FOR EVERYBODY THAT CAN'T GET LIP SYNC ON A HWM MODEL!!!! Right click on the animation set, got to show in element viewer, then click on AnimationSet. Then expand presetGroups. You will see 'Phonemes' is not there. Right click on presetGroups, then select Import Element... Go to the file path for the model (should be: game/tf/models/player/hwm) then tell it to look for all files. Select 'class_phonemes.pre' (instead of class there would be Scout, Soldier etc.).
Dude your awesome! Although I've already started getting confused a couple parts ago, this is going to help soooooo much!! I love how you used your own voice or the dialogue bit, you rock !
it would be much more work to make software know the difference between languages, so no, it isn't. (think about how language speaking AI is really bad still, even in 2021) it works by listening to the individual sounds it hears and moving the mouth exactly where it should be to produce that sound. There is only one way the mouth can produce an S or Th or EE sound. There's only one way you can make an MM or VUH/ F sound. No matter what language you speak, we all use the same mouth and sounds. Lipsync ai works for all languages, however it will always require a bit of additional work to get the plosive and tongue movements just right, especially if it's a quickly spoken phrase.
INFO: This Lipsync-Guidevideo only works with TF2-HighPoly-Charakters...to make this Work on HL2-Characters like Alyx or Barney, select the model in the "Animation Set Editor", make a rightclick on the Modelname and search for the function "Show in Element Viewer", then select the Modelname again in the "Element Viewer"-section with a rightclick and activate "Create phoneme Presets based on HL2 FACS Controls". This will make "Phoemes extraction" work with Hl2 Models!!!
it appears im not the only one that got confused lol if you are having problems with lip-sync on a placed character model, it's because you need a HWM model instead of a regular one: in the model menu select the tf_movie folder or search HWM, and the 9 mercs should appear. HWM models are higher quality and have the mouth controls necessary for phoneme extraction. (I'd recommend using this from now on, these where the models used for the meet the team videos, and are much better than game player models)
You must get the model under the hwm folder, so instead of player/engineer.mdl select the player/hwm/engineer.mdl These models are in alot more detail than the standard ones.
Trust me its not all characters. When you're selecting a model from the Animation Set Editor you need to make sure the model is from HWM. To be specific change your mod filter from "All Mods" to Tf_Movies. And the scout for example his file folder would be player/hwm/scout ! =] I had the same problem as you and this is how i learned
When lipsyncing, be sure that the tab "Phonemes" shows up next to the "emotion" tab. If it doesn't show up, restart your SFM and try again. It works for me.
Let me see if i can explain. Click the big plus symbol in the top left hand corner of the animation set editor. Click Create Animation Set(s) For Existing Elements. Click the camera of which is being used in the scene. scroll or type in scout and look for his character. Once you click on him and press ok he will appear in the Animation Set Editor and you can edit his motion, lipsync, and everything else about that character
Making mouth animations is more detailed and complex than I thought. But I didn't know automatic mouth animation was possible, this is a great feature.
coming in here to say that with eyes, generally when trying to display emotion, it's better to turn off the target view factor entirely and animate eye movement from scratch, and when the eyes look at something in particular to use the viewtarget factor
For no reason, the lipsync stops in the middle of the dialogue (6 sec custom scout sound) I simply did the extract phonemes thing and the entire audio is within one scene and scout just stops talking with his mouth open and the audio still playing. Ok so the model was a recorded model and it looks like he stops when the recording originally stopped creating the still frame.
Some people have missing phoneme tabs with certain HWM models. Even if you add another model with the proper phonemes so the tab appears, you can not modify the phonemes for the original model.
For everyone who don't have lower face and tongue, and if lypsinc doesn't work You're probably using wrong model, search on youtube "source filmmaker tip of the day replacing models" and watch the first video It must help
The more this guy teaches me about SFM, the higher chance I might be able to win in the Saxxy Award's Drama Catagory, that is, if I find the right voice actors.
The camera may be keyframed to move ideally in front of the scout or then the camera's animation set has been locked to with the scout's animation set so it moves correctly in front of the scout.
