More Soundtrack BLUES

edited December 1969 in Carrara Discussion

I have my SOUND (VOICE OVER in Carrara and am trying to lip sinc it up! I can see the WAVEFORM pattern and am trying to Move the mouth to the correct syllables at this time. Problem is I can't really see/tell via looking at the time line WAVE FORM WHERE I need to mark a morph target instance??

I can scrub the time line back and forth but its almost impossible to hear it right.

I think it would be great to have a image of the WAVE FORM pattern say like in Photoshop and then PLACE the LETTERS and syllables in the correct spot above where the time line would reflect this!

Has anyone had any luck doing that OR maybe you have dome this sort of thing before and can give me some advise.

Comments

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,220
    edited December 1969

    bring a Poser dolly in that you have used talk designer on and match your own rigged figure up with its morphs along the timeline
    (I do belive you said once you actually have Poser?)
    or
    gasp!!
    download and instal Daz studio 4.5 (the 32bit version) and use lipsync on the grey Genesis figure and save a .duf scene.
    open it in Carrara 8.5 beta and the same thing I suggested.

  • edited December 1969

    Well I dont have POSER on my new machine and its poser 6 on my old power mac! There has got to be a better way to do lip sinking

  • edited December 1969

    You don't say which figure an original or daz generally I agree with Wendy the 32 bit lip sync feature works fairly well for me in DS 4.5 32bit also depending on what your doing Mimic Live has been quite useable to me I Have tried exactly what you are talking about scrubbing thru the timeline lining up morphs almost impossible for me if you can bring the figure into DS 4,5 32-bit use lip sync I think in daz you can /render/export it as a video clip and get an idea what the lip sync looks like. They have mimic for carrara I have used it in poser for me the one for carrara was not worth the money at the time. Some of it depends on the morphs for the particular model . Best of luck Happy and Safe Holidays to ALL !!!

  • edited December 1969

    UHM my figure(s) I make from scratch in Hexagon and send them to carrara where I do everything else. I use QUADS by the way and the Poly counts are very low and I us Smoothing and it works out great. I couldn't even imagine doing TARGET morphs with object others do from any source with gigantic ploy counts.

    I have been doing some thinking and research and think I came across a winner,

    Get a cheap say 100-150 buck wal-mart video cam corder and use it to shoot my ugkly FACE doing he voice overs THEN record the audio at the same time with my Professional Mic system so the VOICE over is of high quality and and exact match to whats going into the h]cheap cam corder!

    I have recorded a good close up video off of youtube and getting reasd to see it is a good answer of not Will post my findings

    Fulk.jpg
    282 x 463 - 57K
    Fulk_poly.jpg
    218 x 432 - 77K
  • edited December 1969

    good modeling job best of luck

  • edited December 2012

    Well I stayed up till 3 am yesterday doing a linc sink for 20 seconds of Jack Nickleson a few good men rant IT TOOK for damn ever and it is still not good enough.

    SOme of my Morphs are not accurate and the tweeners are not looking like the movement should.

    I did go and copy the section in the movie I wanted to lipsinc up and put it right next to my HEAD object and it works great You got the linc movement and emotions Eye brawl raises, Squinted eye and I t think I may be onto something

    BUT have come to realize to do ANY SCENCE where you want accurate LSing to do it 100% Manually it just cant be done . It would take weeks if one is talking about 10s of minutes or an hour of it, THis would have to be automated some how, ANd with handobjects I just dont know how to pull that off

    Then there is the issue of having to do dozens of Morph targets for dozens of individual charters say like 20 -30 separate MT. One would need a team of people.

    This last one is my fault as I didnt spend enough time getting the MT done correctly.

    One really needs to have a Collection of the VOWLS and certain other type of Speech.

    But I have no idea HOW this could ever be automated in Carrara or Mimic for a new MACRPO in any meaningful way.

    I would have to say if I plan to do more character Lip Syncing animation as a paid professional I would have to consider the limitations of scenes no longer that one typical 30 Second TV Commercial. I spent most of the day doing this 20 second spot and I only considered it half done . I was still going to have to do the movement of the bones rig!

    Post edited by richard.chaos_91798ec102 on
  • RoguePilotRoguePilot Posts: 239
    edited December 2012

    And you've just discovered why any high quality Hollywood style animation costs so much to produce.

    Manpower and time.

    It takes teams like you said, split up into modelling, rigging, texturing, animating, simulation, lighting, software engineering and finally rendering. It's all hand animated, 'rotoscoped' then given the animators particular flair too, no auto lipsynching there.

    The best behind the scenes stuff I've seen that gets the scale of production across is for The Incredibles.

    You're only one guy, I'd be proud of what you've accomplished just by yourself.


    The DAZ characters use 16 visemes so if you just match those in your model you may be able to use Minmic successfuly.


    (Just noticed why the Incredibles came to mind so readily, your model could easily fit into that style)

    Post edited by RoguePilot on
  • Box8068_31c338ee4bBox8068_31c338ee4b Posts: 292
    edited December 1969

    Hey
    My full time job is at a recording studio. Though music is our primary business we have done many voice pre records for Major animated films. The audio recording of the actor is Always video taped for the animators.
    8068

  • 0oseven0oseven Posts: 626
    edited December 1969

    The Mimic for Carrara is a step forward .

    It allows you to type in text which is converted to phonemes as clips on an NLA track.
    these can be moved and stretched or edited .

    you can make custom phoneme clips.

    Of course you dont need Mimic to do that but even after you made your own its very time consuming loading them into
    an NLA track to match the speech. Mimic does that automatically which I regard as a starting point for editing the synch.

    The text also appears below the wave form on the soundtrack and is a bit of help for timing.
    but Its difficult to see unless you collapse everything and there is still the problem of hearing the sound as you scrub.

    I have only just started to explore M for C and havent yet decided if its worth the hefty tag of $100
    but maybe worth you buying on trial .

    I never had much luck with Wendy's suggestion of working in Studio and bringing in duf files. Neither did I have luck with Daz Studio Mimic which saves in poser format but doesnt import properly into Carrara. Tried other things too like aniblocks via aniblock importer .Mimic for Carrara is best so far but not easy or perfect.

    One other thing I do with a lengthy sound file is to split it into short pieces say 4 to 8 seconds depending on the speech.Then (try) animate lipsynch each one . Render a few different cameras for each and put togethjer in Video software.

  • 3DAGE3DAGE Posts: 3,311
    edited December 1969

    Hi Rich :)

    I agree with 0oseven :)

    This is the "complex task" which Mimic for Carrara is designed to "simplify".

    If you're choosing to do it manually,.. the hardest way possible,.. then that's your choice. ...but mimic will analyse the sound file and create the lipsync animation in a few seconds.

    You would create a bunch of "NLA Poses" ,.. one , for each vocal shape, for your custom built figure.
    you assign these (poses) to the Visemes e.g: Ch, Aah, Mmm, P. etc....
    and then mimic will analyse the sound and create the lipsync animation for you.

    Hope it helps :)

  • 0oseven0oseven Posts: 626
    edited December 1969

    AND just to complicate matters - remember adding expressions such as smiles/grimaces alters the facial geometry probably distorting the phoneme pose you or mimic applied.

    In my previous post i said " The text also appears below the wave form on the soundtrack and is a bit of help for timing.
    but Its difficult to see unless you collapse everything "

    I since discovered - the soundtrack can be made sticky so it remains visible when you scroll down through the keyframes

  • edited December 1969

    007 and 3DAGE.

    Thanks, you know I bought MIMIC standalone like 10 years ago and never even opened it. Hell it's on like 3 computers back. Maybe Mac OS 9 cant recall.

    yea If I was gunna do Lip-syncs longer than a few seconds one would have to do something automated. I mean this is nuts

    YES all my objects I make from scratch so all morphs I have to do manually. Getting the correct ones is difficult! Also seems the TWEENS are an issue. Seems the mouth is open too long or not long enough.

    The best setting for the TWEENS set for say the MOUTH MOVEMENT TWEENS set to no less than 25% IN and 75%, BUT am thinking 49 % in 51% Out! Whats your pinion there

    OH by the way I have been doing this in the timeline NOT NLA! I may give that a whirl.

    Also I need to learn to set up the correct what is it calledPHONEME? Mouth Morphs set to certain sounds.

    IS there any resource.

    Ooseven I have in the past TRYED to do a screen shot of the WAVE FORM and in Photoshop PUT the text where for the best guess where that syllable letter sound would be. BUT I think I have figured out and even better.

    Like I said I am doing a SINCE of part of jack Nichelson RANT when he is on the stand in a FEW GOOD MEN. For one I could never get tired of listen to it THATS GOOD since I have listen to it 1000 time yesterday AND for the most part the video is all of him close up. The part I am doing doensn't have tom cruz cut in for a few seconds so thats good.

    I copied it and put in on as a shader of a one POLY objet and play it right next to the head object I am morphing I was HOPING that I could also use the SOUNDTRACK INSIDE CARRARA as well but Its will only play about 20% of the time and when it does it is nearly impossible to understand and is NOT sinced up at all. The SOUNDTRACK in CARRARA for mac is a flop and and completely useless.

    BUT when you have the ACTUAL VISuals of the MOUHTH movement I guess the actual sound is not needed! Like I said I have 20 seconds of lip-syncs here to day and I spent over a day on that 2o seconds and still that was just the mouth movements let alone eyes teeth tune head ect ect ect!

    I am getting ready to redo ALL the MOUTH morphs targets and start again. I will post what I end up with in a few days.

  • edited December 1969

    ALSO looks like MIMIC works with DAZ not Carrara and for MAC OS 10.7 or lower BUT I could use it on my power mac G-5

  • edited December 1969

    ALSO as far as doing accurate LS I was thinking if I do voice overs myself I could take a close up video of myself doing the voice over then use that video in the manner I am using Jack Nicklesons Rant from a few good men now!

    I kano I know thats gunna be some ugly subject natter for sure

  • 0oseven0oseven Posts: 626
    edited December 2012

    If it were me I think I'd do the following.

    Put the jack nicholson video in my editor and slice it up into suitable chunks - how big and how many is up to you - I dont think you mentioned how long the animation is going to be.

    my own project ( which is unlikely to see the light of day at the rate Im going) is a 4 minute song split into 19 audio tracks of varying lengths between 4 and 15 secs each.

    Save the "chunks" first as video and then audio so you have a complete set that match.
    Rotoscope the video as your visual guide as you described in your post ( I never thought of doing that :wow:)and load the corresponding audio track to the timeline.

    With a bit of luck they should synchcronise whilst you do the lipsync animation.

    After you animate, render with or without sound - when all tracks are done recompile the avi/mov files in you video editor.

    If you didnt render sound with the movie reload the audio tracks and they should match, although I would save a copy of the entire audio track to start with and load that for final video.

    Its a chore making phoneme NLA clips but I think you 'll find you dont need too many. Mimic provides a load but its hard to distinguish between some of them. ONce you've got them they can be moved around on the nla track which helps with sync also as i said before they can be stretched and blended or edited in keyframes if you want the nitty gritty adjusments .

    The old Mimic for Daz has been left behind - no plans to update it I'm told - I found this out because i was trying to use Mimic for Daz with studio 4.5 and genesis - 64bit is another issue.Hence my transition to Mimic for carrara. I really dont have time or the inclanation for all the work arounds and flipping back and forth between applications which in itself usually presents problems.>:-(

    hope there is some thing useful for you in the above
    cheers

    Post edited by 0oseven on
  • 3DAGE3DAGE Posts: 3,311
    edited December 1969

    HI Rich

    Mimic for Carrara,... is different from the stand alone Mimic pro (which works with Daz figures)

    The Carrara plugin, allows you to create "NLA poses" (for the Phonemes / visemes) and then assign those in the Mimic controls (see pic)

    so you can mix some Morphs sliders values in your custom figure,. then "snapshot" an NLA Pose, of that current shape, then assign that to a Phoneme in mimic.

    Then when you load your sound file,. and hit go, it uses those NLA poses to create the lip-sync.

    it takes all the hassle and pain out of trying to hand key-frame a complex sequence, and although it's an "Automated" process and it will have some points which you'll need to adjust things,... It does it's job in a couple of seconds. and you have the ability to scrap it , adjust or change some of the NLA poses and try it again.

    Click for full image.

    Mimic_for_Carrara.jpg
    1280 x 1024 - 280K
  • edited December 1969

    Next Question! IS the MIMIC Plug in for Carrara 8.5 PRO? On MAC OS 10.8 Mountain lion?

    Doubtfull

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,220
    edited December 1969

    I beieve so, I got seperate mimic installers for my beta builds for Whine Doze so do not see why Mac should be different, a Mac user like Holly or Cripeman may confirm this for you.
    Anyway does not hurt to wishlist it if not on sale, I only paid $10 for it on sale so you might get lucky.

  • 3DAGE3DAGE Posts: 3,311
    edited December 1969

    From the Beta Thread ....

    About Mimic Pro and this build:
    If you own Mimic Pro and want to start using it with Genesis, you’ll need to download a build just for Carrara 8.5.0.172. You will NOT want to use this with your 8.1 build of Carrara. It won’t work with both builds.
    However, if you install the new version in a different location, you can continue to have both versions on your machine. The same serial number will continue to work for both builds. The DMC files are also included in this new version of Mimic Pro and will work with Genesis.


    So,. Yes,. the mimic stuff has been updated to work with the beta, and it should run on the mac,
    I'm on PC,. so I can't say whether it runs on mountain lion or not, but the updated mimic stuff works for me with Daz3D figures, and my own figures.

    :)

Sign In or Register to comment.