[Nazo]

So I like to use pink noise for things like burn-in. It may not make a huge difference, but I like it because it probably comes the closest of each form of synthetic noise of an approximate average of real music (eg not a huge amount of highs like with white noise or practically no highs and really large amounts of lows like brown noise, but an average of the two that looks fairly similar-ish on a spectrograph to relatively ordinary music -- or at least not so far off as to matter a lot for purposes of burn-in.) I'm not really here to debate how much difference it really makes -- I mostly just figure that if I have to choose between them, I might as well pick the one with the most realistic sound signature. (So in the end, it doesn't really matter a lot how much if any difference it might make since if otherwise one is as good as another, it makes sense to just go ahead and get the one that could otherwise potentially be better. This way if, by any chance, it does make a difference, the effects of the burn-in should be roughly equivalent to having actually listened to music for that period of time. It's very approximate, but for a noise it's probably as close as you get.)

But one thing that bugs me is that random noise compresses really badly for, well, very obvious reasons. It is, after all, just random (white) noise with oddly complex filters applied. Another thing that bugs me about it is it seems to be pretty hard to actually get pink noise. White noise is obviously easy since it's just a random "static" noise and brown noise is quite easy as well since apparently it's just "integral white noise" (well, the math is a little beyond me, but in other words it's ridiculously easy to get brown noise from a good sound editor or whatever.) It's actually really hard to find something that can actually generate pink noise for some reason. (It's probably easier just to take white noise and apply an equalizer or something really.) When I looked around a while back trying to find a generator that could just make a file of x length for me to work with it took forever and I ended up with a none too great shareware program to do it. Lol, since then I've just kept a careful backup just in case.

So here's what I'm wondering and where the more knowledgeable people here come in. I'm wondering if it might not more or less work just as well for burn-in purposes if we just did something simpler. I'm thinking in particular of just generating a bunch of sine waves and then applying some relatively simplistic filter to them. With a simple editor like Goldwave, this sort of thing should be easy really. I'm not really sure of how things like harmonics and such might affect this though. I'm assuming it's best if the waves don't overlap too much?

Anyone have any thoughts on the pluses and minuses of this for things like burn-in and what sort of waves and filters should maybe be used for the best results assuming it would work well? Or is the whole thing just crazy? Lol, I won't lie and claim to be an expert on sound sciences (but then not everything in sound is 100% science anyway right now.)

 

[Mochan]

All of that just went over my head.
 
And not in the sense of "I didn't get it," but rather, "wow there are people who actually worry about such things."
 
Good luck in your endeavour.

 

[Nazo]

Well, I'm not just "worrying" in terms of necessarily "can it be better" as there's also minor details like the fact that I have to use a DAP for burn-in (my old Cowon D2) and I'm running into problems like the fact that one hour's worth of pink noise seems to compress to about 500MB in FLAC (and I'm afraid MP3 compression might mess it up too much, but even that is close to 100MB) and I like to have about four hours worth of pink noise for one hour of silence. I'd be willing to bet that filtered sine waves would compress better than essentially random noise and probably would be friendlier towards lossy compressions like MP3 as well. I'd like to eventually be able to eliminate all usage of the D2 so maybe I can try to sell it since I'm no longer using it to actually listen to music anymore. I'd like to eventually be able to maybe use a much cheaper MP3 player for simple tasks like burn-in really, which may mean horribly limited space among other things (and of course, only the really good ones like Cowon's players generally support anything beyond MP3.)

So it's less that I'm "worried about it" and more that I just need a better way to burn in things like headphones.

That said, I do think this might be something just generally useful for the community as a whole. Everyone here needs to burn things in. If this gives us a better "noise" to actually burn things in with, it might be good for everyone. Of course, for all I know, sine waves are worse for burning in or something (which is why I need to ask the experts.)

 

[Head Injury]

If you want burn-in that works like real music, why not use real music on shuffle?

 

[Nazo]

Because that then begs the question of "which music"? Like I said, I liked the pink noise because it's a good average. It's a lot more work to try to come up with a full range of stuff or whatever when I can just generate an average.

 

[Head Injury]

Put it on shuffle and it doesn't matter which music, over a long enough time period the genre will be as random as your playlist.

 

[Nazo]

As I said, then the playlist has to be wide enough. I wasn't kidding when I said I was even considering usage of cheaper MP3 players for this. That means they could be as little as 1GB. Figuring out just what to put on there seems like a lot of hassle just to avoid using generated noise (and why are we anti-generated noise exactly? Is there something wrong with it?)

 

[Head Injury]

Nothing wrong with generated noise, but you wanted something different. I think intentionally burning a headphone in is stupid in the first place.
 
Do frequency sweeps from 20 Hz to 20 kHz with decreasing volume at higher frequencies, if you want. That might "work".

 

[Nazo]

Nothing wrong with generated noise, but you wanted something different. I think intentionally burning a headphone in is stupid in the first place.

Why on Earth would it be stupid? Heck, right out of the box most sound absolutely horrible. At a minimum you want to get past the initial horrible stage at least... Otherwise how are you going to enjoy listening through them instead of, say, some that already are burned in?

Do frequency sweeps from 20 Hz to 20 kHz with decreasing volume at higher frequencies, if you want. That might "work".

Then it's not really exactly "full" so to speak though since it would only cover a very short range at any given time.

 

[Head Injury]

Quote:

Why on Earth would it be stupid? Heck, right out of the box most sound absolutely horrible. At a minimum you want to get past the initial horrible stage at least... Otherwise how are you going to enjoy listening through them instead of, say, some that already are burned in?
Then it's not really exactly "full" so to speak though since it would only cover a very short range at any given time.

