Les Paul/Marshall tone I’ve ever heard was accomplished by a mic selection and mic placement that I NEVER would have tried. But someone took the time to try it, and found that it really worked.

Another, often ignored reason is that digital equipment makes it difficult, if not impossible, to do much “creative patching.” In the old days, you had a patch bay, with a number of mult points, and you could patch anything to anywhere, combining with other things or splitting the signal along the way. With a digital multi-effect unit, this is simply not possible. Most digital mixers and digital audio workstations impose rather rigid signal flow ideas. Beyond that, it is simply not possible to really explore parameters of the equipment that weren’t programmed into it by the developers.

But whatever your situation, try different things. The next time you are tracking a guitar, use your regular mic and placement, but also set up a completely different mic, and use a different placement. Record it on a separate track so that if you don’t like it, you haven’t cost anything, and compare the two. You might just find that it offers something you didn’t expect, and maybe something you can use alongside the other track in the mix.


Digital Doo-Dads

June 23, 2009

Every time I listen to a digital gadget at the local music store, I am overwhelmed by the mushiness of the patches. Or, maybe a better way to put it is that I am underwhelmed with the usefulness of the gadget. If you assume that the factory patches are the best that it can sound, you probably won’t buy it. Be assured that these factory patches aren’t representative, and are probably almost the worst that it can sound.

Plug a guitar into the latest whiz-bang modeler, select any of the factory patches, and you will assaulted by the combination of every effect known to mankind. All at once. With NO dynamics. How can the people who make these things go to all the trouble to “model” all of those vintage effects pedals and amps, and then have NO CLUE as to what they should sound like when you use them?

This holds true on the guitar modelers as well as digital synthesizers. I’m not saying that they CAN’T sound good. Some of them are capable of sounding very good indeed. It’s just that, if you expect to unpack the thing, select a factory preset, and have it sound good in your band’s mix, you are in for a big disappointment.

There are examples of this in all of them, but one standard factory preset on guitar modelers that seems to transcend brand name is the heavily distorted AND highly compressed patch, with a ton of fizzy distortion, always a lot of chorus, some reverb, and a really scooped EQ. It might sound fun in a music store through some little transistor amp played at 60dB, but it’s not going to work through a 4×12 cabinet, crunched tube amp, played at 110dB, and it’s CERTAINLY not going to fit in the mix. Because of all of the extra compression and time-based effects such as chorus and flanging they always add, there will be NO articulation and dynamics.

But, if you build your own patches from scratch, you can come up with some good sounds. Don’t expect them to sound exactly like your guitar through a cranked Marshall with a pair of 4×12 cabinets, for example, but you can create some patches that sound good in their own right, are quite useful, and will sit well in a mix.

The same holds true for digital synthesizers. You sit down in the local music store to try one out, and every factory patch sounds lush, rich, and full. And almost totally useless in a band setting.

However, if you take the time to learn how to build your own patches, you will probably find that the raw samples and waveforms are pretty usable, and some very good sounding patches can be built from scratch, which are extremely good, are more realistic, and will work well in a mix.

There is something else to consider in the synthesizer patches that claim to be realistic samples of real instruments – and that is that they almost always far less high end and upper midrange than in the real instrument. This adds to the difficulty of getting your instrument to fit in the mix. There are other limitations which, in my mind, are design flaws. One of these is the stubborn refusal by synthesizer makers to put effects in a logical place in the signal chain. A great example of this, and the one that is a huge limitation on Roland synthesizers such as the JV-1080 and XP-50, is Roland’s insistence that a Leslie is an “effect” rather than a “speaker.” As anyone who has ever played a Hammond Organ knows, a Leslie is a SPEAKER, and is therefore the last thing in the audio chain – AFTER the amplifier. The Roland Leslie simulator is not too bad, but by putting it so early in the chain, they’ve made it almost useless. It is such things that make building good patches a real challenge sometimes.

While you can get some patches that will, in a mix, sound “kind of like” another instrument, such as, say, a trumpet; there are some instruments which you will never get close to, no matter how much time you spend creating a patch. A great example of this would be, of course, guitar.

