Hacker News Reader

Comment by almd | Hacker News Reader

This is often used by audio mixing engineers and taught in a roundabout way at schools and studios. We think a lot about where thins “sit” in the mix. Proximity wise, and even height wise in a stereo mix. Eventually you learn how to locate things in headphones and it’s a really weird sensation when you realize you can do it. The kicker is we start out by simulating real environments in mixes, but then end up having to simulate what people expect from the medium as opposed to real life. For example something I learned doing video audio, if someone is writing something on a train, viewers expect to hear the pen on paper. But irl, there’s not a chance it’s audible. Explosions are always distorted because microphones end up clipping due to the volume, etc.

A great book on spatial simulation is The Art of Mixing by David Gibson. Older but forever relevant

View on HN

> For example something I learned doing video audio, if someone is writing something on a train, viewers expect to hear the pen on paper

Just yesterday was watching Territory season 1 where the characters have an intense suspenseful, almost whispering "serious voice" conversation while standing next to a running helicopter, without even raising their voices which took me out of the scene.

So the question is, do viewers want it, or do know it all producers say people do and put it in?

When they say viewers want it, they mean just about 90% won't notice. Most people haven't been close enough to a running helicopter to understand.

I'm having problems watching movies at all, there is so many things breaking my immersion. :-)

> When they say viewers want it, they mean just about 90% won't notice.

Even more than that, they will notice if you don't it the "wrong" way that they've come to expect. This is called The Coconut Effect: https://tvtropes.org/pmwiki/pmwiki.php/Main/TheCoconutEffect

Ah, had no idea it had a name!

My clearest memory of that was me as a kid watching a Bond-movie where a sportscar makes a screaching sound when driving down a sandy beach. I turned of the TV and don't think I ever saw a full Bond movie after that.

The list on the page you linked had one thing that isn't toally correct though:

>The very specific (but entirely unrealistic) echoing thud that is heard when all the lights are turned on in a large spacenote .

That sound is realistic if it is an old building with the heavy type of power relays or whatever they are called. They do make that sort of sound if the acoustics are right. They could be set up with timers so they don't start the lights at exactly the same moment to prevent overloading the fuses.

Have you never walked through sand and heard it screech or squeak? It's definitely a thing.

https://en.m.wikipedia.org/wiki/Singing_sand

Surprisingly, it’s also present in live sports, see for example this article (in Finnish) about sounds in winter sports broadcasts: https://yle.fi/aihe/a/20-10005843

You went on quite many topics there. Could you expand on the proximity and height? Fascinating

The closest analogue I can think of is how due to practice now anyone can close their eyes and imagine typing entire essays how they know exactly where the keys are. Try it.

I have played alot of videogames & at some point identified, how can I guess, the source of sounds. Guess, because it's nowhere near actual approximation. Most often, source is guessed by context. E.g. The door knocking sound illusion, which was used to troll streamers.

Then you have directional localisation based on delay between ears, difference in volume & properties of reverberations. Things to the sides are going to arrive in either ear at different moment. Add source if first echo & you have confirmation that a sound is coming from either right or left. The more directly to the side is the sound, the bigger the delay between ears is, so you get approximate angle.

Now we consider sound muffling, caused by shape of our head & ears. Things in front are going to sound clearer in the opposite ear, than sounds from the back.

The same principle is used for detection of height. Things below are going to get muffled, things above will be clearer. In reality, feeling sounds with the whole body helps in source localisation, which can't be emulated with headphones.

Ive seen a BBC doc that tested this. They had people use putty to change the shape of thier ears, resulting in an inability to judge the height of a sound. Given how differently-shaped ears are, as opposed to the inner structures which are virtually identical, this result points towards a learned skill rather than something genetic. We each must learn how our paticular ear shapes modify sound.

I once lived in a shared 4 story house and I always intuitively knew where everyone in the house was, even if they weren't being loud. You could just tell where everyone was based on how the house creaked in response to footsteps. We had someone new move in once who walked very quietly and it made me feel slightly uneasy because they were sort of invisible to my hearing.

There is a video for the art of mixing. It is indeed fascinating.

https://m.youtube.com/watch?v=TEjOdqZFvhY

If you know how to touch type you should already know how to do this. I know for me it was a requirement in my 7th grade class.

That's not at all what echolocation is. What you describe is locating the source of sound using binaural hearing (similar to how we can gauge distances using stereoscopic vision).

Echolocation is finding out distance to objects (not sound sources!) by sending a sound wave in a direction, and listening for echos that bounce back. Hence echolocation.

The only sound source is you.

It's a form of active sensing: literally how a submarine sonar works (or radar, for that matter). Bats do it, too.

This has very little to do with "locating things in headphones", as that is entirely missing the active part in the first place.

Then, locating sound sources using binaural hearing is not the same as analyzing the scattered echoes when the sound source is you (relative to yourself, you know where you are already!).

It's interesting that this is currently the top comment. I wonder how many people read the article before engaging in this discussion.

> literally how a submarine sonar works

And dolphins and whales, no need to go to submarines.

Interestingly, it took until after the invention of SONAR for the theory that bats navigate by echolocation to be accepted. The theory that bats use hearing for spatial awareness was first proposed in the late 18th century, with experimental evidence, but was rejected by the scientific establishment for more than a century. People didn't know marine mammals used echolocation until the 1950s.

I didn't know this, but the intuition that a tech example will be easier to grasp than an example from biology was why I mentioned sonar before bats in the first place.

Fascinating to find out that the scientific community had this kind of bias as well.

yeah- yikes I did not expect that to happen. My comment was extremely tangential and there were a good few comments already when I added it. I completely agree - it’s not about echo location

This comment captures a lot of important detail about echolocation.

Thanks! I'm glad you found it useful.

I think what you mean is that all your examples don't work when recorded. But a human being in a train may hear the pen on the paper.

You can absolutely hear someone writing in a room.

Whether it's audible on a train, depends on how insulated the train is.

I get the OP's point, but indeed this probably wasn't the best example.

Reply on news.ycombinator.com