Messages in this topic - RSS

Home ? Feedback ? Various issues

pages: 1 2 | the topic is closed
27/10/2013 19:49:42

primaveranz
primaveranz
Posts: 524
Why not add a "Lip synch" "Prepare" stage where you can type in the words you want the character to say rather than rely on the built-in technology which is approximate at best? E.g. Anyone managed to get a character to whistle yet?
permalink
27/10/2013 21:52:24

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
Do you mean text to speech?
permalink
27/10/2013 22:32:59

primaveranz
primaveranz
Posts: 524
MrDrWho13 wrote:
Do you mean text to speech?

Definitely not, TTS is unremittingly awful!

I mean something like Moviestorm's ability where you can improve the actual facial movements during speech by typing in the words to help the engine identify the sounds it is trying to mimic. E.g. I often have to use a foreign language which the engine doesn't recognise, but by typing in the phonetic equivalent the mouth shape is improved. Perhaps though, Muvizu's facial movements are limited to just opening and closing the mouth? In which case my suggestion is pointless.
permalink
28/10/2013 09:48:36

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
That sounds like a good idea!
permalink
28/10/2013 13:44:01

Mike_Num_5
Mike_Num_5
Posts: 165
Muvizu uses certain phonemes to try and distinguish between words when you import sound clips. It's something we haven't reviewed in some time right enough so might be worth us having a look of the system we use.

Thanks for the input guys!
permalink
28/10/2013 14:29:59

fazz68
fazz68
(Account inactive)
Posts: 767
what !! typing in an entire script for mouth movements..... fine if your video is a few mins long and your characters dont say much. typing in an avengers soundtrack phonetically would be a nightmare . im very happy with the way the mouth movements work at the moment. Ok it might miss the odd word now and then but i can live with that.
permalink
28/10/2013 15:15:21

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
Perhaps the option to either have automatic lip sync or choose the mouth shape?
permalink
28/10/2013 17:03:16

urbanlamb
urbanlamb
(Account inactive)
Posts: 1796
MrDrWho13 wrote:
Perhaps the option to either have automatic lip sync or choose the mouth shape?



the best one I have seen is the system that does both. (iclone again) it does it automatically but generates the mouth shapes in a timeline below the sound clip. You then can go in into that line and clean it up to make it more accurate. I dont want a fully manual process this makes little sense the process should be automated to some degree regardless. Muvizu works on the same principle as its using the soundfile to choose something close to the shape it should be but it makes mistakes being a piece of software it isn't always right in deciding these things soo being able to edit this layer would be beneficial unfortunately I dont think this is "muvizu like" this is iclone like and I think other things are more needed then the odd missed sound in the track and want them to spend time on that and leave that on a todo list for when everything else is done for example improving the prop system they just introduced which from the feedback thread looks like its not actually working that well yet. Perhaps in the future they can add a track below the soundfile to reveal the information muvizu is already detecting and allowing us to tweek it, but this seems like something that would go into the pro paid version not the free version.
permalink
28/10/2013 18:57:57

primaveranz
primaveranz
Posts: 524
fazz68 wrote:
typing in an avengers soundtrack phonetically would be a nightmare .



It's really not a big deal to cut and paste if you have already written it in something like Celtx or Word. . And you only need the Phonetic bit for 'unusual' languages depending on the software.
edited by primaveranz on 28/10/2013
permalink
28/10/2013 19:15:53

primaveranz
primaveranz
Posts: 524
MrDrWho13 wrote:
Perhaps the option to either have automatic lip sync or choose the mouth shape?


Yeah, I'd prefer that.
permalink
28/10/2013 20:06:06

fazz68
fazz68
(Account inactive)
Posts: 767
primaveranz wrote:
fazz68 wrote:
typing in an avengers soundtrack phonetically would be a nightmare .



It's really not a big deal to cut and paste if you have already written it in something like Celtx or Word. . And you only need the Phonetic bit for 'unusual' languages depending on the software.
edited by primaveranz on 28/10/2013


the avengers is a pre recorded radio show from the 70's i have been turning into a serial so there is no script to cut and paste just a gazzion spoken words on an audio file and turning it into a readable script would take me forever to do, so the current muvizu way is perfect for what i need.
permalink
28/10/2013 20:14:44

primaveranz
primaveranz
Posts: 524
fazz68 wrote:


the avengers is a pre recorded radio show from the 70's i have been turning into a serial so there is no script to cut and paste just a gazzion spoken words on an audio file and turning it into a readable script would take me forever to do, so the current muvizu way is perfect for what i need.


I feel your pain. I just spent many months producing an animated version of a 30 minute radio show, and it was in Bislama not English. So the optional implementation that MrDrWho13 suggests is the one I would support.
permalink
28/10/2013 20:22:59

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
I might make some mock up interface art of this system.
permalink
28/10/2013 20:51:08

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
Part 1:
permalink
28/10/2013 20:57:33

urbanlamb
urbanlamb
(Account inactive)
Posts: 1796
MrDrWho13 wrote:
Part 1:



i dont think that manual is appropriate in any sense and i dont know why muvizu should create a second system on top of their existing system of recognition.

all they need to do is make the internal system already there visible and we can then edit it this means that people are still limited into completely automated or completely manual. All they need to do is add a toggle to turn on the data that already exists and allow us to change it. We either hide it or use it

Even if they added a TTS system there is not need for a truly manual operation let muvizu do its guessing and we simply edit that work that is how most softwares work why would we remove this automated ability? just turn it on in the timeline below the wavefile and allow us to edit that information by clicking on it and changing the mouth shape or dragging the existing shape to where we need it
edited by urbanlamb on 28/10/2013
permalink
28/10/2013 21:20:57

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
Ok, I'll do another bit to show that!
permalink
28/10/2013 21:42:08

urbanlamb
urbanlamb
(Account inactive)
Posts: 1796
It would just be a toggle in the timeline another line of information for muvizu to reveal. I dont even know if a picture is necessary it would simply be below the wavefile. Muvizu has created its own internal animations for the various mouth movements and generates a set of movements based on what it hears. It just needs to be graphically represented in the timeline is all below the wavefile as its already there (just like the wavefile info was already there) so its just a bunch of vowels and text that is dragged around. This goes back to cartoons the old fashioned way its simply been automated this is very shall we say "old school" stuff that was created by mr walt disney all those ages ago

anyhow we just need a way to change the movments that people use to this day in animation. So we would drag all those vowels to where we need em basically since this is an issue of misplaced vowels and consonants. That way people can mimic things which are hard to recognize in the system without muvizu making an entire animation around it. Whatever is done it should not force us into a fast and hard decision of doing all the work manually one movement at a time or doing it all automatically that is very rigid and unforgiving.

permalink
28/10/2013 22:04:07

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
Part 2:
permalink
28/10/2013 22:06:44

MrDrWho13Muvizu mogulExperimental user
MrDrWho13
Posts: 2221
Is that what you wanted?
You could also use these for basic facial animation (hint hint Dreeko)
edited by MrDrWho13 on 28/10/2013
permalink
28/10/2013 22:26:46

urbanlamb
urbanlamb
(Account inactive)
Posts: 1796
yes sort of but without having to go through recording those movements its not practical people just need to be able to move them around and alter what is already there after its generated
permalink
pages: 1 2 | the topic is closed

Home ? Feedback ? Various issues