Love. Photo (CC-BY acidpix.

Love. Photo (CC-BY acidpix.

We’re not so much in the habit of posting jams on CDM, but this one is especially nice – even through the freak-out visuals. And it comes from friends – Nigel Mullaney, with recording and engineering by Ian Boddy.

Seen in the film:
Elektron Analog4 keys
Elektron Octatrak
Elektron Machinedrum
KORG volca series

Look closely through that shaky video, and you might get some clue as to why people love hardware. There’s plenty of reason not to go the hardware route: computers alone still offer more power, more flexibility, and more sound for your buck.

But have a look at this hardware, and ask yourself – how much software, even in combination with controllers, offers this kind of control? Ableton Push, Maschine, and the like are more exceptions than the rule. Go one step further, and the design of even software/hardware combinations remains fundamentally different than hardware. Hardware is all about constraints: even with more powerful DSP innards and the like, there are restrictions on design. There’s a limit to the number of physical controls you can fit (or afford to manufacture); physical controls themselves can’t have unlimited functions.

All of this results in something that, not only about being tangible, is about a different form of design. Hardware is a different medium, and produces different results. It just happens that those results are particularly well-suited to immediate jamming, where computers often fail.

But given that’s true, I can imagine software rigs that could begin to find some of the same appeal. Some intrepid DIYers, working with tools like Max for Live, have built very specific control/sound combinations for devices like the Novation Launchpad or APC. And even if hardware is a different sort of medium, it is possible at least to take the principles of one-to-one control and restrained choice and apply it to other media.

In short, those ideas would be:
1. One-to-one hands-on control.
2. Design that can incorporate performative gestures and muscle memory.
3. Limited choices in sonic potential that form a specific personality of the instrument.
4. The ability to work without looking at a computer screen.

Hardware can add some additional properties that software can’t:
1. Rigorous timing and low latency (at least, in principle).
2. Instant-on capability. (Getting better with computers, but still not quite the same). Certainly –
3. Immediate music making, without distractions. (Think of how long it takes to load a typical DAW.)
4. Plug-and-play connectivity.

And, just generally, I think a lot of people are eager to spend some time away from the computer and all its psychological associations. But even looking at the bottom four, it is possible to coax computer software into doing the same – and there’s plenty of potential for embedded systems. And the top four are interesting enough that, for some tasks – jamming certainly included – they could be worth exploiting.

(Thanks to Tim Exile, who inspired some of these thoughts. Curious what you think, too.)

Updated: A beautiful quote from our friend Marc Doty. If the above is more practical, this deals with the more poetic reasons hardware works (and is worth considering in any musical design):

Hardware is also attractive because we are the result of evolution, and are genetically accustomed to engaging in physical tasks, manipulating objects, and interacting with our environment for specific outcomes. Our creativity is largely connected to our bodies (on average), and creative tasks are often successful as a result of our unique physiology and psychology.
There is also the aesthetic experience of physical interaction with a creative device. How the device looks, feels, and responds to our interactions is (in the right circumstances) fuel for our creativity.
Lastly, our own physiological quirks and inclinations lend uniqueness to our output when we, as individuals, work with objects.

And lastly, via Twitter – a classic Brian Eno insight.

“muscles..represent several million years of accumulated finesse. …”

Of course, that means a consistently-mapped controller (from monome to Maschine) will also suffice.

  • Jesús Gallego

    I don’t think it’s about hardware or software “per se” but certain hardware and certain software that hit the right spot for me, even regarding jamming. I adore playing Samplr on my iPad or Figure on my iPhone and otherwise despise editing something like a Roland JV synth or some Korg workstation which totally kill the buzz for me. But then again, there are certain nuances which I can only achieve using analog hardware due to natural overdriven circuits, where you can spend hours just finding “sweet spots” while turning knobs and sliders. So I guess in the end it’s not so much about choosing, I won’t make music without either hardware or software if I can afford it.

    • Peter Kirn

      Well said, absolutely. It’s really about design. But that’s part of why I was interested in investigating the underlying motivations.

    • Jesús Gallego

      In design terms it’s called affordances

  • Mister36

    Some very good points about why people use and love using hardware. Guitars, microphones, and MIDI controllers being the only hardware I currently use, I often miss it. The immediacy and limitations of hardware. Composing and performing with software solutions can be very heavy on the mind because of the seeming lack of limitations. Aside from more niche ideas, if you think of it, you can probably achieve it with software, but with hardware, you know what can be done and certain things are not even conceivable. Not to say experimentation and pushing those limitations isn’t possible or encouraged, of course.