So, know the limitations of the technology and your equipment, take the time to build your own patches, and you will get some very usable patches. Don’t expect to fool anyone into thinking that they are listening to a real vintage synth, a real orchestra, or a real guitar. Just get sounds that will accomplish the same purpose, and that sound good in the mix.

After all, the mix is what’s important. Nobody cares what an instrument sounds like soloed, because nobody but you can solo it. Everyone else in the world will only hear it in the mix.

Audio Purists

March 20, 2009

There is a group of sound guys I call “audio purists.” These are people who eschew anything which colors the sound, such as EQ. To them, “purity” of sound is the ultimate goal, above all else.

All EQ colors the sound, not only by varying the frequency response, but also, as a side effect, affecting the phase angle at certain frequencies. This is most apparent when boosting frequencies, and, in extreme amounts causes “ringing,” or resonance at the boost frequency. While this is clearly heard at extreme levels of boost, it is present to some degree at more moderate amounts of boost. This is what purists object to.

Because of this, a “purist” will, when using the 31 band EQ to tune the system, only use it to cut frequencies. In fact, there are a few graphic EQs on the market that are “cut-only.” They do not boost at all. In theory, this should be a good idea. But, what of a situation, where the system sounds pretty good, but has one frequency area that is slightly lower in response to the rest of the spectrum? Most people would simply use the 1/3 octave to boost those few frequencies the 1 or 2 dB needed to smooth things out. The “purist,” however, would prefer to cut all other frequencies, to avoid boosting any band. However, one thing often overlooked is that cutting adjacent bands does NOT result in a flat response. For example, if you cut every band on a graphic EQ by 3dB, the resulting curve would not be a flat response which was simply 3dB lower than the input. Each band has the most affect at its center frequency, and gradual shoulders which boost or cut less and less the further from the center frequency you get. Also, phase shift problems are most apparent in these “shoulder” areas. These shoulder areas are additive, which means that the cut (or boost) of two adjacent bands combine where the shoulders of the filters for those two bands overlap. Therefore, the result from pulling every band down 3dB would be a response which was down 3dB at each center frequency, with ‘ripples’ between bands of lesser or more attenuation. In addition, there would be phase shifts across the entire spectrum. The end result is that, to avoid boosting in one small area of the frequency range, you would be introducing an odd frequency response in the entire spectrum which would be filled with phase anomalies. This in the name of “audio purity?”

A related thought is that they will refrain from using any (ANY!) channel EQ. In fact, they will switch it out of the circuit. This idea actually has some merit, of not carried to the extreme. Their thinking is that EQ is bad, for the reasons I stated above, and since they aren’t going to use the channel EQ, they might as well take it out of the signal path. Since every circuit adds some small amount of noise, you can avoid adding it by not having that circuit in the path at all. Consider that each EQ adds some noise, if you remove the EQs from all 32, 48, 64, or however many channels you are using, this can add up to quite a bit of noise you are avoiding. The thought goes that you should get your sound solely from mic choice and placement. In a situation where you want the most natural sounds possible, and if you are working with acoustic instruments where a “natural sound” is desired, this may be possible. I agree that you should do everything you can to get the sounds you need with mic choice and placement, but in a live situation, it’s not always possible, and you are most likely not working with cellos, violins, violas, etc. So, what is a “natural” sound for a synthesizer, electric guitar, or bass? Odds are, you are going to want to do some EQ to each one, or you will end up with a lifeless and unexciting mix. Can you imagine the average kick drum in a rock mix if you couldn’t have any EQ on it? Cheap EQ sections can sound pretty bad, but if you have good channel EQ, there’s no reason not to use it. Every board sounds different, and the EQ is a big part of this. It is one of the major things you should listen carefully to when shopping for a new mixer.

Another technique that is rather common, or at least was, among the purists,  is to put all of the channel faders at the +/-0 line, and do all of the mixing with the mic trim pots. I’m not real sure where this technique came from, other than the “purists” see a point on the fader where it is neither attenuating or boosting, and they figure that is the “natural” (or “neutral”) spot. However, if you read one of my earlier posts, you may remember that, on instruments that need to be quieter in the mix, this can result in added noise, since you are turning it down at the trim pot, and then running the fader at 0. Whatever noise is added by the channel’s electronics would be better reduced by getting a good strong signal through the channel, and then running the fader at -15dB, or where ever you need it. In my opinion, it is far better to get as hot of a signal as you can coming in to the channel. During sound check, have the player go through his loudest part, and use the PFL meter to set the input level to 0VU and do your mixing with the channel fader. That way, you have plenty of signal to work with, and if you are also sending monitor mixes from the FOH console, makes it MUCH easier to deal with.

A lot of the ideas that the “audio purists” have are based on situations in the mythical ‘ideal situation,’ but don’t often translate well to the real world of mixing a rock band. As always, use your ears and judge for yourself, but keep these things in mind when some “helpful” purist starts making suggestions. Try everything, and keep what works for you – just always consider every aspect and consequence of every technique you try. Otherwise, you may not know what is causing other, seemingly unrelated problems.

Free time, Linux Audio, and Songwriting?

February 4, 2009

It’s been a while since I wrote much of anything. I still keep an eye on the server, but it just seems like I haven’t been able to find the free time to sit down and write much of anything. Two things have appeared that have eaten ALL my spare time lately: A PS2, and a new version of Rosegarden.

Someone gave us a PS2 for Christmas. I played a couple of the games popular today and was pretty underwhelmed. We got a couple of educational games for Paul and he likes them a lot, but as far as I was concerned, I couldn’t see all the interest. I’d see people talking about various games but just figured they weren’t for me. And then, I found a game called ‘Call of Duty.’ The first time I tried it out, before I knew it, 2 hours had gone by. No matter what, every time I sit down with it thinking that I’d play it for 20 minutes or so, before I know it, hours melt away.

As for Rosegarden, it’s at least a much more productive endeavor. For those of you who don’t know, Rosegarden is a multitrack MIDI and audio recording program for Linux, much like Sonar is for Windows. I have Sonar, and it served me well until a couple of things came along. Firstly, I started the web page and blog, which meant that if I wanted to use Sonar, I’d have to take down the server, reboot into Windows, and then when I was done, boot back into Linux, and make sure the server came back up just fine. In the meantime, the web pages and blog would be unavailable, which is something I wanted to avoid if I could, out of my (probably unnecessary) determination to keep them running 24 hrs/day. It was a nuisance, to say the least, and one which conspired to prevent me from using Sonar very often. And if I didn’t use Sonar, that meant that I wasn’t writing songs, or working on recording demo tracks of ones I’d already written. It took so long to get everything set up to do anything that I just tended not to do that very often. And then I started having even more problems than usual with Windows. I had learned my lesson early on about allowing automatic updates, and had turned those off, but one day, my stupid Windows decided to install some “updates” anyway, and those resulted in me no longer having administrator privileges on my own computer. That was a real pain to fix, and really irritated me. A couple of months later, the same thing happened, and Microsoft, in its infinite wisdom, forced another update on me, which totally screwed up audio on my machine. So, no longer was it merely inconvenient to do anything with music on it, it was no longer very productive. I had pops and crackles, and my 2GB machine now claimed that it didn’t have enough memory to load the very same soundfonts I had been using for years. Since I once again didn’t have “Administrator” privileges on my own computer, thanks to Microsoft, fixing it was a real burden. With my bizarre work schedule, where my “spare time” might only be between 2 and 3am, it was just too much of a pain trying to do anything, and trying to keep up with Windows’ repeated insistence on breaking itself was just too much. I started looking for programs for Linux that would do everything I needed. I found quite a few that were available, and started trying them.

I found a couple of command-line multi-track recorders. Now, I’m not a HUGE fan of GUIs, but trying to control a multi-track recording program from the command-line is not something I really want to experience.

