Guest u4910 Report post Posted 10/24/2019 06:26 PM I have noticed in harsh testing that when I make a module fail and get it to repeat, when it generates TTS, it sometimes plays old wav files and sometimes they are from other TTS generations. Please advise. See attached. 1024_1359_vgEngine.txt 1024_ktTts.txt Share this post Link to post
SupportTeam Report post Posted 10/24/2019 10:14 PM Can you please advise which of these TTS generations or calls you would like us to look at. There are 75 TTS plays and 12 calls in attached vgEngine trace. Just browsing through them they all look fine, so not sure which one to look at in detail. Share this post Link to post
Guest u4910 Report post Posted 10/24/2019 10:23 PM Yeah. I was trying it a bunch of times to make it do that. It would be the higher number wav files. I'll try to figure it out or write down which one when it does it again. Share this post Link to post
SupportTeam Report post Posted 10/24/2019 10:29 PM By " when I make a module fail and get it to repeat, " do you mean when the module times out awaiting input and asks for input again? or something else? If a module that uses TTS times out awaiting input it does not make the TTS engine do the TTS generation again, it uses the sound that was generated just a few seconds beforehand which also got placed in a .WAV file. On replays VoiceGuide plays straight from that .WAV file with no TTS engine involved. It's the same sound that was played before. On SAPI TTS systems (which this is) the TTS engine creates a .WAV sound file (which takes a small fraction of a second) and VoiceGuide then plays that .WAV sound file. And when module times out awaiting input it just plays the same .WAV sound file again. Share this post Link to post
Guest u4910 Report post Posted 10/25/2019 11:52 AM So for this test, it seems clear. I deleted the wav files in temp, starting from scratch. Made a new call, did all the correct things to get tts to wav generated in the correct order. Next, I did a call and pressed less than the min number of digits. It played the wav file for the next module instead of regenerating the tts to wav like it should (tts_2_2.wav). I have it set to a 6 for min digits and 8 for max. I have only a path for {success}. Should I have a {fail} to restart the same module even though pressing less than min digits should have the module repeat? 1025_ktTts.txt 1025_0736_vgEngine.txt Share this post Link to post
SupportTeam Report post Posted 10/25/2019 02:10 PM Here is a quick patch + build that should revolve this: https://www.voiceguide.com/release/VoiceGuide_7.6.14_GHRJ_ORACLE.exe Please post traces if any issues still encountered. (bug was present if both the "Message that entered number was not correct or failed validation" was set to a .TXT file and the "Ask Caller to enter number:" mesage was set to .TXT file as well - so if TTS was used for both of those). Share this post Link to post