    For me, however, using hardware (alone at least) is not feasible, financially or functionally. (I do so often wish I could afford some of that stuff though!)

    • Peter Kirn

      Financially? Cough, MeeBlip. ;) More on that this week.

      And I will say, taking off my marketing hat and putting on my musician hat, I’ve played a number of gigs with just a MeeBlip and laptop — it was already interesting enough to me to add that one hardware element.

    • Mister36

      That’s fair enough and cool in itself. But if I was going to spend money on hardware, it wouldn’t just be for the sake of adding hardware to my setup. It would also need to contribute to the setup in a way that other things don’t already do and that kind of functionality is what I can’t afford.

      Certainly not knocking the MeeBlip or other more affordable synths and devices (e.g., Volcas, Microbrute, Rocket etc.) though! They’re just not for me at the moment.
      I think if I didn’t use Maschine so heavily, I’d feel like I needed hardware a lot more than I do, just for that hardware experience if not for the sounds and functions.

  • Sequadion

    One of my friends and I are currently experimenting with building a standalone instrument based on the Raspberry Pi and Pure Data. Plenty of potential in embedded systems, I would say.

    • Peter Kirn

      Yeah, absolutely interested in this. And while the Pi is at the low-end power wise, there are other options.

    • Servando Barreiro

      Do I know him?, hehehe, we should unite..

    • Sequadion

      Unlikely that you know him, but looks like this is another example of great minds thinking alike. :) Looks like you are somewhat ahead of us, looks very promising. If you are serious about collaborating, drop me a line! or

  • Mrst

    I actually like having control exclusively over what I need in a live situation vs. the “knob per function” approach. I use a combination of APC40 + iPad Lemur, which handles only the parameters I use live (return fx parameters, some insert fx, global EQ and some key synthesis parameters for whatever track I’m currently playing). If I had every option possible in front of me without any disctintion of what’s important and what isn’t, playing live would be a nightmare (i.e. a huge FM patch with complex fx routing where only the filter envelope decay in a particular range does something nice and twenty other parameters just make the patch explode). In that regard, I preffer a (customizable) controller rather than eleven billion knobs that represent every option possible.

  • Samuele Cornell

    One big advantage of hardware is that it retains value ;
    A computer is a multi-purpose machine , it is versatile , you can add multiple controllers , it is an indispensable tool nowadays ; but it does not retain value .

    • PaulDavisTheFirst

      I think this post is about why it retains value …

  • Pete K

    I think this is the main difference. Hardware is better for jamming and “playing” music …but when it comes to recording music software wins by a mile.
    Syncing, recording, and resetting all that hardware is a massive pain in the ass.
    Anyone who say hardware is more immediate has never tried to switch from one project to another using a midi studio with 5+ instruments. it is trivial in software

    • Peter Kirn

      Definitely – I think actually that’s part of the shift. Software has become really elegant for production work, which oddly, means it’s also easier to integrate hardware with it. And hardware retains personality and live jamming potential…

  • Ben Heymer

    The hardware/software debate is as simple as this: If you use a computer all day for work, using a computer to make music feels like work. Unless your job is operating the analog control panel of a historic submarine, using hardware is a totally unique experience and is therefore more fun.

    • Martin Yam Møller


    • Foosnark

      That’s very subjective. I can work 40 hours a week developing engineering software, but spending a few hours writing a VST plugin on a Friday night doesn’t feel like work. And making music never feels like work, computer or not.

      Actually I take it back. When I was a taiko drummer, some gigs and a lot of rehearsals felt like work, without a computer in sight.

      Maybe because I’m a gamer, and because I used to do most of my social interaction online — but I really don’t associate “computer” with “work.” Toolbox/toybox of incredible possibilities, more like.

    • schlauchmilch

      You got a point there. I think electronic music was much different 30 years ago or so when you didn’t do digital/electronic things all day. Nowadays, “exploring sonic possibilities” on a screen can feel quite close to writing the next TPS report.

    • heinrichz

      Exactly true, but that’s why you would need to use a controller that is totally integrated with the software from the get go (only Maschine so far) and allows you to use the computer just for additional display functions and file management..which are by all means superior to any hardware. What’s killing the intuitive workflow with the computer is the dependency of the mouse which won’t allow us to establish that ‘muscle memory’ that Eno is talking about and without that we don’t have a musical instrument we can practice. It then all remains too cerebral and non physical, using too much of one side of the brain.

    • pacyderm

      This is exactly how I feel. I just can’t get into writing music on a computer at home after writing code for 8 hours at work.

    • cooptrol

      In this regard, in my case I spend hours editing audio and making music for clients in Ableton. When I want to make music for my artistic projects, I don’t wan’t to grab a mouse whatsoever.

  • beatbbeat

    I think a part of the problem comes from the fact that big companies/manufacturers think of controllers as cheap musical toys (Push and some others are the exceptions). I would be very curious to see some controllers built with the same budget as, let’s say, the monophonic (!) pro 2 for instance, hitting the market at 2K$.

  • mercury

    I think it is about time for a new plug-in standard. Whether it is VST4 or something else, basically, it should have wireless OSC built in to start. All synths should be quickly adaptable to a generic MIDI interface. Instead of having to program every single knob, you should be able to open up a synth and logically map the important parameters with decent defaults setup automatically. It just seems ridiculous that most of the MIDI keyboards have very weak syncing to VSTs…it essentially makes the user automatically use the mouse more than needed. If my Keylab61 automatically linked it’s controls in a logical fashion to each VST, I would prob treat them more like hardware.

    • PaulDavisTheFirst

      Me: “He who doesn’t understand OSC is doomed to imagine that it is suitable for almost any control problem.”

      Apple, circa 2002: “We’ve just done our plugin API and we’re not interested in anybody else’s attempts at answering these questions”.

      Steinberg, circa 2002: “This talk of a new standard plugin API is nice but if anyone thinks we’re giving up control of the VST specification, they are dreaming”.

    • mercury

      i don’t think osc can solve everything but it’s a little ridiculous it’s not built in to ableton.

      all i have noticed is that with hardware i can make music right away and tweak knobs and get results. with software synths hooked to a midi controller, i always waste a lot of time setting up the hardware controller. it kills the creativity. and it’s not that i can’t put in the extra time, it’s just that music comes to me in quick pulses, if i can’t recreate them quickly, the magic is often gone.

    • PaulDavisTheFirst

      More seriously, the control API problem was largely solved by Euphonix, who designed a complex but fairly elegant protocol that does what you want, and more.

      But then Euphonix were purchased by Digidesign/Avid, and their interest in collaborating with others who wanted to use the protocol fell away.

  • wingo shackleford

    This is relevant to my interests! Seriously… I was losing my mind for a long time trying to nail down how to perform with just software, and it was driving me nuts. I kept trying different approaches with Ableton, from just doing straight DJ-style to pure from-scratch improv, and everything in-between. Different controllers, what have you.

    Then I got enchanted by the Microbrute recetly, and it suddenly hit me that I was much more at ease improvising on a relatively simple instrument where I understood the limitations. I grabbed some Volcas, and the rest is history. I’m going full hardware for playing live now. As cliche as it is, it really can be counterproductive to have too many options.

    The Volca beats, in particular is so great. Very limited sound set, but you know what you’ve got, and the design begs you to build stuff on the fly and tweak it. And you can sync it to everything! I always used to view the ‘just hardware’ purist guys as maniacs, but now I can totally relate. I may have crossed over to the dark side already… I’m going to still use Ableton for recording and production, but I’m really digging this live with just hardware business. Working up a new set for an upcoming gig right now, and it’s just fun, and not a chore like it used to be.

  • papernoise

    There’s a couple of great points here! Personally I’m having a hard time with computers when making music for two reasons: first of all it reminds me of work (as has been pointed out by other people here) and second I find that the “hardwiring” of control and function on a hardware device makes it easier to use than a computer based setup. Basically the big advantage and disadvantage of a computer is its flexibility, you can do whatever you want with it, but that comes to a cost: you need to first work out what you want to do. Flexibility takes away many constraints, and that’s what makes us become creative sometimes.
    Plus a tool is best when it becomes invisible, when you don’t have to thing about using it… you just use it. Hardware seems to be easier to take to that lever where you don’t have to think about it anymore, at least when it comes to making music live. Recording is a different thing of course.