I thought you didn't want to get into a discussion on if it matters or not. I'd be willing to, if you're interested.
 
Then generate them all at the same time and play them together.
 
Or just download this pink noise file and loop it, if your problem is availability.

 

[anetode]

Quote:

Why on Earth would it be stupid? Heck, right out of the box most sound absolutely horrible.

 

[mcjorb]

I'm kind of intrigued by what Nazo is bringing up, if only for the sake of discussion. If we get away from the whole "in order to burn things in" motive, I think the whole concept of pink noise in general is pretty interesting. This is after all the sound science forum, what did you expect? 
 
However, I don't understand what's so hard about creating pink noise. The math seems simple enough, why does any given noise generator have trouble producing it? 

 

[Uncle Erik]

I listen to headphones straight out of the box.

They always sound fine. I don't see any reason to do anything besides plug them in and listen.

One thing I've noticed from being around here for six years (I started lurking in December '05) is that people really want to do something to alter heir listening experience in a positive way. People want to be able to say that they love their listening experience and that they, themselves, did something that significantly made it better. Thst they didn't just buy headphones and a couple of boxes, plugged them together and that was it. Some kind of active participation has to be there.

This manifests in two ways:

1. People burn-in, try lots of different cables, buy magic boxes, modify headphones, and get into all sorts of tweaks. They then have some kind of participation in their system and can claim that they would not have great sound but for the significant time they investigated exploring the differen tweaks and tuning their system with them. Most of these people spend a lot of time here discussing their discoveries and tryin to persuade people to use them.

2. DIY. Others - and not nearly enough - start building their own gear. Most follow plans and buy PCB-based projects and put them into commercial enclosures. Others design circuits and some do casework from scratch.

The difference is that DIY involves a bit of science, aside from creativity in casing. I know science is a dirty word to some, but the catch is that science actually works. If I want to design a power supply that delivers 1.5A at 250V, I can go to several established circuits, pick a suitable transformer, and use formulae to calculate the values of the capacitors and resistors. Then when you solder everything together and plug it in, you get the desired 250V. Or close to it; production variation in materials make a difference. But there's nothing mystical about that - you can dial it in just where you want it using known methods.

Electricity has been understood well for about 120 years. Audio has been understood well for around 50 years. It's fashionable to appeal to a "great unknown," but lack of evidence is exactly that. If you think evidence is unnecessary, that puts you in the same category as palm readers, psychics, tarot card readers, and the rest. (You might also notice that those folks have a keen interest in money. Ask yourself why.)

If you really want to contribute to your system, get a soldering iron. Knock off the nonsense. If you want to improve your rig, build a really good amp and DAC. You can get creative and artistic with the casework. You can also experiment with different circuits and devices. Also, look what the Audez'e guys did. They developed an excellent headphone. Maybe you could do something like that, too. Just don't waste your time arguing about how many angels can dance on the head of a pin. It's a waste of your life and doesn't make anything better.

 

[Nazo]

First of all, I do want to point out that this is a website dedicated to stuff like this. Even if you disagree that the electricity ultimately travels through physical components that can be changed over time either due to physical movements, temperature and chemical variances, and etc (though I must admit that I am curious as to why these components supposedly aren't affected by running electricity through them over a long period of time,) you must still surely agree that this is not the website to be fighting people over such things! That sort of talk belongs more in places like HydrogenAudio or whatever.

However, I don't understand what's so hard about creating pink noise. The math seems simple enough, why does any given noise generator have trouble producing it? 

Well, I suppose it's somewhat more complicated to actually do it -- white noise needs only random values between -1 and 1 (or whatever range one might prefer) and brown noise is just integrating white noise. I suspect that brown noise wouldn't be so common in these generators if it were at all harder to generate as most people only think of white noise. And there I think is the biggest issue -- in those things that generate noise, the developers often just don't even think of it. It's ironic because pink noise in some form (not always audio) supposedly occurs a lot in nature -- even in biological systems (though I'm not clear on exactly how that works.) I never really found an actual algorithm to actually generate it -- the 1/f just refers to the actual power over frequency rather than the actual generation itself. I'll admit that even if I found one I'm not sure if I could integrate it into something like Goldwave necessarily.

Besides perhaps making it easier to actually add it to generators like that of Goldwave via some perhaps relatively simple algorithms, I'm still wondering if something that compresses better and such wouldn't be possible. Anything random, by nature, compresses badly. What I was proposing would not be true pink noise at all and, in fact, wouldn't even really be "noise" per se (though if you played it for someone you can be they'd call it noise, lol.) Which is part of why I mistakenly asked in here -- you can't just throw in a bunch of frequencies and be done with it. For one, the idea was to cover the spectrum fairly well, but any overlap (harmonics or whatever) would cause it to no longer be 1/f, but instead it could ultimately turn into white noise instead I would imagine. For another, it would be "always on" so to speak, whereas "noise" is random and only averages out to a certain level, but essentially has fully on and fully off states even in there -- it seems like it should have "dynamics" so to speak (though exactly the best way to do that, I'm not sure. I'm kind of imagining a few scenarios such as having one half of the spectrum covered sporadically one moment, then the other half another moment or something like that, but really I have no idea.)

 

[andrewberge]

For generating pink noise, anyway, Audacity (probably the most popular free sound editor) just generated an hour of pink noise for me in 18 seconds, which doesn't seem too long to me. You can also do what HeadInjury said and loop it if you don't have enough space. Play it 100 times in a playlist or something.
 
In any case, if i were to put my mind into completely reinventing something, i wouldn't exactly start with pink noise.