On Hardware Modelers

Before I speak directly on the topic at hand, I dislike the distinction between so-called hardware modelers and software modelers. Hardware units are in fact specialized DSP computers built into enclosures which are designed for ruggedness and integrated I/O and control. While they give the appearance of being “hardware,” it’s important to remember that in fact the actual signal processing that is running on all of the floor-based modelers is nothing more than software written in a language that the DSP can process efficiently. In fact, some DSPs (like the SHARC processors used by Line6 and Fractal Audio) can be programmed in C++, which allows them to be easily ported to a traditional software environment (aka a computer). Even if other floor units aren’t programmed in C++ or other higher level languages, they are still running software, specialized for execution on the processors in the floor box. Everything in digital modeling is software, it’s just a question of whether its operating environment is a discrete unit or a computer. Advantages of the former (discrete unit) include absolutely 100% known hardware, which improves the ability to efficiently streamline the code since the conditions under which it will be operating are perfectly standardized, and integrated DSP processors and other components which are usually designed specifically for audio and hence which have “low latency” as a goal from the ground up. To get a computer to the same level of latency and ease of connectivity requires a significantly higher investment, and software working in a computer has to be programmed robustly in order to ensure that it operates as flawlessly as possible across a virtually unlimited number of hardware configurations.

However, even with all that overhead accounted for, modern studio computers can be built to be extremely powerful, far more powerful than even the most advanced audio DSP currently in any production “hardware” unit (and I include the AxeFX in that assessment, their more than a year and a half old marketing claim that the AxeFX unit has more power “than many desktop computers” aside). This translates to flexibility that hardware units don’t have. I built my computer with an Intel Q9550 quad-core processor, which I’ve overclocked to 3.4ghz/core with fully tested rock-solid stability (on air, without overvolting anything to ensure as far as possible a long lifetime for my computer). I also have 8GB of RAM in this machine, although when recording I have access to only 4GB of it since I haven’t fully ported my recording software over to my Windows Vista Business 64-bit installation due to the industry lagging badly behind in driver and software support for the 64-bit operating environment. Still, 4GB of RAM, even given as much as 750MB of system overhead, is vastly more than any hardware unit, and with four extremely quick processors and that much memory combined with quality audio supporting hardware there isn’t really anything I can’t do if I want to, as far as my signal path goes. I can concurrently run however many instances of whichever VSTs in any routing configuration I like, flexibility that even the best hardware unit can’t touch. But the expense of that flexibility is quite high compared to hardware units, on par with or even above the cost of the AxeFX when you factor in your audio interface and a software of your choice. So when I say that I have invested in a software modeling rig, I want folks to know that I’m being quite serious – it is a non-trivial investment. That expense should be factored in when weighing the relative merits of hardware units, though some of it is mitigated by the fact that most of us already have studio computers, and if you’re not going to be working with audio professionally – and even if you are, really – you probably don’t need an exceptionally high-performance computer. In most cases if you’ve already got a general usage computer, your big expense to make it capable of amp modeling will be to get a set of decently flat speakers and an audio interface with low latency and good converters. I’m very much in love with the IK Multimedia StompIO for its excellent input section, which I’ve spoken of elsewhere, and they’ve got another unit coming out soon as well, the StealthPedal, which will be bus-powered but midi-operable; the competing Rig Kontrol units from Native unfortunately can’t be used as midi controllers, so they only work with Guitar Rig 3 – IK seems to be gunning pretty hard for Native’s share with StealthPedal, since it basically corrects the single biggest flaw and makes the expression pedal that the interface is built into work in any software you want. I don’t mean to push IK Multimedia so strongly here just for the sake of it, but when StealthPedal comes out, if Native doesn’t fire back and hard IK might have the $200 and above controller&audio interface market pretty well locked up.