Is there a way to zero things out? like in maya when you can grab the name on the slider you are adjusting and enter 0 and it will return to it's neutral state? So I can build a mouth shape with many different sliders, and when I move on to the make the next sound and I want to start from scratch with all the sliders in their neutral positions, I can just hit a hot key or something and everything returns to neutral? Thanks for the tutorials, they are a huge help :)
Jesus, it took me forever to get the lip sync working. I had to copy the motion from my original recording of the scout in the Motion Editor, then delete that scout, bring in a new HWM Scout model, paste the motion back in, re-rig the camera to him again, and THEN it works. In case some of you want to know how to get HWM models working with recordings.
im pretty sure when it becomes easy and "fluent" it would take much less time. I mean, the guy's trying to teach us, he isn't gonna go at his normal pace :P
All of the people for who lipsync doesnt work : You should open the hwm model instead of the standart model when you create animation set for new model, worked for me ! The hwm model is in the "tf_movie" mod while the standart model is stored in the "tf" mod
It won't work for certain models. There are 2 models for each character. For example, medic. The working one: player\hmw\medic.mdl The one that doesn't: player\medic.mdl
If you're using your own wav file, you may have to manually tell SFM what the phonemes are, I doubt it'll take them directly from a custom sound file. This is just a guess though, not tried it myself
ctrl+z is undo (only do this if the distortion was ur last action)... or right click on the audio file, then edit clip, then set timescale, and change the timescale to 1.0
Yeah, it can be easy if you aren't completely savvy with the program. Many times you have to have a pretty steady hand to get things where you want and a bit of technical know-how will help in not making mistakes.
OH I GET IT, it doesnt work on recorded player movements/models? eg the original running scounts + soldier? cause i wanted my soldier to say "if ya know whats best for ya, you best RUN" but theres nothing selectable in his scene
right click anywhere in the animation set window and select show hidden controls........then select the face section and right click, now make sure there is a check beside selectable
Nice....now if only SFM would work for me. :( Whenever I try to play it now, it just says: "Not Responding", and never starts working. Even 5 hours later, it still does that.
I found the solution. Check with his "sound" tutorial. What I did wrong ( and I assume you too) was that I zoomed into the actual video clip, and put the sound there. Now, go check out the sound tutorial (#3 or 4), and see that he opens a separate worktable for sounds. Put dialogue in the "Dialogue" bit, show the waveform, and it should work! If you still cant get it to work, i reccomend re-doing the scene, following the sound totorial first, that's how I got it working. Hope it helps.
Wow alot of people are skipping video because they "dont need" to watch them I suggest watching all of them then reveiw them and even save them for future referance like I am doing.
First Click Scout's face and then right click, click "extract Phonemes" And Click extract and Press Y and say SomeThings and Press Y again and review it.
"G'job, buddeh."
-- RAITT, Bay
goddamnit volvo, you forgot to say in the end "Also, save before exiting"
6:30
Watch as a Blu Scout mystically turns Red, then back to Blu...
Demoman : *Bloodeh Spy*
That part with turning your voice and words into the Scout's was awesome. Didn't think it'd be that easy!
FOR EVERYBODY THAT CAN'T GET LIP SYNC ON A HWM MODEL!!!!
Right click on the animation set, got to show in element viewer, then click on AnimationSet. Then expand presetGroups. You will see 'Phonemes' is not there. Right click on presetGroups, then select Import Element... Go to the file path for the model (should be: game/tf/models/player/hwm) then tell it to look for all files. Select 'class_phonemes.pre' (instead of class there would be Scout, Soldier etc.).
thank you so much
Still don't work
GIVE THIS MAN MORE LIKES
@@purplebanana1433 I forgot I ever even used SFM tbh
I love how you dont have to do a word for word for frame animation for the mouth, the program already does it :D
Although not always. Sometimes it isn't smooth lipsync.
Ok thanks but everybody ate you going to the bathroom
@@Movie-xs4up what
@@Movie-xs4up what
@@Movie-xs4up Ah, you play Gacha. That explains everything...
Dude your awesome! Although I've already started getting confused a couple parts ago, this is going to help soooooo much!! I love how you used your own voice or the dialogue bit, you rock !
Hes a valve employee, these are official valve tutorials.
@@cruelyplayz a bit too late....
he is source film maker
@@TerrificToilet yeah just a tiny bit 🤏
Did someone else noticed that the scout running in front of the pyro is red?
yus
OreMaster Animations This something they need to fix
Right-click on the scout in the animation set editor, then go to set skin and choose whichever one is the blue scout outfit.
i noticed that and found this thing called skin and i set it to skin 1 idk where it is tho
He forgot to set skin to blue on that scout
I guess lip sync in Source Filmmaker is only for english dialouge?
it would be much more work to make software know the difference between languages, so no, it isn't. (think about how language speaking AI is really bad still, even in 2021)
it works by listening to the individual sounds it hears and moving the mouth exactly where it should be to produce that sound. There is only one way the mouth can produce an S or Th or EE sound. There's only one way you can make an MM or VUH/ F sound. No matter what language you speak, we all use the same mouth and sounds. Lipsync ai works for all languages, however it will always require a bit of additional work to get the plosive and tongue movements just right, especially if it's a quickly spoken phrase.
INFO: This Lipsync-Guidevideo only works with TF2-HighPoly-Charakters...to make this Work on HL2-Characters like Alyx or Barney, select the model in the "Animation Set Editor", make a rightclick on the Modelname and search for the function "Show in Element Viewer", then select the Modelname again in the "Element Viewer"-section with a rightclick and activate "Create phoneme Presets based on HL2 FACS Controls". This will make "Phoemes extraction" work with Hl2 Models!!!
My god, thank you. Thank god I was able to import my animations onto the HWM model. I owe you!
Auto lip sync! Yes! Finally! After all these years! This is incredible. The amount of good quality sfm videos will be overwhelming!
I know im late but didnt sfm release in 2012 the same time when this tutorial video got released, so it wouldn't be years it would be months lol
My God SFM is amazing.
These tutorials are still very helpful 11 years later
nah
12 years later and they still
it appears im not the only one that got confused lol
if you are having problems with lip-sync on a placed character model, it's because you need a HWM model instead of a regular one:
in the model menu select the tf_movie folder or search HWM, and the 9 mercs should appear.
HWM models are higher quality and have the mouth controls necessary for phoneme extraction. (I'd recommend using this from now on, these where the models used for the meet the team videos, and are much better than game player models)
Instructions unclear, broke my own jaw...
You must get the model under the hwm folder, so instead of player/engineer.mdl select the player/hwm/engineer.mdl These models are in alot more detail than the standard ones.
i can hear a sound but my charachters lip also closed
same
"GOOD JOB BUDDY" Alright dude.
Trust me its not all characters. When you're selecting a model from the Animation Set Editor you need to make sure the model is from HWM. To be specific change your mod filter from "All Mods" to Tf_Movies. And the scout for example his file folder would be player/hwm/scout ! =] I had the same problem as you and this is how i learned
oh, so that might be why my demo isnt doing anything with his mouth
When lipsyncing, be sure that the tab "Phonemes" shows up next to the "emotion" tab. If it doesn't show up, restart your SFM and try again. It works for me.
This is loads of fun animating TF2 characters, the fact they taught animating your own models in uni was what put me off, because my models suck :p
wat! SFM has autolip-syncing :O GOD FREAKIN AWESOME!
Let me see if i can explain. Click the big plus symbol in the top left hand corner of the animation set editor. Click Create Animation Set(s) For Existing Elements. Click the camera of which is being used in the scene. scroll or type in scout and look for his character. Once you click on him and press ok he will appear in the Animation Set Editor and you can edit his motion, lipsync, and everything else about that character
need new version tutorial
likes to VALVE try to see me
Making mouth animations is more detailed and complex than I thought.
But I didn't know automatic mouth animation was possible, this is a great feature.
I extracted the line, but my spy isn't talking.
cant find any hwm characters
Mr Lenny is it. An wav? Sync only works if you use wav files.
same
It will work if you selected the desired audio, and it must be a WAV.
Did you make the spy happy, mad or frown? Cut if you did he won't lip sync
This is such a good idea for Valve to release this software to us. Can't wait till the L4D2 resources are released to us!
coming in here to say that with eyes, generally when trying to display emotion, it's better to turn off the target view factor entirely and animate eye movement from scratch, and when the eyes look at something in particular to use the viewtarget factor
If you're referring the lipsync, it's not as complex as it sounds - usually you can just use auto-lipsync and a bit of smoothing to make it look good.
thx brodda :) you just saved me a lot of time.
LA Noir: We require 15 HD cameras to make our facial animations badass!
Valve: What cameras?
1:19 Made me laugh so hard "good job buddy!"
You have no idea how hard it was not to skip to this part XD
You have no idea how this is super broken.
For no reason, the lipsync stops in the middle of the dialogue (6 sec custom scout sound)
I simply did the extract phonemes thing and the entire audio is within one scene and scout just stops talking with his mouth open and the audio still playing.
Ok so the model was a recorded model and it looks like he stops when the recording originally stopped creating the still frame.
Valve has created something revolutionary - again.
The voice he did for the scout kinda sounds like his voice actor, but not imatating the scout...
I'm glad I self taught myself how to manually lip sync and edit the face to my likings
That's because LA Noir uses actors. Valve uses science! Thank you to Dr. Paul Ekman and FACS!
Meanwhile, in Valve, a couple dozen employees are wondering how the hell ponies are suddenly more important than the video.