I found Ardour, which is a multitrack DAW (audio only). It was REALLY nice, and worked well, but I needed MIDI for composition (yes MIDI had spoiled me, and I found that trying to write with audio-only tools wasn’t working for me anymore). While it’s possible to sync it with various Linux MIDI sequencers, it was cumbersome, to say the least.

I found QTractor, which is audio and MIDI, and it worked really well, but it was lacking some features I really needed. It is written by one person, and he just can’t add features fast enough to make it a viable equivalent to Sonar.

I found Rosegarden, which did everything I needed, but had some bugs. A week or so ago, a new version came out so I decided to give it a try. I got it compiled and started testing it, and found out that the really irritating bugs had been squashed. I tried various things with it, and found out that I can have it running, plus a couple of software synths, in addition to the web server, blog program, the email server, and all the other things I run regularly (Firefox, Thunderbird, etc), and never use any virtual memory (swap file). Heck, on Windows, running Sonar with virtually nothing else running, and all other services turned off, my 2.6GHz machine was almost maxed out and using a LOT of virtual memory! I’m still fine-tuning the various parameters in the program, but so far, it looks very promising. I’m actually writing music again, which is what I’ve been wanting to do for months.

All of this means that time I was spending listening to what the thieves, crooks, and tyrants in government are doing is taken up by, in one case, something MUCH more productive, and in the other case, by something that is a total waste of time (but addictive). It’s something that I don’t think anyone needs me to point out, and most of the time it’s something that it seems like nobody really even cares about.

My blog is something that I use to simply get things off my chest. If I hear one of our so-called “representatives” say something totally inane, or just simply wrong, which covers 99.9% of everything they say, I would write in my blog to get it off my chest.

Other times, I would just write about things I had learned in my over 30 years of doing pro-audio and recording.

So, in a nutshell, I’ll keep writing in the blog, but I’m spending more time actually writing MUSIC than I have for almost a year, and playing with Rosegarden.

(For those of you unfamiliar with Linux audio applications, there are MANY – probably more than are available for WinXP. Some are excellent. Some are odd. Some are buggy. But, all are free. Linux itself is free, of course. You can put together a system consisting of the O/S itself, an Audio/MIDI recording program such as Rosegarden, a softsynth (QSynth/Fluidsynth), some VST/DSSI synths and effects plugins, a sampler (Linuxsampler), and whatever else you need, for a grand total of $0. And the hardware requirements are FAR less than what you would need for Windows.)

How Loud Should a Mix Be?

January 11, 2009

The answer, of course, is a resounding “it depends.” If you are talking about a live situation, then the answer is somewhere around “whatever is appropriate for the genre of music you are mixing.” You wouldn’t mix a traditional bluegrass band at 120db, and likewise, a heavy metal act at 70db would be pretty silly.

If you are talking about mixing in a studio, then the answer is a little different. There is no “one” answer, for several reasons.

First though, consider “Fletcher Munson.” In essence, what this means is that your ears have different sensitivities at different volumes – i.e. your ear’s frequency response changes with volume. At lower volume, your ear is most sensitive to frequencies which coincidentally are those which add intelligibility to speech (2KHz – 5KHz), and at low frequencies, your ear is not very sensitive at all. At high volume, your ear is much more sensitive to low frequencies, and to a somewhat lesser degree, the high frequencies. So, at 60dB, a mix will sound thin, with no low end. Turn that same mix up to 110dB, and it will have a ton of low end and much more high frequencies.

If you are thinking that there must be a volume range somewhere in the middle where the ear has a fairly “flat” frequency response, you would be right. And that mid-level volume, that is neither too loud nor too quiet, where your ear has the flatest frequency response is around 80dB. If you get your sounds and do your initial rough mixing at 80dB, then your mix will have a frequency response that should be fairly representative. If you get your sounds at 65dB, then when you turn your mix up, you will have WAYYYY too much low end! Likewise, if you start out at 110dB, when you turn it down, your mix is going to be thin and lifeless (sounds like a commercial for a hair care product, doesn’t it?).