With those considerations aside:

My feeling is that, right now, the only hardware unit that can fully compete with software for sound quality, flexibility, and processing power is the AxeFX. The PodX3 is sort of grandfathered into the realm of serious competition because Line6 doesn’t seem to reserve special modeling for the plugins, they just have their modeling cross-platform in that way. The X3 is a powerful unit and Line6 are the oldest hands at digital modeling in the business (second oldest, by the way, being IK Multimedia, who released Amplitube for ProTools in 2001), but it has been my experience that while it can be and has been used professionally by very well respected studio pros, it is not at the top of the heap. This isn’t to say it isn’t a worthy unit, but the AxeFX opens up quite a gap indeed in the hardware realm, and I have never heard any variety of POD or Line6 software accomplish what I can with the software that I use. To me they seem to be somewhat handicapped by their connection to the hardware unit – since they’re cross-platform in their approach, much like videogames that are trying to do a 100% port both on the PC and consoles, they tend to be limited by the weaker of the two systems and in this case the hardware can’t keep up. Line6 does buy and use very nice SHARC DSPs, but AxeFX outclasses with with even higher end SHARC processors (and a great deal of memory, and a well designed, noise-reduction technology implementing input section that behaves like a guitar amp’s input and can be boosted as well as used with overdrives/fuzzes/distortion, and frankly much better DSP algorithms – that is, amps and effects, etc.)… Line6 ensures studio space by making it easier to reamp than with other hardware modelers, including easy to use SPDIF connectivity, and the fact that – though this may have been a limited promotion – for awhile, Pod X3 owners were able to get PodFarm for free to work in the fully digital environment without ever needing to break out of the box at all.

None of which is to say that the POD is a bad unit, by the way. Some very good tones have been got from it, and if the proof’s in the pudding, then the pudding has been a hit product and people are eating it everywhere, sometimes without even knowing that it’s pudding and thinking that it’s some kind of fancy desert that you can’t get just anywhere. Or, to unpack the analogy, as my grandfather used to tell me “don’t question success,” which the Line6 POD in its many incarnations has certainly had. And I really do like the fact that with the PodX3 and X3 Live, they’ve taken on the idea of an all-in-one, low cost studio – it’s a real equalizer for folks who can’t afford a lot of outboard gear and expensive hardware. No, you won’t track the next massive release on a major label by hooking some mics up to your POD X3, but if you want to cut demos and then edit them to achieve results that might have been difficult for an amateur studio five years ago, now you can – and they take the same approach with their PodFarm integrated hardware units, built-in mic pres and project studio oriented effects algorithms beyond just guitar and bass effects. Is it my first pick for a unit? No, probably not, but it should be taken seriously and I hope I’ve explained well enough why that is.

The other hardware units can achieve good results in their own ways when taken for what they are and worked with extensively enough to truly understand them, and my colleague Will Chen (the editor of the http://www.FrugalGuitarist.com zine) has reviewed and used them to good results. Another very good site that you guys should all check out if this is a topic that interests you is forum.thestompbox.net, a site run by a true modeler aficionado who has done comparative tests and the whole works (I even have an article or two hosted over there :)).

I used a Zoom G9.2tt unit for about a year and enjoyed it immensely, I think it is a quite underrated unit on the whole and that it competes with the PodX3 tonally, especially for clean sounds and high gain sounds. Its effects complement is very good, and it includes a number of useful models of boutique and classic stompboxes, including such lofty names as the Klon Centaur, Matchless HotBox, and others. I love the Big Muff model, it really nails that fat, saturated fuzzy tone, and the Zoom Extreme Distortion is a model of the much sought-after Zoom Tri-Metal from their short-lived but very well regarded analog series. Zoom seems to specialize in trying to provide not just the “go to” standards of amplifier and effects, but also some great sounding but somewhat farther out of the way models, though Line6’s shotgun approach of amp quantity saturation has led to them having a selection that overlaps much of what used to be Zoom’s unique territory. I also appreciated the “Z-pedal” vertical & horizontal movement expression pedal that it had (in addition to another traditional up/down expression pedal), and the effects that took advantage of its two-axis movement were really neat and included things stereo effects that did a great job of simulating depth-of-field, like the Tornado filter effect that made it sound like your tone was swirling around across the soundstage, with the center controlled by your foot. Creative stuff that makes Zoom a constant contender despite their relative underdog status (they also design their own DSP chips in-house, while everyone else uses third-party DSPs; not a negative for the other guys any more than code prefers to run on an AMD versus an Intel processor, it’s just kind of cool to know that they’re so into DSP programming they felt like they could build one that would meet all their needs precisely).

