ebrady
New member
Papagayo Vixen Addin v1.3
I plan on adding some lip sync elements to my show this year for both Halloween and Christmas. I have played with Papagayo and it is a useful tool to work out the timing of the lip syncs, however it is somewhat tedious to import a completed Papagayo file into a Vixen sequence as it must be completed by hand.
Over the last week, I decided to write a plug-in for Vixen 2.1 and 2.5 to automate this process and am posting it. At this point I consider the software Beta, but figured I would go ahead and share it now to get feedback and to debug issues. Since it is Beta, it could crash or screw up your prized sequence, so make backups accordingly.
To use:
I have tested the following variations of Papagayo fps and Vixen Event Periods.
Papagayo 40fps, Vixen Period = 25ms ( Works good, but you can't do much at a time with the 500 frame limitation on papagayo)
Papagayo 24fps, Vixen Period = 25ms, ( These are the settings I used on the posted video ).
Pagagayo 24fps, Vixen Period = 50ms, ( Works, can't say I noticed much if any loss of resolution during the conversion process).
Papagayo 24fps, Vixen Period = 100ms ( Works, loss of resolution noticed on some syllables, depended upon the song ).
If you are unfamiliar with Papagayo, information and downloads can be found at http://www.lostmarble.com/papagayo/
A good discussion of lip sync information can also be found in the following thread: http://doityourselfchristmas.com/forums/showthread.php?12249-Singing-Pumpkin-Lip-Sync
The Papagayo Application (not the plugin), while it is a good tool definitely has a few bugs in it. If you are aware of what they are then it is not much of a problem, if you don't it will drive you literally CRAZY. Here are some of the things I found.... (These are bugs in the App itself, which I have no control of, not the plugin.)
1 - Papagayo does not like it when you load sound files of more then 500 frames. It does not warn you and will let you work as if everything is OK, when you go to save it will generate a corrupt pgo file and not warn you about it. This results in all of the work being wasted. I ended up downloading a copy of Audacity and broke the sequence into 15 or second less parts, worked on them individually and recombined them using the offset feature in the plugin on the final sequence.
2 - Papagayo does not like it when you put underscores in the pgo filenames. i.e. ToySack_1.mp3 will cause an error, once again, it will not warn you and act like it saved the file, but did not. I wasted about 45 minutes of work before I figured this out.
3 - Papagayo also has problems when you add phrases that have anything other then A-Z characters. If the phrase you are typing has anything other then that, then it will not save the file correctly.
The author of Papagayo is no longer working on the code, so these will likely not get fixed. However he has made the code open source so someone may one day.
a. To do the sequence, I did the following:
b. Create a vixen sequence mapped to the complete audio file.
c. Broke the mp3 file into workable segments and generated independent pgo files for them one by one
d. Mapped each of these pgo files one by one into the sequence made in "a".
I am open to suggestions for improvements to the plugin at this point, so all feedback is welcomed.
ChangeLog
v1.3
- Corrected issue were a word with no associated phoneme data would cause a plugin crash. This can happen when the app can not process one of the phrase words due to illegal characters or can determine proper pronunciation of the word.
v1.2
- Attempt to preserve voice phoneme->channel mappings when loading different pgo files.
The plugin will now save phoneme->channel mappings by voice name in each sequence. Once a mapping is defined in the sequence for a particular voice, the plugin will attempt to maintain this mapping when other pgo files are loaded that contain voices with the same name. This information is stored in the sequence and will be reloaded anytime that particular sequence is loaded.
- Made Filename text box readonly to prevent some undesired behavior.
- Removed extraneous whitespaces in all voice names.
- Added readme notes to zip file
I plan on adding some lip sync elements to my show this year for both Halloween and Christmas. I have played with Papagayo and it is a useful tool to work out the timing of the lip syncs, however it is somewhat tedious to import a completed Papagayo file into a Vixen sequence as it must be completed by hand.
Over the last week, I decided to write a plug-in for Vixen 2.1 and 2.5 to automate this process and am posting it. At this point I consider the software Beta, but figured I would go ahead and share it now to get feedback and to debug issues. Since it is Beta, it could crash or screw up your prized sequence, so make backups accordingly.
To use:
- Unzip the attached file and copy the dll to your Vixen Addin directory
- Create a Papagayo Sequence, work out the lip sync and save the file.
- Create a Vixen Sequence, it works best to make the frame rate match the fps rate in the Papagayo sequence, however it is not required.
- Open the Add-In "Papagayo Converter", from the Vixen AddIn Menu
- Select the "..." button to load the completed Papagayo file.
- Select the Voice you wish to map
- For each Phoneme in the left listbox select which Vixen Channels you wish for it to map to. It is possible to map each Phoneme to multiple channels.
- Select the offset in the sequence were you wish for the mapping to begin.
- Press convert
- The vixen sequence will show correctly once you close the Add-In box.
I have tested the following variations of Papagayo fps and Vixen Event Periods.
Papagayo 40fps, Vixen Period = 25ms ( Works good, but you can't do much at a time with the 500 frame limitation on papagayo)
Papagayo 24fps, Vixen Period = 25ms, ( These are the settings I used on the posted video ).
Pagagayo 24fps, Vixen Period = 50ms, ( Works, can't say I noticed much if any loss of resolution during the conversion process).
Papagayo 24fps, Vixen Period = 100ms ( Works, loss of resolution noticed on some syllables, depended upon the song ).
If you are unfamiliar with Papagayo, information and downloads can be found at http://www.lostmarble.com/papagayo/
A good discussion of lip sync information can also be found in the following thread: http://doityourselfchristmas.com/forums/showthread.php?12249-Singing-Pumpkin-Lip-Sync
The Papagayo Application (not the plugin), while it is a good tool definitely has a few bugs in it. If you are aware of what they are then it is not much of a problem, if you don't it will drive you literally CRAZY. Here are some of the things I found.... (These are bugs in the App itself, which I have no control of, not the plugin.)
1 - Papagayo does not like it when you load sound files of more then 500 frames. It does not warn you and will let you work as if everything is OK, when you go to save it will generate a corrupt pgo file and not warn you about it. This results in all of the work being wasted. I ended up downloading a copy of Audacity and broke the sequence into 15 or second less parts, worked on them individually and recombined them using the offset feature in the plugin on the final sequence.
2 - Papagayo does not like it when you put underscores in the pgo filenames. i.e. ToySack_1.mp3 will cause an error, once again, it will not warn you and act like it saved the file, but did not. I wasted about 45 minutes of work before I figured this out.
3 - Papagayo also has problems when you add phrases that have anything other then A-Z characters. If the phrase you are typing has anything other then that, then it will not save the file correctly.
The author of Papagayo is no longer working on the code, so these will likely not get fixed. However he has made the code open source so someone may one day.
a. To do the sequence, I did the following:
b. Create a vixen sequence mapped to the complete audio file.
c. Broke the mp3 file into workable segments and generated independent pgo files for them one by one
d. Mapped each of these pgo files one by one into the sequence made in "a".
I am open to suggestions for improvements to the plugin at this point, so all feedback is welcomed.
ChangeLog
v1.3
- Corrected issue were a word with no associated phoneme data would cause a plugin crash. This can happen when the app can not process one of the phrase words due to illegal characters or can determine proper pronunciation of the word.
v1.2
- Attempt to preserve voice phoneme->channel mappings when loading different pgo files.
The plugin will now save phoneme->channel mappings by voice name in each sequence. Once a mapping is defined in the sequence for a particular voice, the plugin will attempt to maintain this mapping when other pgo files are loaded that contain voices with the same name. This information is stored in the sequence and will be reloaded anytime that particular sequence is loaded.
- Made Filename text box readonly to prevent some undesired behavior.
- Removed extraneous whitespaces in all voice names.
- Added readme notes to zip file
Attachments
Last edited: