6 ms·
Thanks - really good to discover this! Interesting to note the size of the samples that Apple is using: * Balanced noise: 1.8MB * Bright noise: 1.8MB * Dark
by mft_ 4y ago
Thanks - really good to discover this!
Interesting to note the size of the samples that Apple is using:
* Balanced noise: 1.8MB
* Bright noise: 1.8MB
* Dark noise: 1.8MB
* Ocean: 62.4MB
* Rain: 63.7MB
* Stream: 72.1MB
That's a lot of effort to ensure your background water sounds are (presumably) high fidelity and non-repeating!
- amelius 4y agoDoesn't the finite size imply that they are in fact repeating?
- throwoutway 4y agoYes. I can detect repetition of ocean “white noise” if the length is too short. But a long enough track is just fine
- natly 4y agoIn this case they're probably recordings, but you could imagine that they'd be procedurally generated in which case not necessarily. A call to `Math.random()` is a finite length 'file' but won't repeat.
- aftbit 4y agoA call to Math.random() will in fact repeat eventually. The backing PRNG has a finite period.
- shhsshs 4y agoYeah, after it loops in several hundred years and I hear a pattern I've already heard, I'll be really annoyed!
- MichaelBurge 4y agoThe human brain is also finite, so even a true infinite random sequence would give the listener finitely many states. So it doesn't matter whether a sequence is finite or not.
- sudosysgen 4y agoIt may be partly affected by fully random effects such thermal noise, meaning there will be no period.
- JJMcJ 4y agoGood ones are loopable so there is no discontinuity when they repeat. And the size of those files implies they are probably minutes in length, at the very least, so you probably don't notice the repetition.
- PartiallyTyped 4y agoNot necessarily. In theory one can construct a chaotic (this is a precise definition) function with very little data that could replicate that class of noise with finite data. Infinite lists are also another beautiful construct in lazily evaluated languages.
- Retric 4y agoNot at a fixed rate with finite computing power or finite memory. Essentially a fixed algorithm needs ever increasing or infinite levels of precision in it’s calculations to avoid loops.
- karmakaze 4y agoSo like emitting 1, 0, 1, 1, 0, 1, 1, 1, 0, ...? Doesn't need much to generate. It would need a wide counter but a MB is a lot of bits.
- jacobolus 4y ago> Not at a fixed rate with finite computing power or finite memory. Essentially a fixed algorithm needs ever increasing or infinite levels of precision in its calculations to avoid loops. If you start talking about true “infinity” then sure. You need to store enough state (or if you like, receive some input) for your computed output to be unique. And counting to infinity takes infinite amounts of memory. Infinity is not accessible to humans, and is best thought of as a thought experiment rather than a part of empirical reality. But if you do 44kHz samples, a counter stored in a 32-bit integer that ticks up once per sample is enough state to give you all unique [i.e. non-looping] samples for about a day. A 64-bit integer will give you all unique samples for 13 million years. With a 75-bit integer you can get all unique samples for longer than the history of the universe so far. So in practice, the “increasing” amount of memory required doesn’t (necessarily) grow fast enough to worry about.
- mejutoco 4y agoCan prime numbers be used as lengths/periods so that there is no perceptible repetitions. I saw this trick and used it to animate different parts of an animation (each part every p frames, where p is a prime) and it works very well.
- andrepd 4y agoIf you're interested in high quality background noise, I'll plug this absolutely amazing site: mynoise.net
- laybak 4y agowow didn't realize background noise can get so sophisticated haha