Claims About Tube Amps Part 2: Using The Speaker Cabinet To Add Tone Colouring

Here we discuss ideas of how to modify the speaker cabinet in order to generate a more pleasing tone colouring for electric instruments, such as a guitar, that lack significant tone colouring.

Ultimate Guitar

Note to readers: the previous article was meant to generate thought and stimulate discussion, not to present fixed ideas. Significant improvement in any field only occurs when we loosen our hold on dogma and begin to try to pick out objectively what works and what doesn't and what has an effect and what is merely superfluous. While time-tested traditions have their place, there is no reason why these cannot exist alongside new innovations and progressive improvement. This is another such article.

Wood has been used for centuries in musical instruments - in everything from flutes to horns to violins to pianos to guitars. It has long been recognized for its ability to impart warm and pleasing overtones to just about any musically generated sound. The most significant use of wood is in the sound box of stringed instruments to amplify and project the sound coming off the strings. In electrical instruments such as guitars and more recently, violins, the sound box is missing and along with it the pleasing, woody overtones. Here it is suggested to re-introduce the pleasant sound colouring of wood further down the line in the speaker cabinet.

One might be tempted to ask, well, why not just amplify an acoustic instrument that already has a sound box? Anyone who's tried to do this knows how tricky it can be. The biggest problem is feedback and is the reason why solid-body electrics were introduced in the first place, and with considerable success. Probably the most effective way to amp an acoustic instrument is with a microphone. This is fine for the studio, but for live performances is less than ideal. Contact pick-ups tend to amplify string sounds - indeed, any sort of pick-up attached to the body often results in sounds that are considerably distorted and to not match the fullness of the unamplified instrument.

Another reason to do this is that the design of instrument sound-boxes is limited by the needs of the human player. They cannot be too large or bulky or otherwise awkwardly shaped. By contrast, the sound-box hanging off a speaker cabinet can made any size or shape we desire. Meanwhile, the instrument actually held in the musicians hands can be considerably reduced in size so that it is lighter, easier to handle and more durable. The body of a solid body electric guitar is really just for show and could potentially be removed and replaced with nothing more than an elongated block of wood. To date, only a few guitars have been manufactured with such a design. Meanwhile, because the sound power is much greater at the speakers, the wood can be made much thicker and therefore more durable.

Of course, speaker cabinets are normally already made of wood. In modern speaker design, however, at least in its theoretical ideal, the wood of the cabinet is not meant to be an active component, and indeed, very thick wood is typically used. The speaker cabinet is meant to approximate an infinite, rigid wall which separates the vibrations coming from the back of the speakers from those coming from the front, the rational being that the two waves are 180 degrees out-of-phase and therefore will destructively interfere. The simplest approximate realization of this ideal is that of the open-backed cabinets used in many guitar amplifiers. Many designs attempt to re-introduce the lost energy from the back of the speaker by inverting the phase and projecting it out the front.

The simplest design to take advantage of wood for sound colouring would take the existing speaker design and simply add an extra, wooden cabinet out the front with a hole in it - just like the sound box of a guitar which approximates a Helmholtz oscillator. A more radical idea would be to use the speakers to oscillate the wood directly, either with a weight for opposition, or another part of the speaker. Here we are using the wood as a sound board - similar to an acoustic guitar - instead of a thin, paper membrane. This could potentially necessitate a significant redesign of the speakers themselves. The membrane could be removed and since wood is much more rigid than paper, the amplitude of the speaker oscillations would need to be reduced. A piezo-electric electric speaker, which generates oscillations through a voltage applied across a crystal, for instance, might be preferable to the more common dynamic speaker which uses electro-magnetic coils.

