~ Elusive angles ~
         to basic   

First published in December 2002     ••     Version April 2009

Elusive angles
"dark" 909 - "light" z404 search challenge
copyright minions and searchers meet once more

by ~S~ vvf & ~S~ jeff / see Credits


Upon heavy request HERE are the music files you'll need to (try to) solve this old riddle:
music/909.wav (2984828 bytes !)     ••     music/z404.mp3 (6156126 bytes !)     ••     music/909.MP3 (2993422 bytes !)


It's a searching roman: Print it, and then read it at ease, sipping something nice. Moreover, there is so much to learn here that I do not know -literally- where to point your attention to. Snippets like "...because using words of your own language is a filter per se : music... musica... musik... musiikki... muziek, and so on" should be worth an essay per se. I am not sure that my 'teaching riddles' deserve such deep work being done on them, but, by jove! I am happy vvf and Jeff did it: It's a great reading for me, and I have learned a lot here: "One of his VERY valuable findings is the amazon trick: change the .com part of the URL to .co.uk or .fr or .de or whatever you like and see the amazing thing: some countries let you listen to more clips than others!"



Credits

Credits to all seekers who contributed - this search is largely a collective work:

lukerone - gs - lopo - Ira_Melanox - ritz - esser - Mordred - amil - Finn61
elim - the righteous one - IM - aapje - +link - loki - il-li - liman -
fravia+ - mark - Boojum - stiletto - anonot - selim - Yeah!!!


A couple of you are listed more than once; or twice; --- but we won't tell :)

Prelude

A seeker's not necessarily a good writer, as I'm sure you will notice. More than a technique, it is a great gift to be able to explain your thoughts clearly, in a logical, interesting and even fun way for others. This essay has to be written nonetheless, in order for the info to exist somewhere, for the benefit of fellow seekers facing similar challenges who, we hope, might find useful information inside.

The whole 909 thing started as a game or as a riddle, like our friend fravia+ has accustomed us to in the past few years, but under this riddle fa�ade lies in fact a searching lab. People join in these searches for various reasons, and not the least because searching, in addition to being useful, is indeed great fun. Perhaps this is why it doesn't feel like learning at all:) But let me assure you, jeff and I discovered quite a few tricks on this search for the elusive 909 and its ubiquitous sidekick, z404, and it would be a pity to let them fall back into the great unknown again, thus this 'scattered' essay that attempts to capture all the things we tried, successfully or not, before we found our targets. And remember, the search took 4 months. So, that other file you were looking for, if you didn't find it yesterday night, although you tried and tried and tried, don't despair. There are more chances it's out there than not. I have a couple of searches myself that I've begun maybe 2 years ago. One of them I was able to complete in one hour or so when I resumed it after the 909 ordeal:) Sometimes, all it takes is patience until you've learned the proper ways...

This new riddle, or rather searching challenge, did not receive the honors of its own messageboard, most probably because its natural place was on the Seeker's. As a result, the contributions to it are mainly stored in several threads which are quite long and difficult to read. If you feel like taking your chances, here they are, in the chronological order of their posting:

01. first occurence
02. language of 909
03. more language
04. slightly off-topic: one after 909
05. music identification
06. the thread that broke the camel's back
07. 909 busted
08. vvf solves z404
09. nine... well... 9 is a good note to end on :)

At first we thought we should copy them here in full, so we could thoroughly show how everything evolved and developed and worked out. A nice plan, very detailed, but which would have resulted in a huge, unreadable and diluted essay. Our thoughts are scattered enough already, without the need for such extra aggravation:), so we decided to just link to those threads and copy here only a few relevant bits and pieces. It was helpful to send that draft to several friends and Seeker's regulars (Mordred, RH, ritz, Ira_M) who pointed out this same thing that we were worrying about. A lot of thanks to them! We hope this time they can read it without falling asleep :)

As I'm writing this intro, I don't have a clear idea of the essay's structure and can only hope it will emerge in the process of writing & collating all the material. I am not a very orderly person, and in my opinion a search, as anything else, should also leave room for intuition and imagination. Although he doesn't seem so ;-) jeff is much more methodical than me, which I admire a lot because I can't quite do that right. He keeps track of the keywords and combos of keywords he uses and tries to further his work accordingly. Myself, when I'm in the heat of a search, I'd be too impatient to write down everything and work methodically. I'd rather rely on my memory and try to chase a dozen rabbits at once, pasting large chunks of stuff in a doc that I'm unable to read 5 minutes later. I have to excuse myself for some things in here that may seem unclear or insufficiently explained. Well, sometimes I forget the exact things I did... I can recall the general context or technique, but the details are gone. I had no idea jeff would talk me into writing an essay and, as I said, I'm not a big fan of taking notes - notes that I never use afterwards anyway.

Outline

This essay was built on a skeleton or outline defined by jeff, consisting of all the questions that came to mind with respect to the 909/z404 challenge. They're the questions we struggled with during the search, trying to answer them as clearly as we could. They are very specific to the nature of our targets because, as you know, the web is first and foremost a web of text and this means most searches have to resolve to text in one form or another. Text is the royal way, even when searching for images, sound or video. However, things move real fast out there and new tools and new approaches become available literally every day, some of which are amazing. We should be aware of them.

Ok, let's see the questions:


So, our 909 target is located here. In addition, there, later, after the hunt for 909 began, was a posting by fravia+ to look into another song he refers to as z404.

For starters, how would you identify and locate a song you hear? Basically, you could: Having any slight problems getting those mp3's in the returns above to work??... read on... :-)

What techniques could you use if you know a few words relating to the song, its title or the artist? Something to keep in mind is that most simple searchs are going to take you to search engines returns that are placed by algorithms in that particular engines 'priorities'. Some may be whoever PAID the most to be returned before someone else... Others such as google will depend upon a system called PageRanking... the short story being that the MORE pages and links that POINT TO your page/site, the more authority your page takes on as a higher return value

You can spend an inordinate amount of time skimming through the commerical returns that may, give you but tiny sound clips... or none at all. So, who is it that may have whole songs? Well most likely it is our joe-public. But how to find joe-public? A general type of search would be to find people who have indexes with music inside (bear in mind that you have requested mp3 in this query, so don't expect to see a lot of wav, or ram type of files jumping to your attention), then you can take your search to the next filtering level, depending on keywords, for instance: mozart. By scanning these files you can begin to get an idea of what Mozart type music files are named. Then again you can get even more specific in your request now, and cut your returns down to be even more specific by, say, a song name. Keep in mind generally if you are looking at a filesize of 2-3-4-5-6 megs you can be fairly confident that you aren't looking at downloading some tiny 60k sound clip... however never be so confident as to not run your virus checkers.

Also bear in mind that most search engines are not going to index and give you returns for pages that are generated 'on-the-fly'... which means you are going to miss a lot of returns that are buried in, and have to be requested and dug out of, databases. There are a few tricks you can use to bypass the main and advertizing pages dynamically: see playing about with URL queries, but these will only be generated by those who want their ?query? pages to be indexed and identified, and you will need to play with the url query-ideas more to find magic in the databases that can be returned via search engines. Ah, but playing with URLs is an essay in itself :)

You could also try other filters such as genre and look for: classical, rap, rock, etc. Some people aren't so forth coming: track1 ... results 95  .... track01 results 401. As well as slightly obscured:  ... 6d%70%33%   ... ok; ok; that's not all that exciting butt, i lub the way google asked me if I'd prefer to search in English :) so don't forget to search things/names in your native language whether it be Swahilli or code :), because using words of your own language is a filter per se : music... musica... musik... musiikki... muziek, and so on.

The bottom line here is to construct your queries depending on your subject; use filter-keywords, depending upon your specific hunt, and always think 'peripherally' - if not out of the norm.

