Power efficiency in the violin | Hacker News Reader

Power efficiency in the violin

newsoffice.mit.edu

65 pointsSandersAK22 comments

I was lucky enough to live next to a master violin maker, Hiroshi Izuka, for most of my childhood. His violins are worth 10s of thousands of dollars at minimum, and sound absolutely fantastic. It was always slightly amusing to see his in-progress violins out to dry on a clothesline with little plastic umbrellas over them to protect them from bird poop and whatnot. Hard to find any pictures online and neither he nor I live where we used to but I did find this picture [0] of him working on what looks to be some sort of viola. He and his apprentice were constantly trying slightly different shapes and testing them seemingly by ear. I'll have to try to get in touch and see what he thinks of this, I imagine his response would be something along the lines of "what violin player cares about power efficiency?"

[0] http://www.orikam.com/img/iizuka.jpg

I'm not a violinist ... More power efficiency means that with the same effort you get more volume, or you can make less effort to get the same volume.

If you do less effort to play the violin, you can probably improve another parts of the performance, like make more accurate notes, or add additional sound effects or ... we should ask a violinist.

I'm not a violinist, but a double bassist.

This is going to sound simplistic, if not crass, but a very good first approximation is that more efficiency means the instrument is louder, and louder is better. There is no substitute for volume. You're paid to be heard. You will use the full capabilities of your instrument. And deep down we're not all that different from rockers. Loud is fun.

Now, other qualities matter too, of course. For a fine violin, I would say that volume is not the point, but the starting point. On top of that, you also need to have good tone quality and whatever else. And the volume that you can achieve in a musical fashion depends on your physical approach to the instrument, along with your bow. I own two bows, and one is noticeably louder than the other.

The historical evolution of instruments has involved a steady increase in volume, not just within families such as the violin, but in the size of ensembles and the choice of instrumentation.

This reminds me of a double-blind study where professional violin players were asked to select a Strad out of six violins that they played while wearing dark goggles. Only three out of seventeen were able to do it.

When asked which violin was their favorite:

"[T]he only statistically obvious trend in the choices was that one of the Stradivarius violins was the least favorite, and one of the modern instruments was slightly favored."

So while there were some brilliant violins produced at the time – certainly an accomplishment based on the resources available – we can also make some pretty good ones today.

http://www.npr.org/blogs/deceptivecadence/2012/01/02/1444828...

Because the violinists were only allowed to try the violins for one minute, that study has the same flaw that the Pepsi Challenge had in 1975:

"In his book, Blink, author Malcolm Gladwell presents evidence that suggests Pepsi's success over Coca-Cola in the "Pepsi Challenge" is a result of the flawed nature of the "sip test" method. His research shows that tasters will generally prefer the sweeter of two beverages based on a single sip, even if they prefer a less sweet beverage over the course of an entire can."

In other words, a saturated sound could initially be more attractive but not as good for an actual concert. http://en.wikipedia.org/wiki/Pepsi_Challenge#Criticism

But I'm repeating myself https://news.ycombinator.com/item?id=7550378

As a violinist, this is exactly right. Violins are not standardized, and are certainly not interchangeable. Everything from the dimensions of the body and fingerboard to the exact locations of the "sweet spots" where the instrument will respond most fully will differ. It takes weeks or months to really learn to play a particular instrument, and a test that only gives players a minute will not reflect how well the instrument can sound in the hands of an experienced player.

Another issue is that the instrument will sound very different to the player than to an audience, due to the way the sound is projected out of the ƒ holes.

That said, I think there is a growing consensus that the Cremona violins are a bit overrated, and that modern makers are making wonderful instruments for a tiny fraction of the price. When I bought my current instrument I ended up with a modern American violin, which I preferred to old Italian and French instruments that cost many times as much.

Indeed. The generally held view in violin circles today is that we're living in a second "golden age" of violin making.

I had an interesting experience with violins about 20 years ago. I saw two concerts with violin solos within a couple weeks. One was in the Seattle Opera House, with no amplification whatsoever. The other was in the Paramount Theater, a Yanni concert where the violist had a mike stuck on the violin and we heard it amplified through the speakers lining the stage.

The former sounded far, far better. Since I presume Yanni could afford the best amplification equipment, it was clear to me that no electronic reproduction can match the real thing, not even remotely. Thus ended my being an audiophile - it seemed pointless.

The Yanni speakers also had a perceptible background hiss. Very disappointing.

It left me with a question - why can't electronics reproduce the sound of a violin?

>why can't electronics reproduce the sound of a violin?

Well, if you're referring to synthesis based on physical modeling algorithms, it's possible, but it's just too complicated. A basic, not-too-expressive sound is easy to achieve, and already existed back in the days of the Ondes Martenot [1]. Getting the body's resonance is not the hardest part, you can just tweak it until you get something that you like.

But the most difficult part to reproduce about a violin is certainly the human and his interaction with the instrument. Sure, you could use a ribbon controller or one of the modern touch-based controllers [2], but it's still just a "spherical cow on a frictionless plane". Interaction between a string's vibration and the player's fingers is infinitely subtle.

Currently, synths based on sample banks are the best simulation of a violin that you can get. Just look up "violin VSTi" on your favorite search engine and you'll find many very realistic synths.

[1]: https://www.youtube.com/watch?v=Yy9UBjrUjwo#t=243

[2]: see the Persephone, the Hanken Continuum, and the ROLI Seaboard

Thanks for the information, but I was interested in why recording the sound with a microphone and playing it back does not remotely sound like a live violin.

Given that the speakers at the Yanni show had a perceptible background hiss, it's obvious that they were not in fact using the best equipment possible. You're working from a single data point and drawing a super-unreasonable conclusion. I have heard amplified and recorded violins that sound exactly as they should.

Without knowing what they were actually using, it's hard to pinpoint why it didn't sound right, but most microphones and speakers don't have flat responses (so the reproduced sound will be different), the signal could have been EQd to hell and back by the sound guy, maybe the actual violin player was sick and they couldn't get a sub so they just put a roadie up on stage, and maybe they weren't using a microphone but rather a piezo transducer and didn't match its impedance with an appropriate preamp. Way too many variables.

Because:

1. You'll usually close-mike the sound, so you don't get any room acoustics. If you're live, you have to close-mike the sound to avoid feedback.

2. If the violin has a bridge transducer, the sound will tend to the edgy and harsh. You won't get the full resonance of the body.

3. Pickup mikes on the body have similar problems.

4. Unless you have high-end neutral studio-grade speakers (~$50k), they'll add an EQ curve of their own. Most speakers are designed to sound a little bright, so that adds extra harshness.

A very expensive mike setup with very expensive speakers will sound very much like a live violin if the mikes are placed correctly.

Anything other than that won't, with varying degrees of tonal distortion.

There's also a psychoacoustic effect: you never listen to live sound the same way you listen to recorded sound.

If you record a conversation and listen back to it you'll hear all kinds of details and distractions. When sound is coming live from a physical source you can see, your brain filters out a lot of sonic detail so it can distill the important content.

I think for those venues, the practical reason is that a battery of microphone stands would be visually intrusive, so they (esp classical venues) rely on mikes suspended from the ceiling or clip on condensor mikes like DPA's. Additionally, violin and violists are notorious for swaying and turning away from microphones.

In the studio, i think they generally start with large and small diaphgram condensors and ribbon mikes for good results.

>I presume Yanni could afford the best amplification equipment

Being able to afford good amplification and having the technical expertise to implement it correctly are very different things. Audio engineering is a fine art, but most musicians know very little about it. They face the same sort of hiring problem as non-technical managers in our industry - they don't know enough to spot the difference between a bullshitter and a bona fide expert. Often musicians rely on house engineers in live venues, or the recording engineers and producers chosen by their label.

The best audio engineers go completely un-noticed, like the best sysadmins. There's nothing to draw your attention to the technology, because it is implemented flawlessly. A great live amplification setup has no noise or distortion, the instruments are all perfectly balanced and placed within the soundstage, and the amplification system works in sympathy with the natural acoustics of the room.

> Being able to afford good amplification and having the technical expertise to implement it correctly are very different things.

I can certainly agree with this, but I'm not certain the rest of the analogy holds. The problem with hiring excellent sysadmins and programmers is the lack of a straightforward test of quality. You can ask them to code something, and that will tell you that they are capable, but are they exceptional? I might be being naive, but I don't think that testing for that is so cut and dry.

However unless taking the 'best amplification equipment', setting it up for a performance takes substantially longer than a few hours, why can't an interview panel of listeners with a known good quality performer act as a test of capability?

- Panel, can you hear any distortions or noise? - No - You're clearly excellent, you're hired!

This certainly isn't my domain though, so I'm sure there are subtleties I'm not considering, but I'd love to hear about them.

Yanni is a professional musician who did music all day every day. I'm a lay person and I could hear it. I find it hard to accept that he couldn't tell the difference between excellent and inferior sound reproduction, and wouldn't demand the former.

He was at the height of his fame and so had a reputation to uphold and the resources to do it.

> But were the design changes intentional? To answer this question, the researchers worked the measurements from hundreds of Cremonese-era violins into an evolutionary model, and found that any change in design could reasonably be explained by natural mutation — or, in this case, craftsmanship error.

I want to see them build and run that evolutionary model and generate some mutant future instruments.

this was kind of interesting, but possibly misguided. I think their goal is to somewhat increase response and projection while keeping the tonal character of the instruments. Otherwise, they could just put higher bridges and replace the Pirazzis and Jargars (strings) with, say, higher gauge steelcore Preludes, the idea being to increase string tension and breakover angle at the bridge to increase downward force at the bridge. Likewise, heavier bow, more hairs, and make sure the shoulder rest is not damping the back plate at all. But except for the last, these will markedly change the attack and tone, or the heavier bow will throw off the players' technique. Also i think the Russian right hand technique lends itself to more volume but more laborious bow recoveries. I'll have to look at Galamian's book about that [2]

If anybody's interested, you can try to understand Strobel's userful measurements [1]. for my student violin and viola I more or less gave up, and just defer to my luthier's bridge and soundpost intuition.

_____________

finally, there's the ergonomic viola with substantially larger top and back plates, for greater projection[3]. i don't think i've ever seen one played. String players are a conservative lot.

[1] http://www.amazon.com/Useful-Measurements-Violin-Makers-Refe...

[2] http://www.sharmusic.com/Sheet-Music/Violin/Etudes-47-Studie...

[3] http://www.rivinus-instruments.com/Pellegrina.htm

An amusing factoid is that most of the Cremonese Master violins in concert use have been substantially reworked since they were made. As I understand it, they now have longer necks, shorter ribs, and steel or nylon cored strings. In addition, bow design has changed pretty dramatically since that time.

Stradivarius could not have known how his fiddles would sound today.

Yes, modern violins differ in various ways from baroque ones, and most of the surviving instruments from that era were "updated." The key differences include the length of the neck, which was shorter as players didn't use as much of the range of the violin. The bridge was flatter, which made chords easier to play. And the bass bar was generally shorter, although I'm not sure what the sonic implications of that are. And possibly most important, the bows were completely different. Modern bows have a screw mechanism that keeps the tension of the hair. Before that was invented (mid 19th c.) players would keep the tension with their hand.

As for strings, modern players mostly use strings made of synthetic cores wrapped in metal. The goal is to get as close to the sound of gut as possible without all of the hassle real gut entails (expense, tempermentality, short lifespan, long break-in periods).

That said, there is a growing movement of "historical informed performance," which is the idea that we should be playing period music on period instruments, instead of our modernized versions.

Tl;dr: Luthiers and violin makers brute forced fluid dynamic discoveries with knives and an inability to perfectly reproduce their wares.

Exactly. I've been re-reading Nassim Taleb's Antifragile, and the first thing I picked up on in the article was how they kept saying that there was no way the slight improvements were intentionally designed. So what?

Seems to be another good example of innovation occurring through trial/error and tinkering, as opposed to discovering the theory in a lab and then implementing it.

Reply on news.ycombinator.com