Recognize Custom Content
Identify your custom content via media files or microphone with internet connections.
Custom Content Recognition Service enables your product (apps or server side software) to recognize your own songs, advertisements, TV shows/series or movies by sounds.
Network Type is the type of bucket you want to use for online recognition projects or offline recognition projects.
Upload my_song_1.mp3 into the bucket you just created.
After uploading, refresh the page, state “ready” which means you’ve uploaded the file successfully.
Create a project and bind the bucket you just created to it.
Save the “host”, “access_key”, “access_secret” of your project. You can find them in the places shown below.
For using our SDKs to scan files, SDKs will use 10 seconds of an audio clip for the recognition of a single request, you will need to send multiple requests to 1 file with different parts of the file if you want to get results from different parts of 1 file.
Basically you can manage everything you see above via the Console API below:
Here are some explanations for the parameters:
Audio Engine: Audio Fingerprinting: Detect the exact same version of the files you uploaded to the buckets. Speech Recognition (Speech-to-Text): Transcribe the text from speech. Audio Source: Choose Recorded Audio for the audio source to deal with noise. Choose Line-in Audio for the audio source if you want to recognize files without noise.
Scanning Policy: Choose Traverse Scanning if you want to scan the entire file for multiple songs in different parts of the file. Choose Points Scanning if there's only one song per file.
Go to the container you just created and click "Upload" to upload or add the files that you want to scan.
Data Type: Audio: You can upload audio/video files; Fingerprint: You can use the Audio File Fingerprinting Tool to generate fingerprint files and upload fingerprint files as they are much smaller; Youtube: You can simply input the Youtube video URL if you want to detect Youtube video;
When the "Status" turns "ready" you are able to click the arrow to browse the results on the page or select the files then click "Action" then click "Export" to export the reports.
Please use the API below if you want to deduplicate the media files.