As further help, Mordred also pointed out that it might be good to mention that one can test a text on unknown language with one of these language recognition tools. I don't know how well they work, so you should test them for yourselves and see what you get, and keep in mind there's always room for error (like when you enter a sentence in Italian and the script tells you it's Latvian:)

If you are coming up empty looking for something 'in English' you may want to take your word to an online translator and try the word in a different language... :) I'm not going to bore you with lists and lists of translators, instead I will give a few sample keywords so you can always find the one you might need, in order to best avoid the trillion sites that are trying to mainly sell you translator software. I have found a good set of keywords to use is "enter text"; this will usualy jump you to pages that allow you to enter text, or a url for online translation.

Don't ever discount people's bookmarks and the work they have already put together. Combing rulez. Always remember that all these people's specific types of definitive interests perhaps identify them; their email addresses; their web pages; their links; as to their knowledge, and their interests, as peeps to comb, kleb and socialy interact with for help advice and answers. I am sure you can come up with better keywords for your returns :)

In another dark hole of websapce we have P2P (peer-to-peer). If you already have clients, great! You can certainly search for something inside P2P networks IF you know what it is; what is called; etc. If you begin grabbing P2P clients, be aware that most music-apps are filled with Adware trojans, spyware, malware and the likes... As just a tiny example, you will discover that your MS media player is recording and sending in the background all info about your music files over the internet to BigBrotherBill. It even records what music you have Burned onto CDs... hummm

Take for instance Napster's newest version which, now, has the code that looks into your music files to see whether or not it is going to allow you to upload and download copyrighted material... Sometimes it is a great idea to retain and use older versions; yes? Butt that's stuff for another essay :) :)

Just remember before installing new apps to go ask questions about [those specific] apps to get an idea about what others think about the particular app AFTER they have installed it... An example: envision WHAT and HOW a 'complainer' would ask a question or express anger about your 'subject' ... and ask it as he/she/it/ may have asked it: pleeeeeeeeeeeeeeeeeeeeeease Help!!! (Don't forget to check into a nice lil proggie called adaware - it listed over 200 ad trojans on my own system :)

Anyway; I am getting OffTrack as usual :) So, what do you do, when things aren't quiet so simple and when all above fails to identify and find a particular song? What do you do as in this case where you are presented with the challenge of finding this particular 909 song where the filename is unknown, the artist is unknown, the language is very questionable, and therefore the lyrics are unidentifiable?

Problem: a song with a language unfamiliar to you; what can you do, and what things might come to mind in order for you to identify a language?
  1. find/ask a recognized expert, and if you can't find one, then
  2. as ritz suggests, find the category of a group or messageboard that may specialize in the subject of your search... i.e. in the case of hebrew/arabic/klezmer/ sephardic/ hasidic / ladino music... see if you can find a group, or a messageboard, with knowledge of these subjects, or a specific subject, and enlist their interest and knowledge in helping you determine and classify. If you can then get a few keywords out of the lyrics, do a text search:) With any luck, someone will know the artist, the song title, the text, and even where you can get it.
  3. weed out people's guessing and 'maybes' and try to evaluate and focus on the positive and definitive. Keep in mind that their ideas may be very very worthy of investigation though. Who seems to be the expert in the group?
In the musical-riddle discussions on the messageboard, the one to watch, and evaluate, is our 'righteous one' imo (pssst---keep yer eye on that ole tzad-ira_M :))))

When someone identifies the lyrics, and no amount of searching brings up an identifiable song --- then what?? Let's see here quickly everything we could use:

a) What kind of identifiers might be within a song file itself?
  1. file size
  2. song length
  3. ID3-Tag in file-properties
  4. artist name
  5. genre
  6. date
b) What tools are available to capture and save stream audio?
  1. asfrecorder - grabs and saves as asf file
  2. Total recoder - Total Recorder Standard Edition records streaming audio, mic input, line-in input, and DVD and CD sounds. The recordings are saved to the PC's hard-drive in either WAV (compressed or uncompressed) or MP3 format. Convert different sound formats to WAV and MP3 (only gives you a 30 second clip -- cause it is broken -- butt the fix is easy to find).
c) What tools are available for identifying music?
  1. Moodlogic
  2. Id3man
  3. there are others too...
d) ... and comparing music? e) Can you hide information inside an audio/music file? How? Where? f) What could be some reasons that a filename is not being found in a 'quick search'?
  1. the author has renamed the file
  2. the file hasn't been indexed (yet)
  3. it's not on the net; or maybe it's not on the "free" net
  4. you need to expand your thinking and tools and entertain using different and various search engine/s
  5. the song/whatever is in a deeper database than any regular search engine indexes and produces
  6. you may be using one too many 'incorrect keywords'
  7. the database just may have it spelled slightly differently than you are spelling it...
  8. you may be searching one database that does not have the song (or song clip), but the same company has it, instead, in a different language/country/URLs database. An example is amazon.com...
  9. well... 9 is important. So there's 9 good reasons! :)
A note about filesharing and downloads: often you will see a small, to a very large difference in file sizes for the same named file --- I'd be very leary.

Filename 'naming'

Names and spelling are vital, as you well know. Everyone has a reason for how, and why, they sat there in the comfort of their puter room and named a file. What is the filename and what reasons can you come up with for the author's naming his/her file that way? What do you know about the author of the site? Did you notice any patterns while browsing the place where the file came from?

If your reasoning path doesn't pan out, don't give up. Make a new plan. What other reasons could a riddler have had in mind when naming a file, say --- 909?

It is likely that your search yields nothing because you are simply not using the correct keywords: too many, to few, or entirely offbase. Or maybe you are not familiar enough with your subject to envision the correct keywords to use. An important note imo: as you study and read any new subject, you need to gather and investigate new words. Make a list. Find their meanings,imagine how they may relate to your subject. In many cases single words in themselves can be the anchor you need to filter out all other returns than the one return that you seek.

A good idea also is to find, and use, various variations of a word. Example: tsadi, tzadi, tsadic, tsadik, tzadik , tzaddic, tzadeket, tsaddiyq, tzaddiq

Keep your eye peeled for variations in spelling of terms. Example: zaide and Zaïde will give us different returns (unless you use alltheweb --- atb ignores the ï); rhue and ruhe can both be found--- but which is right, and which will hold your pearl?

A great example of variations of spelling of an artist is a recent posting by vvf: Vissotski on the web. Look into the subject a little... you'll find files that a search for Vissotski would never bring to light because peeps have indentified him, or named him, differently with many different variations of his name: Высотски, Высоцки, Висоцки, Высотский, Visocki, Visotski, Visotskiy, Visotsky, Vladimir Vissotski, Vissotsky, Vysockij, Vysotskij, Vysotsky, Wisocki, Wladimir Wysotsky.

Back to 909, a word like Ashkenazi comes up in our searches below and is even used in one of my wordlists listed; are their other ways to spell it? Are their other forms it takes? I often think of it as 'cross-referencing' my accumulated wordlists. Often combining and cross referencing the right combination from your wordlist/s will produce nuggets.

Finally, let's say you have identified a song/artist/title; how do you go about finding it? You could use either search engines, or specialty search engines, or other ways to get a slip stream into your puter :), like request it and grab it:)

And now, let's review a bit the notes taken from various people and their ideas who posted on the Seeker's board. Some of it may seem redundant butt you will find that all of the answers to finding 909 and z404 songs were there :)

( from the seekers board)

(special comments to watch for will be in red--)
Re: Re: attn - vvf's compilation of previous posts and ideas on 909 wave (21/08/02 20:32:40)

jeff:
also take a look here :)
vvf:

jeff:
and knowing your riddle author is important:
a time before the battle of Austerlitz  f+ also used a date to create a filename:
1896 & 1897



vvf:

vvf:

Speaking of history, let's try a quick search on "june 5" and history, see what we can find. Maybe that date is important. Here are some events that occurred on a June 5th:

Well, it turned out eventually the riddle had nothing to do with history after all, so I spare you the details...

jeff:

One of the most important posts for 909 that gave a very very big beginning was this one:

jeff & vvf:

This should have been the beginning of the end for our riddle. Once you got the lyrics, the rest is a piece of cake, right?... Right???

Not really. We skip here the whole research in music & instruments & various names, which you can still find in all its beauty in the threads linked at the beginning of this essay. We better concentrate on the lyrics and language instead. Ira gives a triple smile ... so if we think about it not only does IM tell us he knows what this is but he is telling vvf something he said is on right ON the mark; or close ... or dang ... then Ira says: then vvf The h and r etc are mentioned several times again in later threads. Several threads on phonetics written by people who probably know the song and are giving clear clues --- ira (paraphrase): Then mordred Using righteous ones roman numerals (lxxv) and mordreds word, transliteration, you get aapjes' page: Now the righteous one seems to indicate at least now you have the "lyrics". One thing is sure, there is the word 'tzadik' in frav909.mp3.. so, can it be a chassidic song?

jeff:

note to vvf, you're right! :) Had i done my homework up to that point I would have read that hasidic and chasidic were blood brothers... :)

And the word tzadik (qarnothtsaddiyq) is definitely in that song isn't it... :)

vvf:

There's this thread started by Mordred on the Hebrew language. Why the Hebrew language? Probably because it's the language of the song, although no one specifically says so, and some folks even deny it:) Let's refresh our memory a bit: That first link indicated by Mordred contains a very good description, which I will paste here:
So there is an ancient, aka "Biblical", Hebrew, and the modern Hebrew of our present day. They might not sound the same, because the modern flavor is reconstructed. The link is the written form. A very interesting thing must be pointed out here: no vowels in writing, thus need for transliteration...

Which brings us to this mysterious post:
Indeed:) I scanned the results pages and found this little treasure:
It's a discussion in Bulgarian. CDYQ reads "tzaddik" - certainly got my attention. So we have this:
75:11 VKL-QRNY RSOYm AGDO TRVMMNH QRNVT CDYQ


Ha! Looks like code to mee. Or Hebrew without vowels. The consonnants somehow seem familiar, they ring a mighty bell. Let's follow http://www.sacred-texts.com/bib/tv/psa.htm ... It's the Psalms in Hebrew! Here's the line mentioned above, this time with vowels:

75:11 vekhol-qarnêy reshâ`iym 'aghaddêa` teromamnâh qarnothtsaddiyq

Now, 75:11 is lxxv:xi in Roman numerals. If we check cxxxii:xviii (132:18) we get this one:

132:18 'oyebhâyv 'albiysh bosheth ve`âlâyv yâtsiytsnizro

SO, Finally! We have two different lines from two different Psalms ("tehillim" in Hebrew). Both are the final lines of the relevant psalm - interesting. In English translation:

75:11 All the horns of the wicked also will I cut off; but the horns of the righteous shall be lifted up.
132:18 His enemies will I clothe with shame; but upon himself shall his crown shine.


The righteous one (tzaddik) is in there all right! These are the lyrics to 909, hypnotically repeating throughout the entire song. Look at these lines and listen to the song, they'll start to make sense:

vekhol-qarnêy reshâ`iym 'aghaddêa` teromamnâh qarnothtsaddiyq
'oyebhâyv 'albiysh bosheth ve`âlâyv yâtsiytsnizro

vvf: Here, jeff tries some gematria! It turned out it wasn't related to the riddle, but it's a legitimate attempt nonetheless.

jeff:


Here's another idea I put forth to vvf and mordred ... go to the what's new page and look at an essay of the emperors clothes posted just before the riddle -- having to do with 'levi's'...


vvf:
If all these numbers didn't make you dizzy, let's do a bit of language tracking...


Babylonia
a language unfamiliar to you; what can you do?


  1. find/ask an expert
  2. (as ritz suggests) find a group or messageboard that may specialize in the subject of your search and see if you can enlist their interest and knowledge
  3. if you can, then get a few keywords out of the lyrics and do text search
Tricky. I'm speaking for myself here, but I assume it should work much the same for everybody: my relationship with languages falls into the following categories:
a) the language(s) I can speak/read/write
b) languages I can understand or read to some degree, but cannot speak or write
c) languages I can recognize, but cannot understand
d) languages I can associate with some larger language family
e) languages I have absolutely no clue about

The above is not a scientific, but an empiric classification. Let's see the details: a) is straightforward; b) might be a language I have learned in school but never got to practice, or another language that's close enough to my mother tongue or the languages I know; c) is any language that I've heard spoken around me for some time (including on TV - there are many foreign stations available these days), so that I'm able to recognize it when I hear it - even if I can't understand a word; d) those are languages that bear similarities to languages I can recognize, but which I can't define for certain; and e) are languages that resemble nothing I've ever been in contact with before.

From my own personal standpoint, the language of 909 fell somewhere between d) and e). It sounded very strange, and yet its combination of sounds pointed to the Middle East. I was sure it wasn't a flavor of Arabic because I've heard Arabic before, spoken by people from Morocco, Egypt etc., and it clearly wasn't that; but it had to be close! I couldn't pinpoint it, but I would have bet my money on Hamito-Semitic, and more specifically Semitic (the larger family that includes Arabic, Aramaic, Hebrew etc.). Therefore "Semitic" was the last level of detail I could be more or less certain about. In order to go further I needed to either ask other people or comb other net sources for info.

One of the solutions, suggested by ritz, was to find a group or messageboard specialized in the subject of the search (hebrew/arabic/music) and try to obtain their opinion. A logical approach here is to research, evaluate and shortlist several good candidates, then go back and comb the existing info and discussions of these groups, then finally, if the info still eludes you, try to contact some of the people on the board directly. If you ask politely, you may get your answer. Then again, you may not. At this point some people would use trolling (the harmless kind)... I can't comment on that, that's a matter of personal choice, but if you were a knowledgeable person on one of these boards or groups, wouldn't you rather respond to a concern that you feel is genuine and honestly stated than to a question that feels like a troll?...

I did not use this approach, however a good place to start it would be the usenet, nowadays known as "google groups".

Another way to find an "expert" is to find a person who speaks the assumed language of our target and ask that person to decipher it for you. The trouble here is that you have to start with a good assumption in order to reach good results. Actually I considered this myself, in the beginning of the search. I believed the language of 909 had a good probability of being Hebrew or Ivrit or some other language or dialect in that area, so I sent 909 to an old friend of mine who lives in .il now. I got no response, the address isn't good any more, and I took it as a "sign" that I better investigate 909 on my own :) The only other person I knew who could help declared he can't take part in the riddle-solving because he already knew the answer...

Well, having the lyrics deciphered like that would have killed all the fun. But in a RL search I would definitely use a combination of the two above approaches in any way I can, because it yields the fastest results.

What I used instead, as the riddle was "for the sommer" :), was a "lyrics" approach. That's the hard way. The idea is to listen to the song until your ears bleed, try to write down the lyrics as you hear them (phonetically), and then try to match those phonetic lyrics to real words of real languages. This is truly difficult, because:

a) depending on you mother tongue and/or training, you may or may not hear specific phonems of other languages (classic examples: Spanish confusion between the sounds "b" and "v", Chinese or Japanese confusion between "l" and "r")
b) you don't know how to separate the sounds you wrote down into words; you'll have to make assumptions here
c) you don't know how to spell the words in that language; assumptions again
d) in 909 there's a transliteration problem: written Hebrew doesn't use vowels; vowels are introduced during transliteration to latin alphabet, so I found multiple variants of the same word and this is confusing
e) let's not forget our target is a song; words and sentences might be pronounced slighlty different than in plain spoken language (for euphony purposes); this is especially true in z404 (opera)

The only word I truly recognized in 909 was "tzadik". Now, spelling it is not that easy: tzaddik, tsadik, tsaddik, saddiq, etc.