It’s been a long time since I had experience with the ToneLab, but there is a very big “ToneLab Lounge” thread over at the Harmony Central Effects subforum, and quite a lot of community content on the Tonelab over at forum.thestompbox.net, so you’ll not lack for info there. What I can say about it is that I really like Korg’s modeling, and it forms the basis for Vox’s modeling, but that given the new generation of Vox modeling amps which have much improved amp models, I think the ToneLab is really in need of a new product release to get it up to speed with their amp models – their “Valve Reactor” technology is nice and helps in the emulation of the all-important power amp, but it’s a relatively old technology itself and I think they could improve it as well, perhaps make it more adjustable to the user. However, its flexibility and separately adjusted, knob-based Pedal, Amplifier, Cabinet, Modulation, Delay, and Reverb sections make it very easy to dial in a tone quickly, and it succeeds where some other modelers fail in emulating “crunch” realistically. I’ve never met someone who owned the ToneLab and just hated it, while I have met people with strong negative feelings about other modelers, so that says something to me. I also love my Vox DA5 and it’s my take-anywhere, rough and tumble tone machine, even though as far as I can tell it’s built off of their 2005 modeling technology which is a generation out of date!

I have a soft spot for the Boss modelers just because I really respect what they’re trying to do with them. They’re aiming to carve out a spot entirely separate from Line6 for a professional-quality, recording-ready but extremely rugged modeler that can transition from the stage to the studio seamlessly. I think that up to the GT-8, they had succeeded in everywhere except where it counts: the amp models! They sounded, well, frankly not very good. However, the introduction of the GT-10 has seen a big improvement in the quality of their COSM amp modeling, and the other areas where they had already done a great job are still in full effect, so I think this unit is a real contender for “floor-based hardware modeler champ.” It even has an “EZ tone” feature which is sort of like a wizard to let you quickly and automatically set up a tone by entering in a few things about the sound you’re going for. Of course there’s nothing stopping you from using its generated sound as a starting point to make your own. And while the Vox DA5 is rocking my world now, my first love in portable amps was the Roland Microcube 🙂

I am not a fan of Digitech’s modeling, to be honest. Even though I like the concept of their GNX workstations and the latest GSP rack unit, to my ears there is a constant compressed quality to the tone and a buzzy timbre that is difficult to dial out even with a lot of tweaking. I also don’t like that the RP-series, even the higher end ones, don’t allow enough parameter tweaking to really get the job done, though that complaint obviously doesn’t carry over to their much more accessible workstation/rack units. Some people get on quite well with Digitech’s modeling, but it just has never clicked with me. The only Digitech gear I’ve gotten along with are the Metal Master as a silly, uncompromising one-trick pony high gain distortion pedal; the Bad Monkey, which is just an excellent overdrive for not much money; the HarmonyMan, which does great multi-part guitar harmonies with extremely fast tracking and playing-based chord progression detection for intelligent harmonizing on the fly; and the JamMan for providing competitive pressure on other looper manufacturers and because it’s a pretty good product on its own. Apart from the Metal Master, which is a digital distortion pedal but which doesn’t to my ears sound anything like what you can get from their modelers, I just don’t get along with Digitech’s idea of what digital modeling of amp and pedal distortion should sound like.

Keep an eye out for the Rocktron Utopia units, as well – they’ve taken a different approach in their modeling, drawn from their experience making high-quality rack-mounted guitar preamps with all the trimmings. Rather than a bevy of amp models, they have four “general” amps which can be highly tweaked (including power amp characteristics), and a variety of effects which are intended not to overpower but instead to complement your playing.


2 Responses to On Hardware Modelers

  1. tome says:

    I bookmarked your blog, thanks for sharing this very interesting post

  2. DigitalMan says:

    Another great post! I’ve never read a comparison of “hardware” and software modelers before, and it was quite interesting. I definitely agree with you on the fizzy and over compressed tone of the digitech stuff. The Zoom unit sounds interesting, I may give it a try.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: