The AfptMatchStream
task receives audio data as a binary stream, and searches it for sections that match audio indexed in an AFP database.
The AfptMatchStream
task uses templates as opposed to landmarks, which improves robustness to audio mismatches at the cost of scaleability.
Parameter | Description | Required |
---|---|---|
Type | The task name. Set to AfptMatchStream . |
Yes |
AfpDb | The audio fingerprint database (if it is defined in the configuration file). | * |
Delay | The delay in updating results. | |
DelDelay | The amount of time (in terms of audio processed) between the last valid template hit being detected in a match, and that match being deleted. | |
MinDuration | The minimum duration of a section that must match before that match is considered. | |
MinHits | The minimum number of hits that a section of audio must generate before HPE IDOL Speech Server returns it as a match for a clip in an audio fingerprinting database. | |
MinRate | The minimum hit rate threshold for audio fingerprint identification. | |
MinScoreAvg | A threshold applied on the average template hit score, below which a candidate result is rejected. | |
MinScoreRate | A threshold applied to the average template hit score, divided by the length of the section matched. | |
NBest | The maximum number of matching AFP clips to include in the results for a section. | |
Out | The file to write the search results to. | Yes |
Pack | The audio fingerprint database (if it is not defined in the configuration file). | * |
PackDir |
The path to the directory containing the database. |
* |
Rnd | A window size within which a random offset shift is calculated. | |
TimeQ | The degree of quantization to perform. |
The database type for template-based audio fingerprinting is fptdb
(as opposed to fpdb
for landmark-based audio fingerprinting).
http://localhost:13000/action=AddTask&Type=AfptMatchStream&Pack=Adverts&PackDir=C:\resources&Out=Results.ctm
This action uses port 13000
to instruct HPE IDOL Speech Server, which is located on the local machine, to use template-based matching to search the audio stream for sections that match audio indexed in the Adverts
database and to write the search results to the Results.ctm
file.