The other word I concentrated upon was the first word(?) of the song, that sounded like (very approximately) "wekhalka". I tried searching for this word under all possible forms (tried all kinds of ways to write it), in combination with "tzadik" and its variants. Found nothing of value. In retrospect, it's funny that using only "vekhol" and "tzadik" (that's ONLY TWO keywords!) we hit bull's eye. But at the time I was scratching my head to find suitable ways to spell that "wekhalka", then seeing if I get returns from .il or the Middle East. This lead me along many wrong paths, but the method in itself wasn't bad. I just had no way to know that the sequence of sounds I perceived should be split in "vekhol" AND "ka...".
Then I saw this post on the Seeker's message board:
http://www.google.com/search?num=25&hl=en&lr=&ie=UTF-8&oe=utf- 8&safe=off&q=cdyq&btnG=Google+Search (n/t) (05/07/02 12:51:01)
Holy transliterational shit


I scanned the results pages and found this little treasure:
stat.ath.cx/logs/unibg/programisti.2002-06-30
... 48] <chervarium> hmm... imat samo na tanach [15:48] <chervarium> 75:11
VKL-QRNY RSOYm AGDO TRVMMNH QRNVT CDYQ. [15:49] <chervarium ...
62k - Cached


I was already so deep into this holy transliterational, er, thing that some letters jumped out of the screen and right into my eyes:
VKL-QRN Y sounds like the first word I could capture, that "wekhalka"...

Then TRVMMNH sounds like another mysterious word(?) in the song: "tromomenoh"... then QRN again, and then CDYQ - that's our famous tzadik!

So, that "Q" reads "k"!

The whole thing comes from the Book of Psalms (75:11), written in Old Hebrew, and is transliterated WITHOUT vowels like this:

VKL-QRNY RSOYm AGDO TRVMMNH QRNVT CDYQ

The variant WITH vowels that I found is:
vekhol-qarnêy reshâ`iym 'aghaddêa` teromamnâh qarnothtsaddiyq

I took "vekhol-qarnêy" to be a full word. It never occurred to me to use "vekhol" alone and search all +50 links or to combine it with "tzadik" and get 6 links AND 3 HITS! Not bad for efficiency. That's where I hit an ugly brick wall.



jeff:

A note i'd like to make about lyrics: generally the song title is some part of the words themselves, as in Silent Night, Big Bad John, Roll Over Beethoven... if you see what i mean ...

As many dictionaries as I searched I kept coming up empty for the word 'qarney', and you can kinda see here though how a 'q' might be pronounced  :)

We might also take an english jump here in seeing that the 'q' is pronounced as having a 'k' sound: tsaddiyq, ’; from H6663; just:--just, lawful, righteous (man).


Perhaps we should have a seperate z404.htm page here instead??

>Target = (x) z404
>identify and locate song
>1. what steps would you take to locate a song you hear?

Divas Galore / z404
by vvf

First of all, I tried my memory. In this case there's a better chance I heard the piece before, and that's why I think Fravia wrote that "z404 might be easier".

That z404 made me feel real stupid... I wish I had found the file earlier. It just never occurred to me +F was talking about a second file; I thought it was just another clue related to 909. Weren't it for anonot's post, I would never have discovered it. Now, I confess I do not remember having heard "ruhe sanft" before, but I was somehow familiar with Mozart and liked his music enough to recognize his style. When I listened to z404 for the first time I thought "this must be Mozart"... It's the style of the man. It was a hunch, of course, but it proved right. The "style" of the music, its... grace, pushed the Mozart button in my mind.

I also recognized the language as being German, and that isn't easy given the funny way in which opera singers sing:). I can't speak German, however I can understand some, and I also know a few words I can use if I ever lose my way in downtown Muenchen and can't find a bloke who speaks English:). More seriously now, I can decipher some German words if I hear them; having German-speaking friends has helped me here.

So there were two clues I could use: Mozart (a hunch) and German. I put my headphones on and looped the first minute of the song. The words seemed clear enough for recognition. The first sentence I got is "Da, mein Bild". The voice says "tah, mein bilt", but I know that in German that "tah" might be spelled "Da", the same for bilt/bild. Also I already knew "Bild" means "picture". I also deciphered another word: "Leben". And that's already plenty to start a search like:

http://www.google.com/search?q=%22da+mein+bild%22+leben

Mozart's in there! Zaide's libretto. I read about Mozart's unfinished opera, found by his wife years after Mozart's death, sold for a modest sum of money, and staged decades later. Sooooo, zeee 404 is Zaide, more specifically the aria calle "ruhe sanft" (rest peacefully)... It could not be that simple though. German isn't exactly a rare language and many people are bound to speak it and understand the lyrics. The true problem arises when one thinks of the army of performers that probably recorded that aria. How on earth would I recognize our lady? There's dozens and dozens of possible records, all over the place...

Among my first attempts, I visited a friend who's a composer and who graduated the conservatory several years ago (his wife is a soprano). I played "ruhe sanft" and asked them who they think is the performer. Well, they didn't recognize the aria in the first place, which was a bit disappointing. Second, they said that unless you heard it before (and remember it) you can't tell who sings. As simple as that. You have to be a fan of the singer to be able to do it. So it's back to searching.

I teamed up with jeff on this one. We both tried the obvious clue first, namely track length. It's hard to believe that all performers will achieve exactly the same time. Unless the track has been tampered with (easy using CEP or other similar tools), track length is a good clue. We went hunting for info and clips on several sites like freedb.org, gracenote.com (formerly cddb.com), and of course amazon and yahoo. The first two only provide info on the albums, including track order, through their own search engines. The last two are very valuable because they also includes clips.

jeff:

I had no idea what artist or song name z404 was at all ...i couldn't tell you mozart from zartmo ... thankfully vvf wrote me and told me the song is mozart --- the song is ruhe sanft --- artist unknown.

Now you would think this knowledge would make it much easier to find (and perhaps it is?) but as vvf says when you investigate and suddenly see that dozens if not a hundred artists have sung and performed this song--- here alone you would find nearly 40!!! --- a small list i assure you. How to determine which is the one that matches your songfile? Well then you have to find some type of eliminators ...

We began with the idea that track length would need to match. Here's a snippet from messageboard (jeff's post :)

Zaide

* Wolfgang Amadeus Mozart. Singspiel in two acts. 1780.
* Libretto by Johann Andreas Schachtner, after Das Serail (The Serraglio) by Franz Josef Sebastian.
* First performance in Frankfurt on 27th January 1866.

CHARACTERS:
Zaide soprano
Gomatz tenor
Allazim bass
Sultan Soliman tenor
Osmin bass
Zaram, Captain of the Guard speaking rôle
Four Slaves tenors

Gomatz, enslaved, sleeps after his work, watched by the Sultan's favourite Zaide. They fall in love. With the help of Allazim, they escape. The Sultan expresses his anger. The lovers are captured and Zaide acknowledges the Sultan's former generosity, although she prefers freedom. She pleads for the life of her lover, and Allazim tells the Sultan of the brotherhood of man. It may be supposed that a magnanimous gesture from the Sultan would have ensued.

The full text of Mozart's unfinished opera is lost. Written in 1780, it may be seen as the precursor of Die Entführung aus dem Serail (The Abduction from the Serraglio), which he set two years later, when he had established himself in Vienna. The best known aria from the work is Zaide's Ruhe sanft, mein holdes Leben (Sleep softly, my dear one), as she watches over the sleeping Gomatz. Unlike the later work, Zaide makes no attempt at pseudo-Turkish music, although it may be supposed that there would have been some place in the completed opera for the Turkish effects with which Mozart and his contemporaries were very familiar.

Ileana Cotrubas (Soprano)
Edith Mathis (Soprano)
Ruth Welting (Soprano)

In z404 at very beginning of song there seems to be a big difference in z404 than most I am listening to: z 404 sounds more like this one than any others because this one does NOT begin with a rolling "R": http://www.iclassics.com/iclassics/album.jsp?selectionId=899

Other examples of this rolling "R" are like these ... note the difference at very beginning; these roll a sound like an fluttering "R".

So, hummmmmmmmmm... great voice ... don't think so though: Lucia Popp
http://www.cdnow.com/pagename=/RP/CDN/FIND/album.html/itemid=800853/from=sr-2381276-2
This one by Lucia Popp sounds very very much like it... yet a couple words sound a tiny bit different. Arrrrrrg I will never do this:
#2 ...

http://www.cdnow.com/cgi-bin/mserver/SID=1839767486/pagename=/RP/CDN/CL
ASS/muzealbum.html/itemid=800853

Emma Kirkby (Soprano) - nope
http://www.iclassics.com/iclassics/album.jsp?selectionId=2350

Natalie Dessay - nope
http://www.iclassics.com/iclassics/album.jsp?selectionId=9991

Kathleen Battle (Soprano) - nope
http://www.iclassics.com/iclassics/album.jsp?selectionId=715

Less pronounced 'r' rolls. Definitely not this one - not sure who artist is:
http://www.iclassics.com/iclassics/album.jsp?selectionId=3341

Not this one:Lucia Popp (Mezzo-Soprano)
http://www.iclassics.com/iclassics/album.jsp?selectionId=852

Nope: Krista Bennion Feeney (Violin) Renée Fleming (Soprano)
http://www.iclassics.com/iclassics/album.jsp?selectionId=1753

Rennee Fleming again... nope
http://www.content.loudeye.com/scripts/hurlPNM.exe?/~ttt-600005/0164613

_0109_07_0002.ra

The sountrack on this album is 6:25 :)

http://216.239.53.100/search?q=cache:EWxwkim9B5IC:www.iplusmusic.com/music/classical/album_detail.asp%3Falbum_id%3D2851%26artist_id%3D457+%22Ruhe+sanft%22+%226:25%22&hl=en&ie=UTF-8

...............Lott, Felicity

Try to find:
8. Mass No. 17 for soloists, chorus & orchestra in C minor (fragment,

"Great Mass"), K. 427 (K. 417a) Kyrie
Composed by Wolfgang Amadeus Mozart
Performed by Academy of St. Martin-in-the-Fields with Felicity Lott

Another page above said zaide was ( 6:25) minutes long and said track 17... the complete collection shows track 17 as:

17. Zaide, opera, K. 344 (K. 336b) Aria, Ruhe Sanft
Composed by Wolfgang Amadeus Mozart
Performed by Academy of St. Martin-in-the-Fields with Felicity Lott

It ain't this one:
http://handouts.eas.muohio.edu/schabetc/Audio/Mozart%20-%20Zaide%3b%20Aria,%20Ruhe%20Sanft.mp3

Christine Schäfer ... don't think so:
http://www.iclassics.com/iclassics/album.jsp?selectionId=2704

(Nice little index ---http://handouts.eas.muohio.edu/schabetc/Audio/)
recent graduate jess does it here:
http://www.duke.edu/~jmm21/

Sandrine Piau --- i don't think so on this one... she's great too

...The orchestra sounds different too??? This one is: Gottfried von der Goltz
http://www.amazon.fr/exec/obidos/clipserve/B00005UEBE001014/402-8155797-1743353


... Karita Mattila ... mothers to be ...

......Lynn Dawson

........... Kiri Te Kanawa

............... Erie Mills

http://216.239.33.100/search?q=cache:IjxByRr5lWMC:www.classical.net/music/recs/reviews/h/hmu07205a.html+mozart+soprano+zaide&hl=en&ie=UTF-8

......(Though there is one on Orfeo with Judith Blegen, led by Leopold Hager in Salzburg, that I have not heard)

jeff


And you know the funny thing is  i looked for Judith quickly simply because the sentence indicated that it was one... that I have not heard... indicating to me that it might be hard to find ... therefore a perfect song for a riddle.

Arrrrgggg!


(Zaide) Other than the title of this song ... what meanings does zaide have? Well one thing I asked google was for "names"
Surprizing to me it jumped up with arabic and hebrew meaning (not surprizing perhaps after finding that the aria is Turkish inspired).
Zaide == Zaide male Yiddish elder
That kind of thru me off as the third aria zaide seems to be more about a female, but it made me think that z404 might be the f+ clue to think Hebrew/yiddish for 909.

vvf:

The Hall of Mirrors

Not that much arduous lyrics recognition needed for z404. The difficulty in this case is that virtually any soprano on this planet recorded that same piece...

Now, how to recognize the right one without going insane? The logical steps we took were illustrated by jeff (track length, etc.). One of his VERY valuable findings is the amazon trick: change the .com part of the URL to .co.uk or .fr or .de or whatever you like and see the amazing thing: some countries let you listen to more clips than others!

Two other valuable tools I used a lot were www.cddb.com (gracenote) and www.freedb.org. Both have good search engines, while gracenote seems to have indexed more albums than freedb. They're real handy for track title & listing.

We also assumed one possible extra clue, namely that the track number should be "4", as in z404, because the track number of 909 had been "9".

Even with this clue, it's not easy to spot the right performer. As a matter of fact, jeff already mentioned the right lady & record once but did not check it - me neither :) But what got me on the right track, if I may say so, is a clue by Fravia to think what would be a proper place and orchestra for this z404 record. I tried Vienna, then jeff mentioned that Mozart was born in Salzburg (which he left for Vienna at 25), so I concentrated on Salzburg. I don't remember exactly how I searched here, and I never took notes, but what I do recall is that I came across another promising keyword, Mozarteum (check it for yourselves:). Salzburg, mozarteum, zaide & soprano were the keywords I constructed my search on, checking the resulting albums to see what they listed on track 4. One of the records I checked was:
Author:		Mozart, Wolfgang Amadeus, 1756-1791.
Title:		Zaide [sound recording] / W.A. Mozart.
Publisher:	Munchen : Orfeo, [1990?].
Pages:		2 sound discs : digital, stereo. ; 4 3/4 in.
Notes:		Compact discs. Singspiel. Sung in German. Libretto by Johann Andreas Schachtner.

Judith Blegen, soprano ; Werner Hollweg, Thomas Moser, tenors ;
Wolfgang Schone, Robert Holl, basses ; supporting soloists ;
Mozarteum-Orchester Salzburg ; Leopold Hager, conductor.
Reissue of: Orfeo S 055 832 H (1983).
Ed. recorded: Neue Ausgabe samtlicher Werke. 
Recorded Jan. 16-21, 1982.
I found a sample here: http://shopping.yahoo.com/shop?d=product&id=1921562654

I listened and it sounded goooooooood!!! The tempo was right but the voice was slightly different. Could that be due to a difference in bitrates, quality, etc. ? To find out, I launched CEP and did that little acid test. Yes it's the same! Search over.

A little extra here would be to find one of those mp3 recognition engines that already fingerprinted our z404.


note: from jeff ... or extra credit for getting the whole kit&kaboodle here.


vvf:

How To Identify Music, or

CEP 1.2 - Fingerprinting the Suspect

CEP 1.2 is available as a demo version. One can always try to find a 'doctored' version on some repository located in Russia or other Eastern countries that are less concerned with copyright issues for the time being, but for comparison purposes I think the demo version of CEP 1.2 is largely enough. CEP 2.0 has been released and is almost 20 MB (=much larger). Once again, it could be nice, but it's overkill.

If you have a CreativeLabs board there's a good chance you have WaveLab Lite on one of the board's CD-ROMs. It's equally good, and has the same features as CEP 1.2. Actually, the full version of WaveLab 3.0 is a treat and can do amazing things to sound & music... but that's only if you wish to get involved in more serious music editing.

Let's stick to CEP 1.2, easily available. If I remember well, you won't be able to save files with it, so you have to take this into account.

The most useful thing about CEP in a musical or sound-related search is that it can compare two (or more) wavs with an extremely high degree of precision. It lets you compare them visually, at any level of detail, or listen to them in sequence or at the same time, in sync.

How does that work? Just launch CEP, go to the View menu and select Multitrack view (F12). You'll see several empty tracks on CEP's workspace. Place the mouse somewhere inside track 1, right-click, select and then wave from file... Pretty obvious, huh?

One thing though: I had some unpleasant surprises with CEP and mp3-compressed files, so better make sure you convert those to .wav before importing into CEP. My winamp 2.76 does that conversion very easily (go in options/preferences/output and select nullsoft disk writer - and don't forget to press configure to tell winamp were to save the resulting wav files, duh). After this, Winamp will act as a decompressor; you won't hear anything, but it will take whatever file you feed it and convert it to a wav in the location you specified. If you were not aware of this little Winamp trick, you'll see it's very useful for other things too - like converting those mp3s to a format that can be used to write audio tracks. Someone told me the latest Winamp version lacks this facility, so if you don't have Winamp just download an older (free) version like 2.7x.

(IMPORTANT: immediately after you convert the .mp3 file(s) to .wav, remember to switch back winamp/preferences/output to whatever it was, typically "nullsoft waveout plugin", or else you won't hear anything anymore.)

The sound format I use is by default PCM,44,100 Hz, 16 Bit, Stereo - widely known as CD quality.

OK. You got your first file, file A, in the first track. Now let's say you want to compare with file B, a similar piece of music of which you're not quite sure if it's the same... Well, use the same steps to put file B in track 2, or just drag-and-drop it there (it works!).

Once files A and B are in place, all you need to do is play with them until you get them in sync. You do that using the right mouse button: click on the wav and hold it, then move the mouse left or right until you believe both file are overlapping. Hit the rewind button at the bottom-left of the screen. Then hit "Play". Listen... That's all there is to it.

A few tricks:

  • It sounds silly to mention, but I'll say it: you have to use the same portion of music on A and B to be able to draw a comparison.
  • The first clue of a match is visual, exactly like fingerprints, so try to align the "envelopes" of the wavs as good as you can. If there are big mismatches in the envelopes, and especially lumps in the wrong places, don't even bother to align the files cause it's already clear they're not the same (ex. different tempo).
  • For a relevant comparison the amplitude on both files should be the same. If it's not, take the file with lower amplitude (slimmer envelope) and double-click on it. You are now in the "edit waveform view". Select all, then transform/amplitude/amplify and try a 3db boost. Don't go for more, it's usually enough, otherwise adjust as needed. If you're satisfied with the new amplitude, hit F12 and go back to the multitrack editor. If the wavs' envelopes look more similar then you're doing fine.
  • On the left of every track there are "m" and "s" buttons - that's mute and solo. Make sure they are not pressed if you want to hear both files. You can play with them later:)
  • If you hear a constant echo, you're almost there - move a little one of the A or B wavs, to the left or to the right.

If you have a perfect match, you will hear only ONE song (or sound) and NO ECHO! If there are discrepancies, that means they're obviously not the same!

You may still have the same performer, same song, same band or orchestra, but if the recording is not the same, then at some point, in fact really soon, you will start noticing differences. That's unavoidable.

This comparison works a lot like comparing fingerprints, imho: if you got a perfect match then the recordings are  identical . You notice I wrote recordings and not files, because the files might still be different: lower bitrate, hiss, etc. These differences could impact on your ability to recognize the same performer on both files, but this little 'acid' CEP test will tell you for sure what the truth is.

An ending note: you can also modify a wav in order to make it less easily recognizable. You could alter song length, increase or decrease tempo, alter the pitch... make your own personal version of any song, thus obscuring the original. Of course, you'd need a working version of CEP or similar.


jeff:


Hummmmmmmmm, I wonder --- should we maybe not name the songs ... and let people find them with our above clues??

A quote by fravia:

A solved riddle is -as you know- much less interesting than an unsolved one...
(¯`·.¸(¯`·.¸ fravia+ ¸.·´¯)¸.·´¯)


So with that in mind, now that we have pointed out all of the clues and 909 How To's ...?? You should be able to find 909 on your own. Yes? :)

Ok ok ok... we can't really do that without leaving out some of the most phun parts of our quest... as we seek and learn more and more and more about our subject [in this case identifying music] we begin to develope better ways of thinking about the subject. We begin to think of new keywords to use in going that extra step beyond.

In the case of 909 I had already identified it, and found it; (actually someone named Yeah! had posted just before me, saying he had busted it --- butt we know not who he is, nor have we ever heard from him --- the clue in his posting leads me to believe he DID bust it).

So how was finaly 909 solved?

Well ... because of fravias BIG HUGE HINT  (although many still have not solved it) I had written to him and asked if he would pop into the 909 thread and give some encouragement and hints

Open letter to Jeff (20/09/02 09:19:35) >>>> if you could every once in awhile pop into them to give hints or direction it would be better than letting these things DIE a slow death ... and then no one has learned anything <<<<<

You sure you want a hint re 909.wav?
I guessed soon or later someone would have solved it on his own, without further intervention... but if you want another hint... IT IS THERE ALREADY , since the beginning of this search on my site. Nobody seems to have noticed it, though. In fact this "other" hint is CRUCIAL in order to solve the 909 search (without sheer luck or 'zufall'). I doubt that this music exist on any other part of the web (at least not yet :-)

I think I should publish this answer on the msgboard, so that every seeker interested in this "musical side alley" has again the same 'starting conditions'.

This hint, when found, will definitely help to solve the riddle, though. A pity :-) A solved riddle is -as you know- much less interesting than an unsolved one...

I may also add that, even without having noticed this 'other' hint, YOU have all been very near to the solution anyway during an older thread... quite amazing, come to think of it... and worth an essay on EXTREMELY HARD searching, as you'll see once solved.

(¯`·.¸(¯`·.¸ fravia+ ¸.·´¯)¸.·´¯)


..............

This is where I have to tie all the things together that we worked on... I had just wrote mordred and vvf telling them that i thought the clue was the emperors clothes levis essay on the whats new page ... i was sitting there looking at the whole month when i went numb. It is because i had lists of artists names and lists of  nearly everything you can imagine, actually, when i saw the clue I simply went numb ... i recall staring at the word for several seconds:

zahavian


more precisely:
June 2002 (a 'zahavian' month)

Z 404 ... z  zaide  z zahavian?  lots of z's butt WHY a  zahavian month ?
I quickly went to google and asked for zahavian ... my mind clicking... knowing this was IT
Not to good returns ... that dam word means something to me ...
What what what... not hava, not chava, not zahavian ... butt zahava!
Quickly i went to my artists list which had whoaaaaaaa lots of z artists in it ... there i saw 2 names that jumped out at me... Ben Zahava and Zahava Seewald.

I have to disagree slightly to the "sheer luck or zufall" ... it is easy to say, and think now, but i feel, we would have snapped to it with all the info, hard work, and intuition, we had gathered, sooner or later ... simply cross referencing some of our keywords would have kept bringing UP a particular url, or two, that sooner or later would have given you that 'zahavian-numbness' as you asked yourself, "Why does this song (song dummy! slap slap ) keep coming up???"

As another example of using only what we had gathered and not using 'zahava' at all, do you recall we spoke of, and wrote about, 'what type of instrument' is used in the song?

snip...
has anyone indentified WHAT INSTRUMENT is being Used in 909?

jeff


snip...
Now jeff, your question is really important. What about the instruments ? There are percussions, something that uses strings, and an accordion ? If we could find the correct name, these would be keywords for some synecdochical searching :)
loki


snip...
here the list i think

vocal cello viola violin bass accordian percussion 
Liman

snip...
The instruments I find are bass guitar,
Mordred
So, cross referencing your 'instruments list' (bass) with your "lyrics list" (vekhol) would have given you this ...

Simple on monday-morning :0 yes ?

You can see how cross-referencing can be advantageous :) and also with all that we knew we still could have found this without 'zahava'  :)

I can't begin to tell you how many 2 word combo's would have kept taking you too a url or two, that kept repeating itself in your returns... in fact a single word could take you to it also, and sooner or later we would have gone to find that damned song clip to see what it sounded like... I know i listened to hundreds  :), mostly Ruhe Sanft's ... but why did it take me months to find this damned song then? because, in regards to the lyrics, other than looking up the psalms chapters, i simply Never once yo-yo-ed down thru the returns to look for a song --- even though everyone kept telling me i had the lyrics i continued to ignore the verses because I didn't understand the language... so i never used the word "vekhol" as a keyword!

Just a few combos that fravia had sent me after i excitedly wrote him about zahava:

Vekhol zahava

zahava vocal

Zahava Seewald 9

Zahava tzadik

vekhol tzadik

and -incredibly- even vekhol 9 !!

