World's largest natural sound archive is now online | Hacker News Reader

World's largest natural sound archive is now online

news.cornell.edu

142 pointsmhb23 comments

The licensing is unfortunately non-free and is therefore incompatible with free projects such as those operated by the Wikimedia Foundation. It is however a slight improvement on arXiv. At least terms of use are available in the footer of each page.

Basically you just need a license if you want to use the sounds in a movie or toy or whatever, but they're now free for research and personal use. It might not be ideal, but it's leaps and bounds better than it used to be.

Love it.

One of Australia's best http://macaulaylibrary.org/audio/152947/dacelo-novaeguineae-...

I've been giving them shit online for years for not making the archive publicly accessible. They even responded a couple times defending themselves. Great to see they finally did it.

It looks like the site is built on rails. Unfortunately it looks like they are either having site issues or there is still a little work to do on the code base.

http://macaulaylibrary.org/search?taxon=Odontoceti&taxon...

That's a huge red flag. You don't want those logs showing in production. Someone should notify them.

Cornell is an amazing institution considering its effort behind arXiv, this and I'm sure much more.

Now... why use a link shortener in an HTML page? Sir Berners-Lee would shed a few tears.

I'm thinking possibly analytics? I know a bunch of people who do that.

Alex3917, MP3 versions of Macaulay archival material have been placed online over the past 12 years or so, as groups of cataloged items on 1/4" mag tape have been digitized. So I'm confused about what you mean by "not making the archive publicly accessible". Word of advice: before giving people "shit", make sure you aren't full of it! ;-D

The real news is that everything previously cataloged and on tape is now ALL digitized to 96kHz/24-bit digital audio and on the web as MP3.

You can request 44.1/16 or 96/24 versions of MP3 material you hear on the Library website. The material is copyrighted, yes, but not priced to make a lot of money. Contact the library for more info.

kind of related, just sharing a happy discovery: a guy called chris watson http://en.wikipedia.org/wiki/Chris_Watson_(musician) makes recordings of ambient sounds ("field recordings"). they are less focussed than this archive, but longer, and available on cd (or via pirate bay etc i imagine).

his site is at http://www.chriswatson.net/ and you can hear some stuff there. it's pretty good for going to sleep to (some tracks) or working...

Very cool. The Vrindavan temple is great for "ambient music": http://macaulaylibrary.org/audio/161053

Fucking hell, the site is horrible though.

Only 16 samples for Homo sapiens. Does anyone know if something like this exists for spoken languages?

Yes but its really expensive and they have almost a monopoly on the vast quantity of data. They are the linguistic data consortium. Yearly membership fees for companies is $24000, for non profits(e.g education) is $2400 per year. Plus it costs extra for the actual data.

http://www.ldc.upenn.edu/

Wikimedia Commons has a lot of samples, though they're not very strongly curated at the moment: http://commons.wikimedia.org/wiki/Category:Audio_files_by_la...

There are various university projects for specific languages or regions. This site has a number of samples of American English and American German dialects: http://csumc.wisc.edu/AmericanLanguages/index.htm

http://www.meertens.knaw.nl/soundbites/. Hundreds of hours of Dutch dialects, free for research, 'please cite us', contact them for other uses.

From the same institute: http://www.meertens.knaw.nl/projecten/sprekende_kaart/svg/

does anyone know the fastest way to rip a sound file from this flash player for private use?

Chrome Dev Tools -> Network (or Firebug if you're into that :)

Yup. found the file I wanted by looking through the source for *.mp3 and finding the static file asset.

This bookmarklet will add download links. Copy and paste into a bookmark, click it when you're on an ML search/download page:

javascript:{ try{ titletext=document.getElementById("jqmTitleText"); pat=/([^0-9]+)([0-9]+)/gi; id=pat.exec(titletext.innerHTML)[2]; th=id.substr(0,id.length-4); url="http://audio.macaulaylibrary.org/+th+/+id+.mp3; titletext.innerHTML = titletext.innerHTML + " <a href=\""+url+"\" target=\"_blank\" style=\"background-color:#990000;color:white;padding:2px;text-decoration:none\">download</a>"; }catch(err){} try{ content=document.getElementById("content"); title=content.children[0]; if(title&&!title.children[0]){ pat=/([^0-9]+)([0-9]+)/gi; id=pat.exec(title.innerHTML)[2]; th=id.substr(0,id.length-4); url="http://audio.macaulaylibrary.org/+th+/+id+.mp3; title.innerHTML = title.innerHTML + " <a href=\""+url+"\" target=\"_blank\" style=\"background-color:#990000;color:white;padding:2px;text-decoration:none\">download</a>"; } }catch(err){} try{ content=document.getElementsByClassName("catalog"); for(c=0;c<content.length;c++){ if(content[c].children[2].classList.contains("play")){ id=content[c].children[0].innerHTML; th=id.substr(0,id.length-4); url="http://audio.macaulaylibrary.org/+th+/+id+.mp3; content[c].innerHTML = content[c].innerHTML + " <a href=\""+url+"\" target=\"_blank\" style=\"background-color:#990000;color:white;padding:2px;text-decoration:none\">download</a>"; } } }catch(err){} }; void(0);

That bookmarklet is wrong. I think hackernews filters out the "s in those links. Here:

javascript:{ try{ titletext=document.getElementById("jqmTitleText"); pat=/([^0-9]+)([0-9]+)/gi; id=pat.exec(titletext.innerHTML)[2]; th=id.substr(0,id.length-4); url="htt"+"p:/"+"/audio.macaulaylibrary."+"org/"+th+"/"+id+".mp3"; titletext.innerHTML = titletext.innerHTML + " <a href=\""+url+"\" target=\"_blank\" style=\"background-color:#990000;color:white;padding:2px;text-decoration:none\">download</a>"; }catch(err){} try{ content=document.getElementById("content"); title=content.children[0]; if(title&&!title.children[0]){ pat=/([^0-9]+)([0-9]+)/gi; id=pat.exec(title.innerHTML)[2]; th=id.substr(0,id.length-4); url="ht"+"tp:/"+"/audio.macaulaylibrary."+"org/"+th+"/"+id+".mp3"; title.innerHTML = title.innerHTML + " <a href=\""+url+"\" target=\"_blank\" style=\"background-color:#990000;color:white;padding:2px;text-decoration:none\">download</a>"; } }catch(err){} try{ content=document.getElementsByClassName("catalog"); for(c=0;c<content.length;c++){ if(content[c].children[2].classList.contains("play")){ id=content[c].children[0].innerHTML; th=id.substr(0,id.length-4); url="ht"+"tp:/"+"/audio.macaulaylibrary."+"org/"+th+"/"+id+".mp3"; content[c].innerHTML = content[c].innerHTML + " <a href=\""+url+"\" target=\"_blank\" style=\"background-color:#990000;color:white;padding:2px;text-decoration:none\">download</a>"; } } }catch(err){} }; void(0);

http://audio.macaulaylibrary.org/<hundred thousands digit of ID><ten thousands digit of ID>/<recording unique ID>.mp3

Examples:

http://macaulaylibrary.org/audio/163685 Here the ID is '163685'. Raw file is: http://audio.macaulaylibrary.org/16/163685.mp3

http://macaulaylibrary.org/audio/53276 Here the ID is '53276'. Raw file is http://audio.macaulaylibrary.org/5/53276.mp3

Is anyone working on an api.

Reply on news.ycombinator.com