    Hardware already defines how you can interact with it, this makes things easier for me.
    So I think good hardware design is all about figuring out how you would want to play the instrument, and making the controls as intuitive, expressive and usable as possible. Good hardware gives you exactly this. MIDI controllers just give you knobs, it’s then your task to make them useable and expressive…

    One thing I want to try in the near future is to build myself a dedicated controller for the computer. I think I would not want a controller that let’s you map anything how you want, I’d prefer something I design for a certain purpose and where I decide all the feaures beforehand. Again, it’s all about limitations and constraints.
    This can probably replicate some of the features I like about hardware. Though of course as the article says, hardware also is something you just turn on and start jamming with.

  • Yanakyl

    One thing often overlooked is the fact when you play with software you usually look at a luminous screen. I might be wrong but I feel like staring at those give me more of an analytical mindset, I need to stop myself a second to realise I’m building a track as if I was making it with legos.
    Whereas jamming with a couple of machines I move more and I’m more in the feel, I’ll then need to detach myself a sec to think, ok where do I want to go with that now.
    It’s a bit more mixed in reality but those are tendencies I’ve noticed.
    Don’t know if it make sense to any of you?

  • Foosnark

    To me there are two main creative mindsets in music: playing and building. Both are valid arts.

    A DAW, keyboard, mouse, and screen are not very good for playing. But to me, it’s the best thing for building.

    I can pick up one of my hand drums, Wavedrum, or a bass guitar, or sometimes even just a shaker, and lose myself in the moment. (Sometimes a very long moment…) But I will not make music that I want to hear again later, or necessarily share with anyone. Those moments are gone. Sometimes with acoustic drums, it’s a particular combination of temperature and humidity and the acoustics of where I’m playing, as well as my mood and inspirations, and I won’t necessarily get the same sound again.

    To me Maschine is a bit of a fusion of playing and building. Not necessarily the best at either, but good at both. I can easily lose myself just finger drumming. Usually that will disappear into the ether like my acoustic playing, but sometimes I’ll record a bit, tweak it, and start building a track around it.

    But anyway, I think the general idea that people like hardware for the immediacy, rather than for sound necessarily, is important.

  • Gllrn

    “Create Analog Music”

  • heinrichz

    ‘Of course, that means a consistently-mapped controller (from monome to Maschine) will also suffice.’
    And a well designed System like Maschine does not even need mapping in the first place. And then Midi control mapping with the Controller Editor is a breeze. Check out the Ableton V2 template with Maschine Studio or MK2 and you’ll find that it makes for an excellent Live controller as well.
    ‘Rigorous timing and low latency (at least, in principle).’
    Definitely just in principle and musically irrelevant with today’s computers.

  • PaulDavisTheFirst

    To me, this conversation was over in 2009. When I saw Randy Jones stuff, it was clear to me that the problems with making music with software is threefold:

    * physical ergonomics
    * physical properties of the “controller”
    * rich enough synthesis model

    The overwhelming majority of synthesis models used by people who are into “controllers” are completely inadequate to present anything remotely close to the sophistication inherent in even the simplest analog non-electronic instruments. The jam video above is awesomely great (really!), but it is just knob twiddling, not playing an instrument. The subtle non-linearities are not there, and neither is the requirement to develop any deep kind of physical skill (e.g. the kind of skill that some people will never master).

    The same goes for the physical properties of the controllers: low sample rate touch devices, trivial control mechanisms that map individual parameters with linear relationships, etc. etc. You can make some great grooves with this sort of thing, and everyone can have a lot of fun. But terminologically and semiotically, it isn’t related to playing instruments.

    • Taylor

      I’m not sure I understand your point. It seems you’re making a semantic argument about what constitutes an instrument, along with some value judgements about controller’s synthesis methods being “completely inadequate” compared to even “the simplest” acoustic instrument… like a Kazoo?

    • PaulDavisTheFirst

      Take a look at the physics of a kazoo, and how the sound it generates varies under the influence of faster air speed, more air volume, and other more subtle qualities that the player can control.

      That’s more acoustic complexity right there than you can get from a lot of controller+non-physically-modelled-synth setups. Not all, but most.

      And yes, I am judging what an instrument is. If you want to talk about a different kind of thing, which lacks most of the properties of “an instrument”, but nevertheless can still be used to control sound generation, then I’d suggest a new term … maybe … “controller” ….

    • Taylor

      You seem to have conflated physical complexity and interface complexity with musical value.

      And you also have an overly narrow view of what a musical instrument is, which denies controllerists, modular synthesists, etc. the privilege of saying they play a musical instrument. “Just knob-twiddling, not playing an instrument.” Very dismissive of the passions of others.

    • PaulDavisTheFirst

      Knob-twiddling is an excellent way to make (some kinds of) music. I love many knob twiddlers and their output. I might even consider myself as a knob twiddler.

      I just don’t happen to conflate that sort of activity with what happens when someone has spent their 1,000 – 10,000 hours mastering a device which can translate imperceptible motions and tensions into acoustic transformations.

      They are both entirely valid forms of making music. I think it is wonderful that the knob twiddling age is upon us. I just think that maintaining the distinction between these two kinds of activities is useful, respectful and accurate.

      BTW, modular synthesis has absolutely nothing whatsoever to do with this, other than the fact that the synthesis models typically in use in most modulars are not particular sensitive or nonlinear in their response to most of their parameters.

      I’m 50 years old. I grew up listening to Tangerine Dream, Klaus Schulze, Michael Hoenig and many others of the Berlin school “twiddle knobs” and use modular synthesis to break open whole new sonic worlds. I still love (some of) that music. I just don’t have any need to conflate what they or the current generations of controller-ists do with “playing an instrument”. But making music? Absolutely.

    • Taylor

      Modular synthesis is relevant because modulars are instruments (are you denying that?) that generally don’t require a high degree of manual dexterity to operate well (and one of your requirements was “deep physical skill”). Also, a nonlinear response is mathematical jargon you will not find in any commonly held definition of a musical instrument. Nor will you find any requirement for the underlying physical complexity of the instrument’s operation.

      “I just think that maintaining the distinction between these two kinds of activities is useful, respectful and accurate.”

      Its perfectly fine to draw some sort of distinction between instruments sometimes (perhaps rarely) played with virtuosity and others. But this should not involve deciding one type is not a musical instrument. Besides, just a little searching will lead you to some impressive (and perhaps virtuosic) use of controllers.

      By any reasonable definition of “musical instrument,” such as the one you can find on Wikipedia, your implied definition is overly narrow.

      Furthermore according to your view one can make music without using musical instruments, which is a very odd proposition!

    • PaulDavisTheFirst

      Furthermore, according to your view, one can make music without using musical instruments, which is a very odd proposition!

      You don’t seem to be aware of the entire electro-acoustic tradition which started in the 1940′s (maybe even the 1930′s), or the earliest days of computer-generated music, which involved no instruments of any kind (not even knobs). Very definitely music, very definitely no instruments since there wasn’t even any real time interaction involved in its construction. One of my favorite (late) “commercial” examples of this is “Computer Experiments I” by Synergy (Larry Fast, who did lots of “commercial” work with Peter Gabriel and others), which is just a PDP assembler program. Nice stuff.

      Non-linear is a handy phrase to describe things that work in more complex ways that “when I turn this knob, it gets more <X>”. Blowing into a trumpet or french horn or trombone doesn’t work that way, and horn players know it. Ditto for reed instruments. Hitting the keys on a piano doesn’t just make things louder, it alters the harmonics in different ways depending on the gauge of the strings, and the precise nature of the effect depends on other physical properties of the piano. Tapping the head of a djembe or duhmbek drum lightly produces a sound which is almost unrelated to what happens when you actually “thump” it. These are examples of commonly known subtleties of acoustic instruments – commonly known to people who play them – and not just “mathematical jargon”.

      I love music and the people who make it. I see the need to equate controllerism with instrumental performance as a bit of a wierd inferiority complex. Is there something wrong with attempting to distinguish between subtle differences? There’s nothing wrong with using controllers and related equipment to make music. The result is still good or bad independent of how it was created. But the process, the skillset, the possibilities .. these differ in important ways from playing acoustic instruments, and I think that our culture, our language and our music are all richer for retaining the differences in the terminology we use about it.

    • Taylor

      As a software engineer, I’m well aware that you can write code to use the computer AS a musical instrument to generate music. And as someone who has written DSP, graphics, and other numerical code, I’m also quite familiar with nonlinearity.