So, what do you say now?
Abi Gezint!
:-)
(¯`·.¸(¯`·.¸ fravia+ ¸.·´¯)¸.·´¯)


As vvf points out (dang!! and he kept telling me over and over that I HAD the lyrics :) simply using the 'lyrics' would have kept bringing up these same url's to a song that kept being returned... vekhol, tzadik

Also ... you may wonder why I left #7 in our outline above, empty --- blank --- it is because one of the very first returns i went to has 'vekhol karney'  --- not ONLY mispelled but also listed at another place on the wrong track number ... As can be seen if your thought process is that the naming of the 909 may be because of the track number this could really throw you for a loop... and if they mispell it you'll never find it ... as a matter of fact because 'qarney' is spelled karney you won't find it if you use 'qarney' ... as I did so many times ... this is a case where less keywords is more :) ... any number of keywords now could find you the artist, the song, the song title--- etc ...  but all of these returns do not produce a working sound clip to compare to see if we are correct ... you now have to take the info that you now know into a database to search and find an actual sound clip or two... note that the song track Length of "3:07" would have stood the hairs up on the back of your neck as you wondered if this vekhol was to be IT.

Amazon france, canada and japan had the sound clip 3 weeks ago, although USA, germany, and uk did not-- indicating that amazon was running 2 different databases ... today however, track 9 at .fr .ca and .jp no longer have a track 9 sound clip :(

We had decided almost immediately, upon identifying 909, that the 9 probably represented being on track #9 ... that, just perhaps, ment that we could eliminate, and concentrate, on zaide, ruhe sanft songs, on track 4 only... this allowed us to trim our list of 50-60 ruhe sanft artists down to 4 or -5 ...

We are still not sure why the 90 part of 909 was used in the file naming,  nor the 40 part of the z404 ... we are confident that the 9 was track 9; and the 4 was track 4 ... the z probably represented "zaide"...

One theory: the 90 could mean the Label of TZADIK, because, 90== Tzadik in Hebrew alphabet ... maybe - " Tzadik, track 9"==90,9.
However, we haven't been able to tie this therory into 404 ... 40 in Hebrew alphabet == M or mem; and in some pages it is mim. Now the last 3 letters in Psamim are 'mim' but 404 wasn't tied directly to Psamim at all ... so we still brude over this...

In the case of z404 vvf had found and identified it within a few days of 909... the search was over, the riddle of 909 z404 was solved. And yet it seemed to us that there simply had to be a better and faster way to identify a songfile. Granted we had / have / learned a great deal along the way taking a pioneers trail thru this riddle that tested our thinking caps.

And yet ... still ... unsatisfied ... I began to to think about the outline we have drawn up ... ask an expert.

Who is the expert for identifying songfiles? It occurred to me it would be the Music Industry itself and their invasive products that threaten to come into our computers and read our files without our knowledge and in their own literature they say they can identify and DELETE your files...

HOW do they do it?

In my searches for experts and identifying music ... a program kept coming up in my searchs that rang a bell; a program called clango ... that name rang a bell because i recalled a conversation with Stilleto dealing with this program:

909/z404 - Music Identification - Eureka? (05/09/02 00:15:11)
    Check these out... (dead or alive - dead, use archive.org)
    http://www.clango.com/
    http://sourceforge.net/projects/freetantrum/
    http://www.relatable.com
    http://www.canoe.ca/CNEWSTechNews0201/03_kendall-can.html
    http://www.audiblemagic.com
    http://www.freetantrum.org/
    http://www.tuneprint.com/
    http://www.etantrum.com/

    Theory....
    Use program to identify z404.mp3. If it works (possible), use to identify 9009.wav. And voila!

    Clango was taken offline. Can anyone find it? :)

    Stiletto
Stiletto

I did at the time try to find this proggy but didn't spend to much time on it.

Although I hadn't found clango I refreshed my search for it now because i wanted to gather keywords to expand my search into hopefully identifying some OTHER proggie that advertized producing similar results.

And as i investigated it [not finding it anywhere btw :( ] I began to read the idea behind it and certain keywords kept popping out ... audio fingerprinting.

Slap me twice for not paying more attention to Stilettos postings ... I wasn't understanding, and so after a few attempts to figure it out I moved on ... :(

One of the things I had thought of early on, was to look for "name that tune" as keywords. I figured those words might lead me to someone on the web or googlegroups who was trying to find a tune... and in the answers i might find some clues (it sortof worked) it led me to ideas like melodymonkey and the parsons codes idea and lots of white papers and articles on radio stations... arrrgg ... had i only crossreferenced another word with "name that tune" I would have been lots closer :) ... "name that tune" fingerprinting

After 909 was solved and we began writting this essay i just wasn't satisfied that a 250k How To outline that took 4 months to solve was good enough, and so I kept looking for the cherry to put on the top of this essay ... this audio fingerprinting just kept coming up...

I am not going to take you into all of the notes and links and places i went too ... suffice it to say if you look it up you begin to see that the music industry has INDEED hired various companies who have developed proggies to identify acoustic fingerprints of songs. These fingerprints can be identified with as little of 5 seconds of sound. These fingerprints in their databases can ignore real or fake ID3 tags; they can ignore songfiles names and other disguises; they don't even need to read watermarks or any other tricks

Napster is one such company who had to hire and use a company called Audible Magic to fingerprint their databases so that songs could be identified, and when deemed copyrighted were prevented from being transfered or downloaded. Guess who Audible Magic is? The same company that used to give a program called clango away for free...

I leave it to you to look up audio / acoustic / music / fingerprinting / fingerprint / fingerprints / thumbprint / etc. and learn how vvf came to find a program that found and identified the 909 songfile in 2-3 minutes (and another that found it in about 15 seconds).

I never found clango though I know someone somewhere has it, since it was a free download ... I'm sure we'll find someone who has it someday... probably in some index named 'kahlango' or sumtin :) - or possibly through a file sharing utility on someone's home computer.

But there are other programs that provide the same theme :). The sad thing is that until today as i am writing this essay and including Stiletto's postings I didn't even realize until today that all of the url's he posted (I only looked into the clango one) all relate to my hours and hours of searching this theme on my own ... I recognize them now immediately as places i have been to and read!! I could have just clicked on his links and saved myself a lot of searching :) :) [ brilliant thinking stiletto ; kudos ) ]

A few days after i wrote vvf to look into this fingerprinting stuff [ I was busy building stairs outside grrrr], he once again busted 909 in an incredibly beautiful way !! :)
Here is what he wrote:


Music in the Dark

As jeff put in somewhere in this essay, the searches for 909 and z404 are natural complements: 909 is unique, whereas z404 has plenty of "cover versions" around; the former hides in the dark, the latter lies at the heart of a hall of mirrors.

What tools do we use on a search? Whatever the methods and the creative thinking, they basically boil down to our two good old friends, induction and deduction. A radically different tool would be, say, magic. Magic is another avenue of human thought which considers that under certain circumstances a specific action on an item can produce a similar action on an analogous target (that's "voodoo 101";-). Finally, something else that can work but that you can never rely on is serendipity. Some call it coincidence, some call it luck, some call it synchronicity, some call it zen and some call it zufall... You can't count on it for sure, but you just love it when it happens.

There I was, browsing through a newspaper, looking for the tech pages. I seldom read papers, and my favorite are the weekend editions, but I do it nonetheless because it's a good way to find out about new stuff I may not have been aware of. You see, I visit all sorts of news sites online and look for the things that interest me. I usually search for those things actively, that's why I tend to find the same stuff all over again :). I need a couple of papers at least once a week to bring new things to my attention, things I would otherwise not be searching for.

So I went to that weekend's edition tech page and what do I see? Hmm? Bingo! A review of a trialware called "Moodlogic" that identifies or recognizes the mp3 files on your hard drive using fingerprinting (keywords in bold). Clango and the recognition/fingerprinting discussions immediately came to mind. I had no idea this technique was possible online, for free (well, free trial anyway). I had previously read an article in August or September about mobile phone companies who wanted to provide a paying song recognition service to their customers: hum the tune or play it in the receiver for a few seconds then wait for the answer. No watermark involved, just plain recognition. The fact that such service is already available online was a big surprise, things sure move fast on the web.

