To Save the Sound of a Stradivarius, a Whole City Must Keep Quiet | Hacker News Reader

To Save the Sound of a Stradivarius, a Whole City Must Keep Quiet

nytimes.com

65 pointsdreamcompiler48 comments

I am sorry, but I don't believe that they were able to hear a glass break outside of the building they were in, much less down the street or a few block away. This feels like lazy reporting or something. Am I crazy?

They say they're using ultra sensitive microphones, but it does seem sensationalized.

No, it's a typical puff piece.

Sample libraries are nothing new.

But now you know of "Cremona", and that "Antonio Stradivari" made some violins there (and some other things), and that they have the "Museo del Violino", and other things.....

Yes, if the background noise from down the street is a problem, then how come the players heart beat, other body noise and the chair noise are not an issue.

You're not crazy... inverse square law. Unless the noise was conducted (or ducted/focused) through a solid (horn/reflector) it would fall off by 80dB at 100m. In order to get it into a solid you need amazing impedance matching (or alignment)... which also reduces signal.

Makes a good story though.

I don't understand why they would use a hall in the first place and not a studio where they can both control outside noise and any color the hall environment adds via uncontrolled echos. It's very odd, the sound engineers seem to know what they are doing but it baffles me why they made this choice.

Interesting question. I don't know the answer, but the whole Stradivarius mythos is unscientific [1], so I wouldn't be surprised if there is some kind of unscientific (or "artistic") answer to this too.

[1]: http://www.sciencemag.org/news/2017/05/million-dollar-strads...

> the whole Stradivarius mythos is unscientific [1]

Seems to me a lot like new Coke. In blind tests people prefer new Coke, but if you tell them what they're drinking...

I suppose so long as people can tell the difference, recording the sounds of the Strads makes some sense, even if people prefer the new violins in blind tests.

Also I thought part of this had to do with the size of a sample vs a full serving. Sweeter or stronger flavors tend to do better in small volume taste tests (like the Pepsi Challenge). Drinking a whole glass can be another story.

Yeah, I feel like an acoustic anechoic chamber is a better place for this.

An anechoic chamber is not a good recording environment. The musician won’t be able to play well.

In this case they would just be playing individual notes and scales, so would it be as important?

Yes, you can tell the difference between a virtuoso and a regular player just by how they tune up their instrument. A lot of violin is in the right hand.

Hardly. You give the musician IEMs (just like at a concert) and you can juice their sound however they like.

2nd and 3rd order reflections are extremely important to psychoacoustics. Unless simulated, it will absolutely affect a nuanced performance.

The goal here is, allegedly, to generate reference-level recordings. Or, to say it another way, you can add to existing sound but it's much harder to take away.

I can use Avid to make your reference recording sound like any one of hundreds of concert halls in under an hour. I can also generate almost that level of fidelity in real-time with an Alesis rack mount box into your IEMs.

An anaechoic chamber is not a good place for human mental health. I believe no one has ever been able to stay for more that 40 minutes in the worlds quietest anechoic chamber in the US. Recording sound samples can take hours upon hours.

And even in an anechoic chamber, the best of IEMs would still 'leak' a little.

Thats a bit of a myth. They kick you out of the guided tour after 40 minutes.

That’s an urban legend, they’re fine: https://youtu.be/mXVGIb3bzHI

It's a bit spooky, but it's fine.

Because they’re trying to promote the city and its history and its mystique. There are better sounding violins than Strads and better recording environments than this but would that get any interest from people?

> the acoustics of the auditorium, which was designed around the sound of the instruments, had to be studied, as well.

Seems like the room was made specifically for these instruments.

So sample the impulse response of the auditorium, at as many locations within the auditorium as you like, and apply a convolution reverb to the dry samples of the violins. If you sample both the violin and the auditorium together then your virtual violins will always have to play in the exact same location they were recorded. You're losing flexibility for no good reason.

The point is to record the violins in a hall that was designed for them in the town that they were made. Think there are already enough clean studio samples available.

The violins are not omnidirectional point sources. They have complex, frequency-dependent energy radation. A regular impulse response isnt going to cut it.

There are ways to capture non-point-source IRs, but none are the same as playing the instrument in the space.

Additionally, and perhaps more open to argument, the movement of the performer is part of the sound of an instrument. If the violin is locked in place, with a fixed IR applied, that's not really "what a strad sounds like" when its played for real.

I would suggest that having many more microphones would be a nice idea though.

Similarly to how farm cows get pregnant, with bull sperm being harvested with a machine and cows later manually inseminated, it is perhaps more efficient than the traditional way, but arguably lacks some... oomph.

The violin was not built for a studio.

The museum hosts dozens of instruments it does not own, probably worth 10’s of millions and left there by the owners to be kept in a controlled environment, seen and heard (the instruments are regularly played on by local musicians and conservatory students). It could also be that taking these expensive instruments out to a studio without the appropriate security logistics is a security/legal nightmare and they were never going to be able to pull it off.

A couple of years ago I was driving through the area with a couple families and on a whim decided to check if Cremona had a violin museum. Kind it and spent the afternoon. To this day I think about it and I think it’s one of the top 3 museums I’ve ever been to. It’s mind blowing, exceptionally well done, super interesting and worth every second. An absolute, almost unreal gem, highly recommended!

Just a side note, but the auditorium they're using to record is gorgeous. Its lines themselves remind me of a violin.

That's exactly why they chose it. The mythos of the Stradivarius is 100% about look and feel as opposed to anything real about the sound. It's the same reason expensive wines are tasted in fancy tasting rooms.

True. It is called priming: your mind is preparing itself for something special and it makes you enjoying the thing more.

>The mythos of the Stradivarius is 100% about look and feel as opposed to anything real about the sound.

Nope. The Strads do have a great sound. The fact that select modern instruments can make equally great sound does not in any way diminish that for most of the last two centuries, Strads were among the very best-sounding violins.

Nope. There have consistently been instruments which are indistinguishable or better for roughly 100 years. Stradivarius:violin::Rolex:timepiece

Well, Stradivarius was exceptional at the time, and the fact that these violins still hold their own after more than 2 centuries means that it's not just look and feel. But yes, there are (some) better instruments available nowadays and we are also able to simulate any instrument available with outstanding precision.

According to the article, they used a 1615 museum piece for this.

I remember reading somewhere that you have to play those instruments regularly for it not to change its sound.

Could anyone clear that up for me? Could you tune an instrument, that wasnt properly tuned for decades or centuries?

Yes that’s correct. If there isn’t serious damage to an instrument you can always tune it.

As far as playing an instrument regularly for it not to change its sound, that’s also true, although I don’t think anyone knows why exactly. It probably has something to do with the way the sound energy of the instrument, when played, is absorbed by the wood in a patterned way over time. Kind of like a used baseball mitt.

If the instrument isn’t played, something happens where the vibrations of the sound don’t flow correctly through the instrument and it can sound, and feel “stuck” to play.

https://www.sciencemag.org/news/2017/05/million-dollar-strad... disputes "A Stradivarius violin, viola or cello represents the pinnacle of sound engineering, and nobody has been able to replicate their unique tones."

It isn't possible to create a database storing all the possible tones a non digital instrument can make.

They are possibly not going for the individual tones, but rather taking a range of impulse responses (IRs) plus the original timbre. From that, they may be able to digitally reproduce the sound - or else use an ordinary violin, and colour the sound with the Strad's IR to recreate similar waveforms.

How is this relevant? The recordings only need to be good enough that a (trained) human cannot tell the difference.

Both harmonic and inharmonic overtones and a changing mix over time. Would be a big database.

https://www.pianobuyer.com/Articles/Detail/ArticleId/40/Digi...

Above reference discusses some of the issues for digital pianos - which are well-established and it is easy to find demonstration models in music shops.

Some fun dendrochronology references:

"Stradivari, violins, tree rings, and the Maunder Minimum: a hypothesis" https://s3.amazonaws.com/academia.edu.documents/30407555/bur...

"Chemical distinctions between Stradivari’s maple and modern tonewood" https://www.pnas.org/content/114/1/27

The Stradivarius' fame is not based on any technical superiority. Like all old things, it starts to be appreciated purely because it's old (and we like to collect old stuff, nothing bad about that itself).

In its time, its quality was remarkable. The change is anyone with aged wood, superglue and a cnc can do better now. Then, it was a master craftmans opus.

Amusingly, rescuing a Soil Stradivarius happens to be a side quest in Fallout 3.

http://fallout.wikia.com/wiki/Agatha%27s_Song

Are there any recordings of the same piece played with a Stradivarius compared to a "regular" instrument? I have always wondered what makes their sound unique.

Rarity, ego and cost. Most blind tests have pros perferring cheaper modern insturments.

There have been some studies comparing Strads with modern vilins: https://www.livescience.com/44651-new-violins-beat-stradivar..., not sure if there are any publicly available recordings.

Reply on news.ycombinator.com