93 Commits

Author SHA1 Message Date
Hoan Luu Huu
30f2189986 support playht3.0 (#119)
* support playht3.0

* update language

* wip

* update top_p and repetition_penalty

---------

Co-authored-by: root <root@af6633a5cfe8>
2024-10-09 12:12:59 -04:00
Hoan Luu Huu
ffde303446 fixed azure transcribe crashed if both endpointId and alternativeLang… (#124)
* fixed azure transcribe crashed if both endpointId and alternativeLangs are used

* fixed review comments

Signed-off-by: root <root@af6633a5cfe8>

---------

Signed-off-by: root <root@af6633a5cfe8>
Co-authored-by: root <root@af6633a5cfe8>
2024-10-07 09:52:58 -04:00
Hoan Luu Huu
7b4520c070 Playht delete circular buffer with mutex check (#123)
Signed-off-by: root <root@af6633a5cfe8>
Co-authored-by: root <root@af6633a5cfe8>
1.2.33
2024-10-05 10:02:07 -04:00
Lyle Pratt
9f7a06ce56 Update README.md (#9)
Updated ElevenLabs module to include an example of how to use the module as well as links to params in ElevenLabs docs.
2024-10-03 08:06:32 -04:00
Hoan Luu Huu
f7f8f52283 fixed google asr max duration exceeded or no audio raised jambonz_transcribe::error (#120)
Co-authored-by: root <root@af6633a5cfe8>
2024-10-03 08:04:40 -04:00
Dave Horton
3f06a24b5d not clearing mark memory properly (#118)
* not clearing mark memory properly

* race condition where mark followed by audio and buffers for mark not completely allocated
1.2.32
2024-09-30 14:44:41 -04:00
Dave Horton
8a3c001b59 wip (#110)
* wip

* #103 wip

* wip

* wip

* support both grpc (legacy) and websockets api for aws transcribe

* renaming
1.2.31
2024-09-24 09:51:06 -04:00
Hoan Luu Huu
d17a2aa9be env variable for tts stream connection timeout (#114)
* env variable for tts stream connection timeout

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* env variable for tts stream connection timeout

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* env variable for tts stream connection timeout

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-09-17 10:05:57 -04:00
Hoan Luu Huu
5fd58ba6e5 fixed different voice quality while using playht (#111) 1.2.30 2024-09-16 15:56:10 -04:00
Dave Horton
4ee08a310a Feat/mark bidirectional streaming (#102)
* initial support for mark feature in bidirectional streaming

Signed-off-by: Dave Horton <daveh@beachdognet.com>

* wip

* allow max of 30 marks on any connection

* fix send multiple json in same ws text frame (#108)

* fix send multiple json in same ws text frame

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* fix mark is not sent without more bidirectional audio

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Dave Horton <daveh@beachdognet.com>
Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
Co-authored-by: Hoan Luu Huu <110280845+xquanluu@users.noreply.github.com>
1.2.29
2024-09-04 16:10:41 +01:00
rammohan-y
d01991ed0f feat/106: converting playht speed option to float instead of integer (#107)
* feat/106: converting playht speed option to float instead of integer

* feat/106: Using strtod

* feat/106: Using strtod

Signed-off-by: Rammohan Yadavalli <rammohan.yadavalli@kore.com>

---------

Signed-off-by: Rammohan Yadavalli <rammohan.yadavalli@kore.com>
Signed-off-by: Rammohan Yadavalli <rammohan.yadavalli@gmail.com>
Co-authored-by: Rammohan Yadavalli <rammohan.yadavalli@kore.com>
1.2.28
2024-08-23 10:49:55 -04:00
Hoan Luu Huu
eec4df4b77 support variable to enable ws ping pong for jb transcribe (#101)
* support variable to enable ws ping pong for jb transcribe

* update ping/pong time duration

* fix review comment

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-08-22 07:33:05 -04:00
Dave Horton
b003ab0875 revert change that added support for aws language model, as this does not work on our earlier version of aws-sdk-cpp and recent versions have hugh performance issue on debian 12
Signed-off-by: Dave Horton <daveh@beachdognet.com>
1.2.27
2024-08-13 10:53:46 -04:00
Antony Jukes
81ceddf3d2 Added AWS_LANGUAGE_MODEL_NAME (#99)
Co-authored-by: ajukes <ajukes@callable.io>
1.2.26
2024-08-12 11:03:26 -04:00
rammohan-kore
110a12d5a5 feat/856: added "=" for version parameter at line #175 (#100)
https://github.com/jambonz/jambonz-feature-server/issues/856