      And as someone who has played in rock bands, I’m also aware that if I told the other guitarist that my amp is highly nonlinear, he’d look puzzled.

      I hate to resort to dictionary definitions, but here’s the definition of instrument from the dictionary on my mac:

      “1 a tool or implement, esp. one for delicate or scientific work: a surgical instrument | writing instruments.”

      “3 (also musical instrument)an object or device for producing musical sounds: a percussion instrument.”

      There’s nothing in there about real-time performance, nonlinearity or physical complexity. There’s nothing about whether the device is a grid of buttons, six strings, or a computer. There’s nothing about whether it can be played with virtuosity.

      Its perfectly fine to make subtle distinctions between instruments. We could talk about how acoustic instruments tend to have more complex underlying synthesis models than electronic instruments.

      However, it is not fine to blatantly misuse terminology.

    • PaulDavisTheFirst

      You’re the one who is now quibbling. I never claimed that “nonlinear” is a term that would be used or understood by musicians – I referred to phenomena that they are familiar with that would often be described by physics geeks as nonlinear.

      You also stated that it was a very odd proposition to claim that one could make music without instruments – I gave you examples where people did that – Larry Fast would not have claimed that the PDP-11 he had access to was an instrument, and if he even went that far, he would certainly have declaimed the idea that he was “playing” it. The “music concréte” created on tape machines was also generally done without calling the tape machine an instrument, or imagining that “playing the tape machine” was a useful concept by analogy with acoustic instruments.

      Fine, you like the idea of using the term “instrument” for anything that, even if it requires a D/A converter in the signal chain, can be used to create pressure waves in the air, and that anybody who uses such a device is “playing” that instrument.

      I think this is an unhelpful way to think about and grapple with the differences (and similarities) between digital, computer-based music creation and playing acoustic instruments. Clearly, you don’t agree. Not a problem.

    • Taylor

      Wikipedia has a nice description:

      “In principle, any object that produces sound can be a musical instrument—it is through purpose that the object becomes a musical instrument.”

      So for example a turntable, if played with the intent to make music (scratching, mixing, etc), is indeed a musical instrument. You make want to confine the notion of instrument to just traditional musical instruments with hundreds of years of history or some other arbitrary value judgement, but it is inconsistent with a reasonable, useful, and widely held definition.

      Regardless of whether Larry Fast at the time thought of the PDP-11 as a musical instrument, he was using it as one. It became a musical instrument through purpose.

    • PaulDavisTheFirst

      You may want to confine the notion of instrument to just traditional musical instruments with hundreds of years of history

      Never said this.

    • Taylor

      Right, you didn’t say that, but what you’ve articulated is tantamount to that. That is why I wrote “may” and added “or some other arbitrary value judgement.”

    • PaulDavisTheFirst

      “Making the air vibrate by making an object vibrate by stroking, touching or blowing it” isn’t an arbitrary value judgement.

      It is a fundamentally different process from “Making the air vibrate by creating a set of electrical activity that can be transduced into sonic pressure waves”.

      Not better, not worse, but fundamentally different.

    • Taylor

      Good, we’re making progress! Now, all you have to do is not be a bigot and realize that both are musical instruments!

    • PaulDavisTheFirst

      “Getting from here to there by exploding liquid hydrocarbons inside a chamber that makes metal rods move and turn wheels”.

      “Getting from here to there by exploding liquid hydrocarbons in a container which forces a turbine to spin and generates enough thrust to fly”.

      Sure, sure, fundamentally different processes but they both involve “getting from here to there” so we’ll consider them the same. And of course in some ways, they are. But in many other ways, they are not, and I just happen to think that the ways things are dissimilar is just as important as their similarities.

    • Taylor

      Sigh. Cars and planes are both vehicles. Acoustic and electric/electronic/digital musical instruments (including controllers, drum machines, etc!) are both musical instruments.

      You’re a smart guy, so I can’t believe I have to tell you this.

    • PaulDavisTheFirst

      Which part of driving a car translates over to flying a plane?

      They are both tools for transportation. How does ability or technique at one of them having any bearing on the other? We have drivers and pilots, and we don’t consider their activities as the same.

  • cooptrol

    Also there’s something different in the sound of hardware coming out of a club system, I haven’t yet figured out why, but it sounds better to me.