However, every once in while, you need to vary the volume you are mixing at, again for a couple of reasons. One is to double check your mix at different volumes, but the other is because the ear develops “ear fatigue” when listening to something at the same level for long periods of time. Never listen for too long at any one volume – it will all start to sound the same, because your ear is developing fatigue. In fact, you should completely take a break fairly often. Leave the control room, walk outside (unless the sun is shining!), get a drink, play a video game. Just do something different for a while. Then, when you return to your mix, you will “have a fresh set of ears on.”

Along with checking your mix at different volume levels as you go, you also need to check it on different speakers. Try some near-fields for a while. If you have a different set of monitors, use those for a while. Some people even have an average (i.e.-crappy) car speaker, and will listen to a mono mix on that. Play it on a system in another room. All this is to verify that the mix holds up well on a variety of speakers, at different volumes, and in different rooms. Sometimes it will sound pretty good in every situation, but sometimes you will hear something on another system that just sounds somehow wrong. Go back to the mix, fix that problem, and try again.

Don’t drink alcohol before mixing!!! Alcohol does horrible things to your hearing, PLUS is makes you much more susceptible to hearing loss. The same thing holds true to a lesser degree with caffeine, but there are limits to what you need to put yourself through to mix, and caffeine deprivation is beyond that limit. Also, lack of sleep can affect your hearing.

Notice that I have said nothing about what the band has to say while all of this is going on. There is a very good reason for that – it’s a whole ‘nuther can o’ worms which would take many more paragraphs to go into.

Transformers . . . .

December 13, 2008

No, not “Robots in Disguise,” but rather the electronic kind.

Often misunderstood and mysterious, their basic function is to do as their name says: transform. Their primary purpose is to change something – a voltage or impedance. They can be very useful in a variety of ways, and are present in nearly every piece of electronic gear.

The most basic of transformers consists of two coils of wire wrapped around a core, usually of metal. The two wires are not connected to each other electrically. They interact due to the magnetic field they create and their close proxomity to each other. One other, but equally as useful result of their construction is that they will not pass any DC at all. Any DC applied to the “primary coil” is not induced into the “secondary coil.”

They change voltage in a ratio that is determined by the number of turns each of these coils has. For instance, if the primary has 10 turns, and the secondary has 100 turns, then 1VAC applied to the primary results in 10VAC signal on the secondary.  This would be a 1:10 winding ratio transformer, and could be used as a “step up transformer.”

At first glance, it would seem that this would be an amplifier, but there is a very important law that comes into play. That is the law which states that “energy is never created or lost, but only converted into something else.” Therefore, while the voltage may be “stepped up,” the current available at the secondary is reduced by the same ratio. If the transformer were 100% efficient, these figures would be exact, but nothing is 100% efficient. A little energy is “lost” (mostly converted to heat) through the resistance of the wires used to construct the transformer, and various magnetic losses – including stray magnetic fields that ‘escape’ from the transformer and can affect nearby circuitry).  A fairly standard efficiency rating for a transformer would be in the 85%-90% range.  So, any time you use a transformer, there is actually some loss involved, but the benefits usually outweight the losses.  A transformer is what is called a “passive device.”  It draws no external power, and can generate no power gain, and in fact, always results in a slight power loss.

What, you may ask, are transformers used for in the audio realm?

Probably the most common use of a transformer (besides the internals of your equipment) is the lowly direct box (usually called a “DI“).  It interfaces the high impedance (10K-12K ohms or so) of a guitar or bass to the low impedance (600 ohms to 1K ohm) of a mic preamp.  In this function, it is called an “impedance matching transformer.”  Without this, if you plugged a guitar or bass directly into a mic preamp, the relatively low impedance of the mic preamp would “load” the guitar’s pickups too much and would adversely affect frequency response.