Signed-off-by: rammohan-kore <rammohan.yadavalli@kore.com>
2024-08-12 09:09:23 -04:00
Hoan Luu Huu
fe1e4dcf11 deepgram tts support on-premise (#95)
* deepgram tts support on-premise

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

* fix review comment

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-08-08 07:42:47 -04:00
Hoan Luu Huu
f828171b3b support jambonz transcribe with multiple sampling rate (#98)
* support jambonz transcribe with multiple sampling rate

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-08-07 14:40:22 -04:00
Dave Horton
e717ca7dd3 fix: gain was being ignored in sayOnTrack (#97) 1.2.25 2024-07-31 15:19:11 -04:00
Dave Horton
3ce819b7c9 fixes for resampling and handling odd-number byte stream (#88)
* fixes for resampling and handling odd-number byte stream
1.2.24
2024-07-19 16:45:45 -04:00
Hoan Luu Huu
56df923cdb support elevenlabs previous_text, next_text (#75) 2024-07-19 10:22:59 -04:00
Vinod Dharashive
911ace221c Implementation SpeechServiceResponse_PostProcessingOption (#87)
Signed-off-by: Vinod Dharashive <vdharashive@gmail.com>
1.2.23
2024-07-07 09:17:20 -06:00
Dave Horton
47c4de4791 clear cached bidirectional audio when we get killAudio message (#86)
* clear cached bidirectional audio when we get killAudio message

* wip
1.2.22
2024-07-02 21:47:07 -06:00
Dave Horton
3459188bb6 use trylock on mutex (#85) 1.2.21 2024-07-01 08:16:31 -06:00
Dave Horton
d6e246d84c fix: discard incoming binary or text frame if we are shutting down to avoid deadlock (#84) 1.2.20 2024-06-28 07:28:35 -04:00
Hoan Luu Huu
de676ddc81 append null to ostringstream will make c_str return part of a string. (#83)
* append null to ostringstream will make c_str return part of a string.

* fixed review comment

Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>

* fixed review comment

Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>

* fixed review comment

Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>
2024-06-28 07:08:40 -04:00
Dave Horton
d41bd15816 enable audio logging if env AZURE_AUDIO_LOGGING is set (#78)
* enable audio logging if env AZURE_AUDIO_LOGGING is set

* wip
1.2.19
2024-06-25 20:58:28 -04:00
Dave Horton
8bd20703b8 Fix/azure tts no device output (#79)
* enable audio logging if env AZURE_AUDIO_LOGGING is set

* wip

* per discussion with microsoft, add nullptr to creation of speechSynthesizer to ensure it knows we do not want it to play to device

Signed-off-by: Dave Horton <daveh@beachdognet.com>

* logging

* fix bug in creation of config string

* fix ticket 230 - Microsoft TTS having configuration data as part of audio generation

* azure transcribe, resuse existing cap_cb if azure configuration is changed

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* clean up azure code for how to re-create gsstream when configuration is changed

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* fix review comments

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* fix review comment

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* fix review comment

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wrap function in try catch

---------

Signed-off-by: Dave Horton <daveh@beachdognet.com>
Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
Co-authored-by: Hoan HL <quan.luuhoang8@gmail.com>
1.2.18
2024-06-23 14:54:36 -04:00
Dave Horton
2e553631dc implement silence as remove followed by add, fixes https://github.com/jambonz/jambonz-feature-server/issues/768 (#80) 2024-06-21 08:56:08 -04:00
Hoan Luu Huu
2a94213668 support single azure connection for same channel (#73)
* support single azure connection for same channel

* wip

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
1.2.17
2024-06-01 06:59:24 -04:00
Hoan Luu Huu
6ad663f9d9 whisper_tts should use rate to do resampling (#70)
Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
1.2.16
2024-05-29 08:13:38 -04:00
Hoan Luu Huu
c945a2de5a verbio stt/tts mods (#65)
* verbio tts mod

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* mod_verbio_transcribe

* wip

* wip

* wip

* wip

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* verbio stt

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-05-29 08:05:54 -04:00
Hoan Luu Huu
466dec7a6f mod_custom_tts (#53)
* mod_custom_tts

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* add custom_vendor to mod_dub

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-05-28 19:27:17 -04:00
Hoan Luu Huu
097ac337cc mod_vad_detect (#69)
* mod_vad_detect

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* mod_vad_detect

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-05-28 10:32:39 -04:00
Hoan Luu Huu
45ecf151c3 Delete tts tmp file if playback_start event has not been sent (#68)
* Delete tts tmp file if playback_start event has not been sent

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* Delete tts tmp file if playback_start event has not been sent

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-05-26 09:29:05 -04:00
Hoan Luu Huu
d1d0f5dd77 jb_transcribe correctly unlock mutex before cleanup (#66)
* jb_transcribe correctly unlock mutex before cleanup

Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>
2024-05-23 07:48:02 -04:00
Dave Horton
e4a27ae133 minor logging
Signed-off-by: Dave Horton <daveh@beachdognet.com>
1.2.15
2024-05-14 06:59:11 -04:00
Hoan Luu Huu
19f20bf0e7 Fix/azure white noise (#63)
* azure return odd bytes audio, that azure tts mode generate white noise

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-05-14 06:55:44 -04:00
Hoan Luu Huu
b019a634bd Fix playht timeout while downloading audio (#61)
Signed-off-by: Quan HL <quan.luuhoang8@gmail.com>
2024-05-10 12:18:32 -04:00
Dave Horton
f0b304b8a1 session mutex not properly unlocked causing stuck channels (#60)
Signed-off-by: Dave Horton <daveh@beachdognet.com>
1.2.14
2024-05-02 10:35:54 -04:00
Hoan Luu Huu
be3714465b support aws sessionToken (#58)
* support aws sessionToken

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* fix code style

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
1.2.12
2024-05-02 08:28:01 -04:00
Hoan Luu Huu
b495dba126 reset response when cached speech interface feeded new text (#59)
Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-05-02 07:24:57 -04:00
Hoan Luu Huu
33750b5420 mod_deepgram_transcribe keep alive connection (#56)
* mod_deepgram_transcribe keep alive connection

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* wip

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* read use single connection configuration from env var

* remove deepgram media bug buffered data

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* fix review comments

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
1.2.11
2024-04-30 08:00:59 -04:00
Dave Horton
f8604656e3 minor logging 2024-04-29 18:09:23 -04:00
Hoan Luu Huu
916332d91d elevenlabs to use http 1.1 (#57)
* elevenlabs to use http 1.1

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* update the rest of mod tts

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* add env variable for disabling http2

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
1.2.10
2024-04-29 07:51:09 -04:00
Hoan Luu Huu
33fee93ec7 free maloc variable on mod_audio_fork (#54) 1.2.9 2024-04-23 09:15:11 -04:00
Hoan Luu Huu
41aebafd1c fix eleevnlabs callsession stuck (#55)
* fix eleevnlabs callsession stuck

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* fix read sample rate from session is not needed

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
2024-04-23 09:14:43 -04:00
Andrew Golledge
bd69d476e7 Add Error Code Check to Google V1 Implementation (#51)
* Put the check for `grpc` error code 0 in the Google Speech-To-Text v1 as well.

* Distinguish between two types of error object in `grpc_read_thread`

* Improve naming of JSON field

* Correct error in JSON field name.

* Add sign-off to previous commit

Signed-off-by: Andrew Golledge <andreas.golledge@gmail.com>

---------

Signed-off-by: Andrew Golledge <andreas.golledge@gmail.com>
2024-04-22 19:20:15 -04:00
Dave Horton
fea51d5ecf initialize variables to avoid possible junk values
Signed-off-by: Dave Horton <daveh@beachdognet.com>
2024-04-22 16:06:32 -04:00
Dave Horton
83a2d1d730 modify cache folder name and various fixes from testing tts streaming (#50) 1.2.8 2024-04-18 11:28:17 -04:00
Hoan Luu Huu
3f642467eb add properties to tts span for mod_*_tts (#45)
* add properties to tts span for mod_*_tts

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* support deepgram tts span

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* support tts span for playht

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

* support rimelabs tts span

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>

---------

Signed-off-by: Hoan HL <quan.luuhoang8@gmail.com>
1.2.7
2024-04-14 09:13:54 -04:00