core.proto
path phonexia/grpc/common/core.proto
package phonexia.grpc.common
Phonexia common message definitions for gRPC API.
Messages
Audio
Audio message in either raw or one of the standard formats.
In the case of raw format, this message contains raw audio data that
are encoded in a format specified in the configuration.
The audio content is represented as raw audio if the
raw_audio_config field is specified.
In case that the raw_audio_config field is not specified,
this message contains audio data in one of the supported standard formats
and encodings. Supported formats are WAV (various encodings)
and FLAC.
For best results, the audio source should be captured and transmitted
using a lossless encoding (FLAC or LINEAR16).
The accuracy of the speech technologies can be reduced if lossy codecs
are used to capture or transmit audio, particularly if background noise
is present. Lossy codecs include ALAW or MULAW
Important: all Phonexia microservices that are audio-based, with the
exception of MediaNormalization and TimeAnalysis,
only support mono-channel audio.
See https://docs.phonexia.com/products/speech-platform-4/microservices/audio-formats
| Name | Type | Description |
|---|---|---|
content | bytes | The audio data bytes encoded in the supported format. In case of raw audio, this field does not include the audio header, but only the data. Note: as with all bytes fields, proto buffers use a pure binary representation. |
time_range | TimeRange | Time range of the audio to process. If omitted, the
whole audio is processed. If |
raw_audio_config | RawAudioConfig | Specify raw audio metadata in case that the |
Matrix
Data type representing matrix of values. The matrix is represented as a linear array in row major format.
| Name | Type | Description |
|---|---|---|
rows_count | uint64 | Number of rows in the matrix. |
columns_count | uint64 | Number of columns in the matrix. |
values | repeated float | Sequential list of matrix values in row major format. |
RawAudioConfig
Configuration message for raw audio data used in Audio message.
| Name | Type | Description |
|---|---|---|
encoding | RawAudioConfig.AudioEncoding | The encoding of the audio data. This field determines how the raw audio data is interpreted. |
sample_rate_hertz | int32 | The sample rate in hertz (Hz) of the audio data. This field specifies the number of audio samples per second. Most Phonexia technologies resample the audio into 8kHz or 16kHz as a first step. |
channels | int32 | The number of channels in the audio data. Note: if not specified otherwise, technologies only support mono-channel audio recordings. Setting this field to values other than 1 will result in an error for such technologies. |
TimeRange
Data type representing time range.
| Name | Type | Description |
|---|---|---|
start | google.protobuf.Duration | Start time of the time range. |
end | google.protobuf.Duration | End time of the time range. |
TimestampRange
Data type representing timestamp range.
| Name | Type | Description |
|---|---|---|
start | google.protobuf.Timestamp | Timestamp specifying the beginning of the range. |
end | google.protobuf.Timestamp | Timestamp specifying the end of the range. |
Vector
Data type representing a vector of float values.
| Name | Type | Description |
|---|---|---|
values | repeated float | Sequential list of vector values. |
Voiceprint
Represents the result from voiceprint extraction.
| Name | Type | Description |
|---|---|---|
content | bytes | Voiceprint data bytes in one of the supported formats:
|
Enums
RawAudioConfig.AudioEncoding
Enumeration of supported audio encodings for raw audio data.
| Name | Number | Description |
|---|---|---|
PCM16 | 0 | 16-bit signed little-endian linear Pulse Code Modulation |
PCM32 | 1 | 32-bit signed little-endian linear Pulse Code Modulation |
MULAW | 2 | μ-law encoding |
ALAW | 3 | A-law encoding |