Welcome guest, is this your first visit? Click the "Create Account" button now to join.
Page 2 of 4 FirstFirst 1234 LastLast
Results 11 to 20 of 32

Thread: Audio Problem

  1. #11
    Super Member
    Join Date
    Dec 2010
    Location
    USA
    Posts
    152
    Rep Power
    44

    Default

    My voice:
    [Only registered and activated users can see links. ]

  2.    Advertissements


  3. #12
    Important User Audio Problem
    Audio ProblemAudio ProblemAudio ProblemAudio Problem
    osiris4isis's Avatar
    Join Date
    Apr 2012
    Location
    Darkside of the Moon
    Posts
    2,891
    Rep Power
    1364

    Default

    Ok, I was only able to check against 14th street, but this mod seems to work ok. And since it didn't need modding dictionary.voice, it should work for all numbers. Please check against 1st,2nd,3rd.... I also fixed the Ln->Lane instead of Ln->Line
    Spoiler: .

    [Only registered and activated users can see links. ]
    Attached Files Attached Files
    • This post requires you to click the "LIKE this post" button to see Attachments.

  4. #13
    Super Member
    Join Date
    Dec 2010
    Location
    USA
    Posts
    152
    Rep Power
    44

    Default

    I'M REALLY IMPRESSED!!! Your changes worked. I tried it on 1st St. and it said First Street. 2nd said second. How far should it go now? I ran another simulated trip in NYC and 50 said fifty. I'm kinda excited. Thanks. BTW, what did you change that I missed?

    Also, I hope you used my C320 unlock. Also C310.

  5. #14
    Important User Audio Problem
    Audio ProblemAudio ProblemAudio ProblemAudio Problem
    osiris4isis's Avatar
    Join Date
    Apr 2012
    Location
    Darkside of the Moon
    Posts
    2,891
    Rep Power
    1364

    Default

    Try at least one street in the hundred range (ie: 105th) The changes are what I had posted earlier. In config_transforms.lua, comment out L"^%d+", which allow correct logic to be used
    Code:
    Please Login or Register to see the links
    I never deleted any existing routines ( just commented them out between --[[ ... ]] ) The benefit of this kludge is that you don't have to manually add all entries (1st,2nd,3rd,4th...100th...) in dictionary.voice I hope I didn't break anything. For example, would it announce house address of "23 Applewood Dr" or "Route 66 Pkw" correctly? I also got a fix for "N Ave" not announcing "North Avenue"

    BTW, there are lots of TTSPro features (under diMkaWAaP133 skin) that your voice file didn't activates (ie: Read Out Route Summary) which make testing easier.

    Didn't have time to play with C310/C320 yet. Still got to look for a SD card (less than 4GB) to test with.
    Last edited by osiris4isis; 23rd December 2015 at 06:29 AM.

  6. #15
    Super Member
    Join Date
    Dec 2010
    Location
    USA
    Posts
    152
    Rep Power
    44

    Default

    Download those unlocks before they kill the links. In the event they do, I have back up. FYI, I ran another simulated NYC trip to 100 St. It said one hundred street. I can't think of anything above 100th St. LOL. So I made up an address on 126th Street and it said one hundred twenty-sixth st. Pat yourself on the back 'cause you deserve it.

  7. #16
    Important User Audio Problem
    Audio ProblemAudio ProblemAudio ProblemAudio Problem
    osiris4isis's Avatar
    Join Date
    Apr 2012
    Location
    Darkside of the Moon
    Posts
    2,891
    Rep Power
    1364

    Default

    Good, glad it worked. I can't take credit for it because the original code was already there, I just un-comment the unwanted part I think there are others that had tried to work on fixing various problems because I see the logics being chaotic. That is, there is no designed, just kludges to fix problems. The street naming in US is different from Europe/Asia so I guess people just fixed problems they found. I'll post the "N Ave" fixed and duplicate cleanup of config_transform_tables.lua later after some more testing.

    I've downloaded all of your Mio postings.

