ACTNow provides various different types to classify audio data recorded from any
device (e.g. microphone, telephone, Voice over IP...)
|
Technologies |
|
Phonetic Index Search: |
ACTNow can search extremely fast for keywords or
phrases in any recordings. In a first pass ACTNow creates a phonetic
index for the audio data. Once an index is created you can search
for any word or phrase that was mentioned in the audio data. The search speed is more than 55.000 times real-time, which means that
you can check a sound file that is 15 hours long within 1 second on a standard PC.
Currently supported languages:
- US English
- UK English
- German
- Hungarian
- Modern Standard Arabic
|
|
Speaker Detection: |
The Speaker classification type allows you identify the person currently speaking on an audio stream. ACTNow allows you to train any number of different speakers. ACTNow detects if any of these speakers can be found in the audio data.
This technology is language independent. |
|
Audio Clip Detection: |
This classification type allows you to spot pre-recorded audio
clips (e.g. jingles, advertisings, songs...) even in very distorted audio
channels. ACTNow even detects audio clips when only parts of the specified
clip can be found in the audio stream.
This technology is language independent. |
|
Speech/Music Classification: |
ACTNow can automatically distinguish between music and speech. It will inform you if the audio data
fed into ACTNow contains either music or speech.
This technology is language independent.
Please click here
to download a White Paper about Speech/Music Classification
|
|
Music Track Change Detection: |
You can train spoken words or phrases and search later for them in audio data. This can be used to spot pre-recorded words phrases in an audio stream.
This technology is language independent. |
|
Key Phrase Detection: |
You can train spoken words or phrases and search later for them in audio data. This can be used to spot pre-recorded words phrases in an audio stream.
This technology is language independent. |
|
Concurrent use of any detector item: |
|
ACTNow allows you to specify any number of different detector items at the same time. E.g. you can search for some audio clips and/or speakers while still being informed if the audio data fed into is rather music than speech. |
|
Confidence: |
|
You can specify for each item a specific required confidence level. By using this feature you can choose if a detector item should be rather easily detected or only if ACTNow is very sure that this is really found. |
|
Result Information: |
|
Each ACTNow result contains information about the position of the classification type found (start time, end time in milliseconds), in case of multiple audio channel input the channel ID, as well as the probability if the detector item was really found. |
|
Supported Audio Formats: |
|
PCM 16kHz 16Bit Mono or Stereo: Typically used for broadcast or microphone recordings
PCM 8kHz 16Bit Mono or Stereo: Typically used for recordings via telephone |