Moodlogic is a client app you have to download and install. It lets you select any mp3s that you need to identify, then it samples them and queries the online engine/database. If found, it returns the artist name and the song title. Neat! It worked with 909 in seconds (for me). Another similar service is id3man, and I'm sure you will find others as well, not to mention the ones that aren't up yet because this is still a relatively new trend. Danger also lurks in the dark, so be careful what you download. As jeff pointed out, there's a very high probability that this technology will shortly be used to censor whatever mp3s you listen to on YOUR OWN computer. After all, it's clearly built for copyright purposes, cause that's where the money comes from.

These trialware services would typically give you a number of free recognitions (25 for moodlogic and id3man), after which you can always create a new account from a different computer, or find more creative ways to overcome the limitations. Anyway, 25 is largely enough for a seeker imho.

The second and most important limitation of these mp3 recognition proggies is that they don't recognize all songs. They may know all there is to know about Britney Spears and Mark Anthony, but it's surprising moodlogic recognized an "obscure" song like 909. Well, not all that surprising after all. Tzadik is a US record label if I remember correctly. It would be interesting to check if you can recognize songs from tiny European/world labels. They claim to receive input from human contributors; that's a rather statistical approach and they'll surely miss lots of stuff that way, especially if their main concern is to provide services to main copyright holders (again, the people most likely to pay for the development of such huge fingerprint databases).

As a good example, Moodlogic didn't recognize z404! Their focus is probably not on classical music anyway.

Therefore I would say all our deduction/induction based searching, lengthy as it was, is still the only sure way to find almost any target.

vvf:
(paraphrased)
    Go to:
    http://www.moodlogic.com

    - download / install

    - Feed the folder that contains 909.mp3 to it (moodlogic only works with mp3 files ... convert extension if you need too)

    - Press "Activate"

    - AUTOMAGICALLY retrieve the name of the performer & song...

    Limitations: shareware version allows only 25 queries for song recognition. Didn't recognize z404, though.
(Note from jeff: a pop-up box pops up giving you TWO options in 3 steps... Pick the second option of browsing to the folder that contains your 909.mp3 songfile ... if you pick the first radio option it will peruse your entire hardrive listing all of the songs on your drive and as it is only good for 25 songs you may find that 909 won't be in that list and that your 25 songs limitations is over ... I had to restore my computer BACK to the pevious days date and reinstall to get my 25 trys back again).

Meantime ... now knowing that such an animal existed I went and found a couple more of these types of moodinhancers ... here is another that identified 909 in about 15 seconds ... including Track, Track length and Date:

http://www.id3man.com/

The readme file was one i could finaly read and figure out --- very striaght forward ... had to register a name and an email addy at id3man ... personal stuff was all optional anyway so you can just skip through it ...

Screenshot looks like this: http://www.mp3machine.com/software/ID3Man/screenshot.shtml

It found 909 in about 15 seconds ... it even filled in the track number... it also does not recognize z404 however :( obviously it simply is NOT in their database...
We know however that z404 IS in the loudeye database ... so we probably need to find which company of fingerprinters uses loudeyes database and then we will probably retrieve z404 ... just a thought.

Long story short ... wow!!! That dang moodlogic found & identified the 909 file!!
vvf you rock and roll!!... and you put the icing and cherry on the 909 cake.

NOW ... now, this ben zsona is done ! :)

...or is it?

Gahdaye mates! Abi Gezint !

Beertime :)

jeff / vvf /

December 2002



jeff's trick for searching amazon.com:

Searching by your personal country/preference:
Be wary --- be very wary my friends.
Let's do a search for chava alberstien's album, Foreign letters in germany: amazon.de
".de" finds the album, and lists the tracks, but their are zero sound clips to listen to and test...
Change the .de in the url too .co.uk and retest ...
No sound clips available.
Now goto .com and you see it's a different database that has 5 sound clips in it ... I can tell you that the song 909 is just the opposite of / or reverse of / this above example ... in most of my tests France had a much larger 'soundclip' database than that of even .com :)

Note: the above is no longer true ... UK now has 5 song clips and germany still has none...
909 is no longer available at .fr and .ca and .jp ... obviously this amazon database at this time is being worked on and changed ...


DEFINITIONS:

Ash·ke·naz·i

Pronunciation Key (äshk-näz) n. pl.
Ash·ke·naz·im (-nzm, -näzm)
A member of the branch of European Jews, historically Yiddish-speaking, who settled in central and northern Europe. [Medieval Hebrew ’aknzî, from ’ aknaz , Germany, adoption of Hebrew ’aknaz, name of one of Noah's grandsons and of a neighboring people, perhaps alteration of earlier *’akûz, Scythians; akin to Akkadian aguzai, ikuzai, from Old Persian Saka-, Ska-.]Ashke·nazic (-näzk) adj.
Source: The American Heritage® Dictionary of the English Language, Fourth Edition Copyright © 2000 by Houghton Mifflin Company. Published by Houghton Mifflin Company. All rights reserved.
Hasidic
Ha·sid or Has·sid also Chas·sid
Pronunciation Key (äsd, ô-, hä-) n. pl. Ha·si·dim or Has·si·dim also Chas·si·dim (ä-sdm, ô-, hä-)
A member of a Jewish mystic movement founded in the 18th century in eastern Europe by Baal Shem Tov that reacted against Talmudic learning and maintained that God's presence was in all of one's surroundings and that one should serve God in one's every deed and word.
[From Hebrew sîd, pious, from sad, to be kind. See sd in Semitic Roots.]Ha·sidic adj. Ha·sidism n.
Source: The American Heritage® Dictionary of the English Language, Fourth Edition Copyright © 2000 by Houghton Mifflin Company. Published by Houghton Mifflin Company. All rights reserved. Has·sid Pronunciation Key (äsd, ô-, hä-) n.
Variant of Hasid.
Klez·mer
Pronunciation Key (klzmr) n. pl. klez·mo·rim (klzm-rm)
1. A traditionally itinerant Jewish folk musician of eastern Europe performing in a small band, as at weddings. 2. The Jewish folk music played by small, traditionally itinerant bands. [Yiddish, from Mishnaic Hebrew klê zemer, musical instruments : klê, pl. bound form of klî, instrument; see kll in Semitic Roots + zemer, music, song ; see zmr in Semitic Roots.]
Source: The American Heritage® Dictionary of the English Language, Fourth Edition Copyright © 2000 by Houghton Mifflin Company. Published by Houghton Mifflin Company. All rights reserved.
Ladino
La·di·no Pronunciation Key (l-dn) n. pl. La·di·nos
1. A nearly extinct Romance language, descended from medieval Spanish, spoken by Sephardic Jews especially in the Balkans, Turkey, and the Near East. Also called Judeo-Spanish.
Sephardi
se·phar·di
Pronunciation Key (s-färd) n. pl. Se·phar·dim (-dm)
A descendent of the Jews who lived in Spain and Portugal during the Middle Ages until persecution culminating in expulsion in 1492 forced them to leave.
[Medieval Hebrew spraddî, Spaniard, from sprad, Spain, adoption of Hebrew sprad, placename of disputed location (mentioned at Obadiah 20).]
Se·phardic (-dk) adj.
Source: The American Heritage® Dictionary of the English Language, Fourth Edition Copyright © 2000 by Houghton Mifflin Company. Published by Houghton Mifflin Company. All rights reserved.
\Se*phar"dic\, a. [From Sephardim, a name applied to the Spanish Jews, fr. Sephard, name of a place where Jews were held in captivity (--Ob. 20).]
Of, pertaining to, or designating, the Jews (the Sephardim, also called Spanish or Portuguese Jews) descended from Jewish families driven from Spain by the Inquisition.
Source: Webster's Revised Unabridged Dictionary, © 1996, 1998 MICRA, Inc.


Petit image

© III Millennium: [fravia+], all rights reserved, reversed and revealed

Page optimised for Opera. Other browsers? Couldn't care less.