22 comments sorted by best / new / date

    Much better, however still need to stop the spread of misinformation: One might be tempted to ask, well, why not just amplify an acoustic instrument that already has a sound box? Anyone who's tried to do this knows how tricky it can be Sure, it can be tricky, especially in the early twentieth century. Yet many would argue that Mr. Kaman (a fellow aerospace guy) solved the problem commercially in 1966 with the Ovation Roundback. These days electrically amplified acoustics of all kinds are commonplace, even in high-dB live performance settings like arena rock concerts. The artists (and/or their engineers) manage to pull off the trick without feedback most of the time--except when they want it or if they mess up. Combinations of mics, transducers, and pickups can achieve "natural sounds" well enough to meet the requirements for live settings. In-ear monitoring for the performers (vice foldbacks) can also help avoid feedback. This sounds to me like a solution looking for a problem. Not that something good couldnt come from it. But it seems unlikely that it could produce a more pure acoustic guitar sound than an actual acoustic guitar.
    okay, two things: 1. for your next article: get straight to your point. if you have an idea for a knew concept then stop talking about how and why things are the way they are for like 5 paragraphs. 2. have you build one of those cabinets/speakers? does it sound better, or is it even usable?
    Acoustic players wear my ass out with all of their esoteric bullshit. Taylor, Martin, Fishman, etc all do a dang fine job of replicating the raw acoustic sound electronically, and if pick attack/string noise is your problem, change your pick to either a much thinner design for strumming, or use a thick one CORRECTLY so it glides over the strings instead of "picking" at them like a chicken. As to amp designs, They're build the way they are to accommodate sound needs and something the road warriors like to call PORTABILITY and DURABILITY. A soft wood cabinet isn't going to hold up well in the back of the truck, nor will it hold up an amp head for too long. A Bubinga or Koa cabinet will give everybody back spasms. The biggest part of tone on an amp is knowing how to dial it in to get what you want. Marshall and Boogie do just fine by me.
    I don't understand how this is about tube amps, or where you are leading with this. I wanna guess that you're suggesting a new speaker cabinet design that places the speakers within the cabinet, rather than at the edge, so a sound hole can alter the sound further. Where tube amps go into this I don't know, same design could be applied to SS amps.
    I've heard plenty of performances by professional or accomplished musicians using amped acoustic guitars where the sound coming out of the speakers was crap. 'Combinations of mics, transducers, and pickups can achieve "natural sounds" ' "In-ear monitoring for the performers (vice foldbacks) can also help avoid feedback." Neither of which sound like particularly simple or easy-to-implement solutions. Electromagnetic pickups are nothing but a magnet and some coils. They are extremely reliable--they don't pick up a lot of string sounds and they don't generate much feedback, at least when rigidly mounted. The idea wasn't to get a natural acoustic guitar sound. It was actually to get something potentially much better. Remember, we're not limited by the needs of the human performer in regards to the size and shape of the sound-box.
    I typically use in-ear monitors even for open mic sessions with my electric guitar. The main reason is to prevent feedback from the vocal mic, yet still be able to hear ourselves. It's extremely simple--vocal mic goes to a Y-cable with one end going to the house PA and the other end going through a $30 pre-amp to the transmitter (tx). I use the other channel for guitar and tap it from the end of my effects chain into the tx. It takes about an extra 30 seconds of set-up time. In larger venues we can feed the transmitter from the mixing board. No big deal at all. Carvin makes a great system at a surprisingly reasonable price (I paid ~ $500 for 1 tx and 2 rx, but you can buy Shure, Sennheiser, etc. professional systems for tens of thousands of dollars if you want). The "combination of mics, transducers, and pickups" is also very simple; you can buy guitars already set up like that and you just dial in the blend you desire with the pre-amp. Just one instrument cable coming out of the guitar, like most electric guitars. No big deal at all and not complicated. Both of those solutions are certainly more simple than reinventing guitar speakers, not to mention building them. It doesn't matter how many performances you've heard where in your opinion the "sound coming out of the speakers was crap." There are many things that can cause that, and lots of them (if not most) could also sabotage the setup you propose, too. So many more influential variables are involved in this, and youre not addressing any of them. As for improving tone; I believe we already have achieved something much better than the natural acoustic guitar sound with the current spectrum of amplifiers, cabinets, cab micing options, and effects. Remember, they already aren't limited by the needs of the human performer. We can already do so much to the tone with high resolution digital manipulation, that I don't see further experimenting with cabinet design (i.e., size, shape, construction, materials) moving us much further. And certainly not coming up with anything that would be exclusive to the approach you suggest. Any tone your approach would produce could be emulated digitally, given high enough fidelity in the modeling. In fact, digital could be better because you could escape the inconsistencies inherent in wood from temperature and humidity. Paper speaker systems are capable of reproducing sound accurately enough that humans cannot tell the difference in blind tests. There just isn't a need for what you're proposing.
    You commented on my previous article: "There are some solid state solutions that sound almost as good (and sometimes better) than tubes, but sadly they either don't work well in real time or they cost as much or more than some tube amps." Now you're proposing to model an entire acoustic enclosure, even though (you claim) a simple electronic component can't be modelled in real time? Something doesn't wash here... I've tried a couple of acoustic modellers for electric guitar and they didn't even come close to the real thing. In fact they sounded worse (and less like an acoustic) than the simple clean tones coming from the guitar.
    I didn't claim modeling cannot be done it real time. I said that real-time modelers that are as good as or better than tubes could cost the same or more as tube amps. Since you didn't get it, I'll assume you don't have experience with what I was talking about, so I'll explain it more: Virtual Studio Technology instruments (VSTi) can take a raw guitar signal in your Digital Audio Workstation (DAW), (i.e., recording software for your computer) and make it sound as good as most tube amps (and better than some). But if you use it with a cheap interface and/or a slow computer, then you're probably going to have too much latency to play it in real time. If you have state of the art equipment and software you can pull it off, but that will cost you more than a low-end tube amp. The other option is hardware modelers. The better ones are usually rack units like the Fractal or the Elevenrack. Some would include Line 6 Vettas as well. These high-end hardware modelers generally start around $1,000 and go up from there. That still costs more than a lot of tube amps, and you still need amplification (PA or you can run some of them through tube or SS guitar amps). I wasn't talking about using the couple of acoustic modelers you've used. Those clearly would not be relevant to this conversation. I was talking about if you produce a good sound with your new cabinets, someone can build a high-resolution, high-fidelity model, and if mass-produced, it would cost a lot less than your cabinet concept. And if they do a good job, you wouldn't be able to tell which was which in a blind test.
    Sounds interesting. I'm not suggesting using the acoustic modellers I've tried--no one with an ear would bother with them--just pointing out that good acoustic modelling can be tricky, but it seems you already know that. On the other hand, it's pretty clear that whoever designed or coded these modellers didn't know what they were doing. Moreover, I'm not convinced that good modelling is as out-of-reach as you claim. As I pointed out in a previous post, the cheap $350 netbook I'm typing this on has a dual core, 64 bit, 1.66 GHz processor. Using a 40 KHz sampling rate, that leaves 83000 clock cycles to play with per sample. A talented programmer can do a lot with those 83000 cycles--I wouldn't even rule out limited first-principles modelling.
    PM me when you get around to buying an interface, a DAW, and some modeling VSTis or other plug-ins, learn how to use them and let me know how it works out for you. It was disappointing with my TASCAM interface and my Dell XPS dual core 2.2ghz with a 32 bit OS, and also with the Dell Studio with its dual core 2.4ghz on a 64 bit OS. I haven't tried it on the overclocked quad core HTPC I built; we only use that for home theater and video editing. BTW, before you even buy all that extra stuff for your netbook, you already could have bought a tube amp for the $350 you spent on the netbook. That's part of my point from your previous article.
    Ah, well here's the clever part. There's this wonderful thing called open source for people like me who have no money. Even if they haven't got it yet, they're bound to catch up sooner or later. I was considering giving it a go myself...
    That would be awesome if you do give it a go. I'd genuinely be interested in your results (PM me when/if you do). So what would you attack first, an accoustic sim for electric, an improvement over tube distortion for electric guitars, an emulation of what your cabinet ideas would theoretically sound like (or are you going to build one first?), or something else? Also are you looking to write a plug-in compliant with the specs for common DAW standards, a stand-alone product, or both? You may wish to consider writing it to be capable of DAW integration; if it produces a worthwhile tone, then you'd want to be able to record it efficiently (as part of an industry standard professional workflow). Again, if you actually do this, feel free to PM me about it.
    Seriously? As someone that has extensive experience with VSTs, I am calling BS. You can dial in good/ok clean sounds, but distorted sounds are very sterile and undynamic...and crunch sounds are just a joke. It seems that you are just on an anti-tube amp crusade.
    Calling BS on what? BTW, If you still have the same 4 amps you sometimes list in your sig, my tube amp is one of those. I've had it around 3 years and played it live and recorded with it. Just for clarification, when I'm saying VSTs can sound better than tube amps, I'm comparing the best VST against the crappiest tube amps. I've hardly played with guitar VSTs because I didn't really like the initial results or the latency, but I still got tones I liked better those I've gotten on a Fender Champ, a Valve King, or a even a Line 6 Spidervalve (all at music store volume and without pedals). I didn't say VSTs were better than all-tube tube amps like a VH4 or even a JVM. But I will say that hypothetically they could be; it's scary what teams can accomplish with high-fidelity modeling and simulation in most any field. The neat part for music is that hypothetically (given proper design), you can control it to emphasize whatever aspect of the real thing that sounds good to you while minimizing the counter-productive attributes. Im not saying that there is any VST currently on the market that does that--just that it is possible. Also in this article I mainly focused the VST discussion on the fact that if the TS (i.e., article author) got good results building an unconventional cabinet, then someone could emulate it for a lot less than what it would cost to build the physical cabinets.
    One thing that simulation on a general purpose computer gives you that an amp and a couple of distortion pedals, or even a digital effects console do not, is unparalleled versatility. If the algorithm is framed in general enough terms, you can add a new effect, any effect and by this I mean not limited to modelling existing effects, simply by coding a new module. One development I'm particularly interested in is the conversion of guitar string sounds into MIDI codecs. In this way a guitar can act much like a synthesizer keyboard. On Godin guitars this is done using a separate pick-up for each string, but in theory you can analyse the same result from only a single pick-up, via Fourier transform etc.
    Graphtec's approach is a Floyd Rose-style bridge with a piezo sensor in each saddle for the MIDI. Roland also has a MIDI pickup which has a sensor per string. The advantage of these systems is they are polyphonic. The only system I've seen that can convert output from a normal pickup to MIDI is the Sonuus G2M:
    But the problem with the Sonuus is it's monophonic; i.e., you can't do chords with it or even have one string ringing at all when you play a note on another string. Clearly polyphonic computer analysis is possible (even through one common pickup for all the strings) because Rocksmith does it. Although it's not perfect. It can't tell the difference what string you're playing on; it can only tell the note. Also it's easily fooled when you're playing chords (and doesn't always give you credit when you play a chord properly). I imagine that's indicative of the same challenges that keep manufacturers from marketing polyphonic MIDI hardware using conventional pickups yet.
    Acoustics is a science. Not one manufacturer today makes an enclosure that is engineered properly to efficiently and effectively harness the speaker's potential, especially not for a stage monitor or combo amp. Enclosures can make an incredible difference when it comes to sound. Shape, volume, duct size & position, and materials used all have large impacts on how a speaker sounds and performs...amplifiers and monitors don't capitalize on any of these, relying instead on circuitry, power and the actual speaker to do all the work. It's not wrong, it's just not as effective as it could be. Enclosures are meant to be abused and survive...they manage.
    good article, although as said earlier it has nothing to do with tube amps.