You can't call artist/audio anymore; it doesn't exist. Change artist/audio to artist/news or similar.
↧
ENAPILibrary API key not allowed
↧
Spotify Artist can't be found
Works like a charm when I provide an item_id, thanks.
Two thinks. a) Is there a limit of items in an update? b) I was so clumsy and posted my API-KEY, can you please remove it from my post (I can't find an edit button)
↧
↧
Spotify Artist can't be found
No enforced limit on items in an update. I typically do updates 1K items at a time.
API key removed. -- Paul
↧
How to create a static playlist both seeded by and limited to a catalog?
I am also trying to do something very similar and getting either empty or almost empty playlists. My query is:
http://developer.echonest.com/api/v4/playlist/static?api_key=MYKEY&results=50&type=catalog&format=json&seed_catalog=CAPYHWY14291E7B5CE&bucket=audio_summary&bucket=id:CAPYHWY14291E7B5CE&limit=true&max_tempo=100&max_energy=0.3
I see from the previous posts, they were some time ago. Has the same bugs re-appeared recently? The above query used to work a few months ago with great results.
Am I doing anything wrong?
↧
Static Playlist Generation
what is the correct way to encode the request for this situation?
↧
↧
Echoprint-Server accuracy if only one song is ingested
Hi Guys,
We have installed echoprint server and also ingested "1" Song to it using fastingest.py as documented in readme.md on the github.
We are trying if echoprint server identifies the recorded (noisy) version of the ingested song using http://localhost:8080/query?fp_code=. But the weird part is echoprint server is not recognising any recorded version. I tried with recording lengths from 5 seconds to 2 minutes. I am getting error codes of type-2 and tyoe-7 (only :().
And the most surprising thing is, results are not consistent even if I query the same song (exactly same file :)). Sometimes it returns code type - 6 (Single match) or type - 7.
Please let me know if I am doing anything wrong in terms of ingestion (I doubt ingestion phase because I am ingesting only one song).
For installation of server and codegen, I followed exact steps mentioned in readme.md on github.
System: We are using Ubuntu Precise 64 bit.
Please let me know if you want any other details.
Regards Amar
↧
Get an artist's main genre
Thank you!
↧
Can I fetch albums?
Hi,
Can I fetch albums by artist id?
Thanks. Satoshi
↧
Can I fetch albums?
Sorry, no - we don't expose album info currently in our API -- Paul
↧
↧
Echoprint-Server accuracy if only one song is ingested
I tried with recording lengths from 5 seconds to 2 minutes. I am getting error codes of type-2 and tyoe-7 (only :().
Use a length of 20-30 seconds, which it's been tuned for. You can try longer than that but you probably won't see much benefit.
Sometimes it returns code type - 6 (Single match) or type - 7.
These return codes don't make sense within the context of your results, i.e., 6 means a positive match was identified from several potential candidate matches, and 7 means no match was able to be identified from several potential candidate matches. Did you ingest the same track multiple times? Can you try querying again?
Finally, the Echoprint Google Group is probably a better place to try to resolve these sorts of issues, i.e., if you're setting up your own server. This forum is fine, but more suited to questions specifically about the song/identify API hosted by The Echo Nest.
Andrew
↧
song/identify times out
Hi,
since recently , my calls to the identify API have been answered with a timeout.
curl -i -F "query=@code""http://developer.echonest.com/api/v4/song/identify?api_key=xxx"
HTTP/1.1 500 Internal Server Error
Date: Wed, 04 Dec 2013 09:54:52 GMT
Content-Length: 94
X-Ratelimit-Remaining: 19
Server: TornadoServer/3.1
X-Ratelimit-Limit: 20
Cache-Control: max-age=86400
Access-Control-Allow-Origin: *
X-Ratelimit-Used: 1
Connection: Keep-alive
Keep-Alive: timeout=15, max=100
{"response": {"status": {"version": "4.2", "code": -1, "message": "The operation timed out"}}}
My other calls are working fine.
I may have exceeded the call limit recently. But if that's the case, shouldn't I get a 429 error with a message telling me I've exceeded the call limit ?
Thanks
Tim'
↧
adding a list of songs to a taste profile?
Hi there!
I'm testing out the Taste Profiles functionality and would like to create some profiles with a fairly large number of songs in each one. The problem is that I don't have Echonest ID's for the songs. So I need to use song_name and (artist_id/artist_name) combinations to identify the songs.
I followed the directions to create a taste profile, but when I try to update it with the list of songs I get an error saying: "Echo Nest API Error 5: Invalid item block in the input: a song_name/artist_name combination can only appear once per update."
I just want to check whether I'm doing something wrong or whether what I'm trying to do is just forbidden.
The docs have a section about creating a profile with a JSON list of songs identified with name/name pairs. But I can't figure out any way to send over JSON in the create step.
Do I just need to send updates one-at-a-rate-limited-time?
↧
song/identify times out
Thanks for the report. We'll look into it, and will get back to you once we have an update.
Andrew
↧
↧
song/identify times out
We've identified the cause of the problem and are fixing it now.
Andrew
↧
adding a list of songs to a taste profile?
You should be able to add songs in chunks of a 1000 in a single catalog/update call. The error you're seeing is letting you know that you tried to add the same song and artist twice in a single update call.
You should either deduplicate the songs you're adding before you try the update, or include item_ids with your items.
↧
Track misidentified
I found another track where the identification failed.
"Wolf Heath" by Egil the Eagle gets identified as "Oh Shit (Crazy Pills)" by Abstract Mind. If you need to debug it the track can be downloaded for free at http://www.last.fm/music/Egil+the+Eagle/_/Wolf+Heath
↧
Track misidentified
liftarn,
I get no song match when I upload "Wolf Heath", which is what I'd expect since we have not ingested that song from any of our partners. Can you post the response you are getting?
The Blondie mashup has enough in common with the original that I'm not surprised that's a match. We try to make a best guess on track/upload identification since the tracks can be encoded and distorted in 1000 different ways.
DD
↧
↧
Track misidentified
Thank you four your help.
This is interesting. When I upload the file I got directly from the Mama Viol site I get the reply
{"response": {"status": {"version": "4.2", "code": 0, "message": "Success"}, "track": {"status": "pending", "song_id": "SONFXHO13E93FA08D5", "title": "För dina blåa ögons skull", "artist": "Mama Viol", "id": "TRVIGSB142C315A0CD", "analyzer_version": "3.2.2", "release": "För dina röda ögons skull", "artist_id": "AR32MWS11C8A415150", "bitrate": 192, "audio_md5": "6435aea601e260a800c9369be83cd4ee", "samplerate": 44100, "md5": "c27ef2351bc85ad5a882c1920744157a"}}}
So it's like it should be.
For a fresh download of Wolf Heath I get
{"response": {"status": {"version": "4.2", "code": 0, "message": "Success"}, "track": {"status": "pending", "artist": "Egil the Eagle", "title": "Wolf Heath", "analyzer_version": "3.2.2", "release": "", "audio_md5": "75cac7d928e729b5948374c0f3e63f70", "bitrate": 128, "id": "TRVPVAN142C2CDB21E", "samplerate": 44100, "md5": "da2a99e9193a6cf0d7f227700ae728e1"}}}
Also like I would expect.
But if I instead do it with the track I tagged using MusicBrainz Picard I get:
{"response": {"status": {"version": "4.2", "code": 0, "message": "Success"}, "track": {"status": "pending", "song_id": "SOLRBXA13E3300611D", "title": "Oh Shit (Crazy Pills)", "artist": "Abstract Mind", "id": "TRJBZRN142A32FF3A7", "analyzer_version": "3.2.2", "release": "Buttoned Up - EP", "audio_md5": "75cac7d928e729b5948374c0f3e63f70", "bitrate": 128, "artist_id": "ARCJNKY13B864F295F", "samplerate": 44100, "md5": "8a7b5ad37b11958d7e9bc5f4644418f9"}}}
I tried to tag my freshly downloaded mp3 with Picard and it produced the correct result. Notice that the audio md5 in both cases is 75cac7d928e729b5948374c0f3e63f70.
For my old version of the Mama Viol track I got
{"response": {"status": {"version": "4.2", "code": 0, "message": "Success"}, "track": {"status": "pending", "song_id": "SOBKFSB13EA9DE5C07", "title": "Whitebird", "artist": "Roy-Erik Andersen", "id": "TRPTFIQ1424789DA6F", "analyzer_version": "3.2.2", "release": "Whitebird - Single", "artist_id": "ARSPTTL13194EF3B10", "bitrate": 192, "audio_md5": "ec8a6ffa4976f055cdb47358bfb41f73", "samplerate": 44100, "md5": "0299e3f93bdc56b35fa058adda13f6ae"}}} But when I tried the fresh download tagged with Picard it works as it should. Here the audio md5 is different. Perhaps because one was downloaded and one was ripped from the CD.
This is very strange. I have listened to the tracks several times and I can verify that they contain what it says. So there must be something odd with my files that causes problem for the fingerprinter. Let me know if you want to have the files for analysis.
↧
song/identify times out
This issue should be fixed now.
Andrew
↧
jEN API: get song/track with Rdio ID
I noticed I had probably posted this in the wrong section, so here it goes again...
I'm using Rdio and jEN on android. I use the Rdio api to search for songs, and then I use jEN to try to get some EchoNest info on a specific song. The problem I'm having is when I can't seem to find a way to search for a song in EchoNest with an Rdio ID.
The way it is right now I search for a song with the title and artist I've got from Rdio, but since a song can be on more than one album (Greatest Hits etc.) It is not guaranteed that I get the correct song and in the end the correct track with a foreign_id matching my Rdio ID.
Any suggestions?
↧