Another use of transformers is in the splitter systems used to split the signal from a microphone and send it to two consoles (FOH and monitor).  In this case, the transformer has one primary winding and TWO secondaries, and each winding is isolated from every other winding.  In this design, the transformer is a 1:1 ratio – there is no voltage step up or down, and there is no impedance matching function.  This not only splits the signal, it also isolates the two consoles from each other so there is no interaction as can otherwise happen if the consoles have a particular preamp design.  If you simply used a “Y Cable” to split between the two consoles, the mic preamps of the two consoles would be in parallel, which would reduce the “load impedance” seen by the mic to one half, which may be low enough that it would begin to affect the response of the mic.

One thing to be aware of when using transformers is that they not only create magnetic fields, but they also will interact with magnetic fields that may be nearby.  For instance, if you set a DI on top of a bass amp, the magnetic field created by the bass amp’s power transformer may be ‘picked up’ by the DI’s transformer and cause a 60Hz hum in the low impedance (mic) output of the DI.  Simply move the DI.  I would suggest putting it on the floor.

The coils of which a transformer is made are not purely resistive.  They present what is called an “inductive load.

The design of a good quality transformer for a specific purpose is a skilled art that requires extensive knowledge of electronics, magnetics, metallurgy.  And, just like everything else, there are good quality transformers, and there are crappy transformers.  It can be the difference in design or materials.  Also, using a transformer for something other than it was designed for will result in poor results.



December 6, 2008

I see a lot of audio products for sale recently that have a tube inside, and claim to provide “that vintage sound.” The odd thing is that the “vintage” period they all talk about is a period where transisters were already widely used in pro-audio equipment.

Tubes have their place . . . . in guitar amps. Tubes add something to the sound, due partially to the way they distort, but primarily because they add second-order harmonics to the original signal, and this tends to sound “fuller” and “warmer.” These are good qualities in a guitar amp, but not necessarily in pro-audio gear. There is one major difference between guitar amps and pro-audio gear. A guitar amp is a part of the sound generation (production) process. It is the various distortions, non-linearities, and non-flat responses that make a guitar and guitar amp sound ‘good.’

Conversely, pro-audio gear should add nothing to the sound, the frequency responses should be flat, and the dynamics should follow the original exactly, with no compression (except what you purposely add) and no ‘ringing.’  It is sound RE-production.  It should neither add, nor take away, anything from the original – unless you purposely do that.

A guitar plugged directly into a hi-fidelity system sounds horrible, and likewise, playback of a mix through a guitar amp is even worse.

Beware of claims that a tube mic preamp will provide you with cleaner sound somehow. If used on someone’s vocal mic, it may indeed make it sound somehow “warmer.” But, don’t be fooled. What you are hearing is simply distortion (“distortion” technically is anything that changes the waveform).  You are using the tube mic preamp as an effect. It (as an effect) is not a normal part of the signal chain, and you wouldn’t want to use it on everything (just as you wouldn’t stick a chorus on everything in a mix).

Tubes would have some serious problems if used in pro-audio gear: size, weight, heat, reliability, durability, as well as sound quality.

There is a reason that all of the manufacturers went to solid state as soon as it was possible, and this is what made portable sound systems as we know them today possible. Don’t get sucked into the marketing hype. Certain companies hope that you buy into their marketing hype so that you’ll shell out a couple of thousand dollars for a tube mic preamp. There are an awful lot of six and seven figure consoles out there, (both recording and live consoles) such as Neve, SSL, Harrison, NeoTek, MCI, Soundcraft, Allen & Heath, Yamaha, that have been the top brands for the past 30-35 years, and not one of them has a tube preamp. Most of the major recordings you heard between the late 70’s and about 5 years ago were made using one of the first 5 in that list. (they were also being tracked on Studer, 3M, or MCI 24 track machines using (probably) Ampex 456) . THAT is the “vintage sound,” and no $2000 tube mic preamp is going to recreate that.

If you are used to using cheaper consoles, the internal mic preamps are probably not all that good quality-wise, and a $200 tube mic preamp is very likely a better preamp – NOT because it is a tube preamp, but simply because it is a $200 mic preamp.

If you really want to spend a couple of thousand to improve the sound of the vocals you record, I would suggest a Neumann U-87 might be money better spent.