  8. #17
    Master
    Join Date
    Nov 2014
    Location
    USA
    Age
    51
    Posts
    172
    Rep Power
    68

    Default

    Good stuff! Nice working getting that fixed up!

    What's interesting (to me, anyway), is that the Nuance voice files in WinCE are completely different (they don't even have the "roadnumber_typ_patterns" function in the config_transforms.lua file (I wanted to check to see how mine were setup in WinCE). I would have thought that the voice .lua files would have been the same since I thought that LUA is a "language" used within iGo. If so, I wonder why the .lua files are so different between WinCE and Android - I would have thought that LUA was LUA - regardless of the OS. Obviously not though! :-)

    EDIT: I'm sorry - it *does* have the "roadnumber_typ_patters" function - I just missed it the first time. Interestingly, the first line is NOT commented out in WInCE. May be worth comparing the config_transforms.lua file between WinCE and Android - maybe there is another difference that causes it to work properly in CE (without commenting out that first line).

    EDIT2: Actually, *some* of the WinCE Nuance voice files have that function and some don't. It seems like the TTSPro Nuance voice files have the function and the non-TTSPro Nuance voice files do not. For example, the Samantha voice does not have that function at all. But the Ira voice file does have it (but that line is NOT commented out). Come to think of it, when I tested this issue on WinCE, I tested with Samantha - not Ira. Let me test with Ira to see what happens on WinCE.

    Sent from my HP SlateBook 10 x2 PC using Tapatalk
    Last edited by jtrosky; 23rd December 2015 at 12:28 PM.

  9. #18
    Important User Audio Problem
    Audio ProblemAudio ProblemAudio ProblemAudio Problem
    osiris4isis's Avatar
    Join Date
    Apr 2012
    Location
    Darkside of the Moon
    Posts
    2,891
    Rep Power
    1364

    Default

    I believe Nav N Go (maker of iGO/Primo/NextGen) uses a subset of LUA. My original mod was to change dictionary.voice file for an old voice running on MyWay. Then I checked Primo with the Joe Cool voice file (probably from Mike Hunt) had the LUA routines and after playing around with it, the "digit check" need to be remove in order for the routine not to intercept and change the output. The commenting out was my fix I like to use Samantha because it has TTSPro functionalities, so I'm porting the fix over to it. Who ever originally created the function format_road_number, was thinking to handle highway with naming of <digits> or <alpha><digits> or <alpha><alpha><digits>. US highways have format of <alpha><digits> like "I-75" (so having it not pronounced "Aye" 75 th is correct) or "US 101" (pronounced "U S one hundred one") However, I don't recall any highway having just <digits>, only streets. So commenting it out should be correct.
    Last edited by osiris4isis; 23rd December 2015 at 01:29 PM.

  10. #19
    Super Member
    Join Date
    Dec 2010
    Location
    USA
    Posts
    152
    Rep Power
    44

    Default

    Quote Originally Posted by osiris4isis View Post
    Good, glad it worked. I can't take credit for it because the original code was already there, I just un-comment the unwanted part I think there are others that had tried to work on fixing various problems because I see the logics being chaotic. That is, there is no designed, just kludges to fix problems. The street naming in US is different from Europe/Asia so I guess people just fixed problems they found. I'll post the "N Ave" fixed and duplicate cleanup of config_transform_tables.lua later after some more testing.

    I've downloaded all of your Mio postings.
    I'll be completely honest here. I found it amazing that you understood the computerese and knew what to comment out. I was kinda "lost."

  11. #20
    Master
    Join Date
    Nov 2014
    Location
    USA
    Age
    51
    Posts
    172
    Rep Power
    68

    Default

    So I tested Ira and Tom on WinCE and the whole 1st, 2nd, 3rd, etc thing works fine on WinCE, even with that one line un-commented. I think that there must be a difference in the TTS engine itself that is the root cause of the issue, not the voice file (although the change to the voice file is obviously a good work-around).

    I'm just wondering what else commenting that line might affect, if anything.

    Now we see what the folks at NNG have to deal with due to different platforms, TTS engines, etc. :-)

    Sent from my HP SlateBook 10 x2 PC using Tapatalk

 

 

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •