Hrmm. I'll try on my machine tonight - I am again suspicious of Windows 7 vs. Windows Server though...
Best, Thor
↧
cAction import error module not found
↧
cAction import error module not found
So can you get any ffmpeg command to work? For example,
en-ffmpeg -i input.wav -vn -ar 44100 -ac 2 -ab 192k -f mp3 output.mp3
Best,
Thor
↧
↧
cAction import error module not found
Nope. Same error regarding DV input stream.
↧
Energy Calculation Change?
Hi, We also noticed that the energy values currently being returned are significantly higher than they were back in mid-June. Can we rely on those new values?
Sample results: Luke Bryan; Suntan City; Energy as of 6/2013 = 0.615197590677048; Energy as of 7/29/2013 = 0.88867277 Bruno Mars; When I Was Your Man; Energy as of 6/2013 = 0.149327560024323; Energy as of 7/29/2013 = 0.271887 Mumford & Sons; I Will Wait; Energy as of 6/2013 = 0.501349644524231; Energy as of 7/29/2013 = 0.7938127 The Lumineers; Ho Hey; Energy as of 6/2013 = 0.343775706198965; Energy as of 7/29/2013 = 0.47053275
↧
Minor change in HTTP Response Status Code
Hey all - We are making a minor change in how we notify your application that it has hit the rate limit. This is not a change to your rate limit, it's merely a change in the format of the response we return.
Currently, when you reached a rate limit, we return an error response that included an HTTP status code of 403. This status code is changing. On August 1, we will start returning an status code of 429 when you've exceeded your rate limit. If you rely on the HTTP status code to determine if you've exceeded the rate limit, you should update your code to check for the new 429 response code. If you are using pyechonest, you should update to the latest version.
-- Paul
↧
↧
Change in default for the DMCA parameter in the playlisting API
We are making a minor change in how we support DMCA compliance in the dynamic playlist API. Specifically, we are changing how we interpret dmca=true. Previously, when a playlist was created with dmca set to 'true' we used our 'stylea' set of rules to enforce DMCA compliance. Starting this week, if you create a playlist with dmca set to 'true' we will use our 'styleb' set of rules to enforce DMCA compliance. With 'stylea', skipped songs are not considered to have been played for DMCA conformance purposes. Whereas, with 'styleb', skipped songs are considered to have been played for DMCA purposes.
'styleb' is a more conservative interpretation of the DMCA rules. Most developers will not need to make any changes to their code as a result of this change. If you are creating playlists with the dmca parameter set to true, you will see slightly more conservative scheduling of songs in the playlist if the user skips songs.
-- Paul
↧
Outdated Results in Song search
manodrum - Sorry for the delayed response, somehow I missed your post.
We currently don't expose song release year yet. It is coming soon. Until then, filtering by song_hotttnesss is the best way to go. We also have some pseudo-genres that you can use to get current music. Here's an example call:
http://developer.echonest.com/api/v4/playlist/static?api_key=your_api_key&results=100&type=genre-radio&genre=current
-- Paul
↧
Preview Url
NinjaCoder - The preview audio is provided by 7Digital. You need to agree to 7 Digital's TOS to use their assets. Read this for all of the details:
http://developer.echonest.com/sandbox/7digital.html
You need to add the 'limit=true' flag to your calls. This will limit results to songs that are in the 7digital catalog, which will always have the preview urls. Since the preview audio is maintained by 7Digital, their lifetime is set by 7Digital. Historically, they've been stable, but there's no guarantee.
-- Paul
↧
Energy Calculation Change?
Hi slegroux, guitarded. We are always working to improve our analysis algorithms, so you may see changes to date now and then. If you look at the detailed analysis (pointed to by the analysis_url in the audio_summary), you'll find a metadata block that includes the analysis version:
"meta": {
"analyzer_version": "3.1.1",
"detailed_status": "OK",
"filename": "/mnt/mp3/025545615.mp3",
"artist": "",
"album": "",
"title": "",
"genre": "",
"bitrate": 208,
"sample_rate": 44100,
"seconds": 207,
"status_code": 0,
"timestamp": 1341602264,
"analysis_time": 33.09159
},
You can use the analysis version to learn if and when the analysis has changed. We will also post to announcement forum on any future analysis updates with details on the changes so you won't be surprised if/when values change.
↧
↧
Energy Calculation Change?
Folks,
In addition to what Paul noted, we constantly get new versions of songs with the same name (e.g. live versions, remixes, extended or shortened). These will usually have different audio attributes from the "original". Take for example the Mumford & Songs song "I Will Wait":
http://developer.echonest.com/api/v4/song/search?api_key=API_KEY&artist=Mumford%20&title=I%20Will%20Wait&bucket=audio_summary
You can often use other filters like song_category or duration to find a particular version.
= David =
↧
Possible memory leak in ENMFP Codegen
I'm using ENMFP 3.15 (C++) under Ubuntu (64bit). I have a class generating fingerprints in a separate thread for an entire song. Whenever I create a Codegen object:
Codegen codegen(sampleData, size, 0);
and it runs out of scope (or creating a pointer and immediately delete it afterwards), about 20-30 mb of memory is never returned to the system. I've run this test about 20 times with the same result. The moment I remove the single statement that creates the Codegen object, the memory loss doesn't occur.
I haven't run Valgrind on it, but the memory is lost - it can't be reused by the program afterwards. Since ENMFP is closed source, I can't dig through the code to find the "possible leak". Anyone else experiencing this problem?
↧
How does the Infinite Jukebox get its track IDs?
So, seven months later, I'm at the point where I can open up a page's source and see how it gets its information. (in Python, no less)
http://hasteb.in/nucefeyegi.py
Turns out there's a json api that gets called with all the information in it (example: http://static.echonest.com/infinite_jukebox_data/TRYDDPR1401CA91AAF.json )
Sorry for not replying sooner, and sorry for the necro, and sorry for the slightly-off-topic post.
↧
cAction import error module not found
I found that the version of en-ffmpeg was very old. After updating to a newer version I was able to get passed the error. The script now runs successfully :)
↧
↧
Uploading private songs
Sorry for the basic question, but I couldn't find information on this.
I am looking to to analyze my own songs and then experiment with them using echonest-remix. But it gave me pause when I realized I needed to upload them. What does it mean to upload the mp3s? Are the mp3s saved somewhere on other servers and is it possible for other people to then access or download them? Up until now, the mp3s have been only on my private hard drives. Anything I should be concerned about if I'm trying to keep my songs private until I sell them on cd?
↧
Uploading private songs
tunesmith,
The audio you upload is only used for analysis and not available to anyone else (or even you) for download. The track/upload method returns you a unique track ID you can use in the future to get at the track profile (summary and metadata) and detailed analysis.
= David =
↧
Uploading private songs
Hi, thank you. Is the same true for streaming and streaming-previews? Under what circumstances can someone hear the audio if I upload?
↧
Uploading private songs
tunesmith,
The Echo Nest does not serve or stream audio. Several of our Rosetta Stone partners provide streaming services of audio based on Echo Nest playlists, but user-uploaded tracks do not participate in this.
Under no circumstances can anyone listen to audio you upload to The Echo Nest.
= David =
↧
↧
Uploading private songs
Great, thank you for the clarification. When I saw that url field for preview, I wasn't sure if it was something that echo nest generates. At any rate, you might consider updating the API documents for track upload to make this clear to people who are uploading their own originals. Thanks again!
↧
Setting up Remix: en-ffmpeg not found
Same here. How to work in around? I'm using python 2.7.5
import os os.system("en-ffmpeg")
returns 32512 - i don't know how to interpret it.
↧
Setting up Remix: en-ffmpeg not found
OK, i found it - it cannot find the file. I tried copying the en-ffmpeg to my python's default path, changing it's name to ffmpeg with no result.
↧