If I may still say it more straight, the camera's animation set is "bonemerged" when animation set is locked with another animation sets parts.
Some people have missing phoneme tabs with certain HWM models. Even if you add another model with the proper phonemes so the tab appears, you can not modify the phonemes for the original model.
For everyone who don't have lower face and tongue, and if lypsinc doesn't work
You're probably using wrong model, search on youtube "source filmmaker tip of the day replacing models" and watch the first video
It must help
6:29 blue scouts
6:31 red scout o.O
6:32 blue scouts
good job buddy :P
Later on, when I got pretty good with SFM. I realized that not all models work and only certain Valve ones do.
The more this guy teaches me about SFM, the higher chance I might be able to win in the Saxxy Award's Drama Catagory, that is, if I find the right voice actors.
Damn i didn't know there are auto lipsync
this is the feature I'm most excited about.
Thanks works for me, also please update sfm so it runs a newer version of tf2 (currently it is a 2008 version) and fix the many bugs
"Good job, buddy!"
Even if it does have to do with ponies, remember that the pony's wife is cheating on him and he is drunk all the time. But don't tell that to him
The camera may be keyframed to move ideally in front of the scout or then the camera's animation set has been locked to with the scout's animation set so it moves correctly in front of the scout.
It used to take 3d animators days to make their character's mouths move, now its like...pow
him after you finish this tutorial vid: 1:20
Yepp, not working for me either. Does it require any of that BS that Faceposer needs? I had massive problems with that as well.
Actually the problem was in Meet the Heavy, only one audio track even works period.
I was getting ready to spend 30 months lip syncing for "RUN COWARDS!", but it's just does it automatically? Good god :U
Is there a way to zero things out? like in maya when you can grab the name on the slider you are adjusting and enter 0 and it will return to it's neutral state? So I can build a mouth shape with many different sliders, and when I move on to the make the next sound and I want to start from scratch with all the sliders in their neutral positions, I can just hit a hot key or something and everything returns to neutral? Thanks for the tutorials, they are a huge help :)
Yup, you right, i have this problem too now :(
i wish he told us that the lip sync doesn't work for the placed models like the pyro. took me like 20 minutes to realize
wait pyro has a mouth?
This is insane automatic lip sync i love it
oh my god thats so damn awesome
Thanks man. Didn't realize I was using the wrong model. Now I have some help with the lip sync!
Good job buddy!
its like character customization in a bethesda game
I can't thank you enough. Thank you sir!
Jesus, it took me forever to get the lip sync working. I had to copy the motion from my original recording of the scout in the Motion Editor, then delete that scout, bring in a new HWM Scout model, paste the motion back in, re-rig the camera to him again, and THEN it works. In case some of you want to know how to get HWM models working with recordings.
im pretty sure when it becomes easy and "fluent" it would take much less time. I mean, the guy's trying to teach us, he isn't gonna go at his normal pace :P
How strange. :o
biggest problem I have with SFM is that it crashes rather often.
All of the people for who lipsync doesnt work : You should open the hwm model instead of the standart model when you create animation set for new model, worked for me !
The hwm model is in the "tf_movie" mod while the standart model is stored in the "tf" mod
Wait, when he recorded his voice, the SFM AUTOMATICALLY synced the Scout's mouth with the recording? YUS!
its best to make sure SFM glitches and freezes your system before you save, that way you know its telling you you did it right.
One problem I have is the fact that for some reason the Medic and Spy have no lower face category..
Wow. I´m glad it still works for me
It won't work for certain models. There are 2 models for each character. For example, medic.
The working one: player\hmw\medic.mdl
The one that doesn't: player\medic.mdl
If you're using your own wav file, you may have to manually tell SFM what the phonemes are, I doubt it'll take them directly from a custom sound file. This is just a guess though, not tried it myself
Is the recorded sound programmed to blend in with the class voice?
ctrl+z is undo (only do this if the distortion was ur last action)... or right click on the audio file, then edit clip, then set timescale, and change the timescale to 1.0
if anyone is saying "MY PRO WON'T MOVE HIS MOUTH" HE HAS A GAS MASK NOT A MOUTH
Yeah, it can be easy if you aren't completely savvy with the program. Many times you have to have a pretty steady hand to get things where you want and a bit of technical know-how will help in not making mistakes.
Wow. That is just... Amazing!
"Good job buddy!" :D
Is it any good? I.e. worth checking it out to get familiar or did they change a lot of stuff?
OH I GET IT, it doesnt work on recorded player movements/models? eg the original running scounts + soldier? cause i wanted my soldier to say "if ya know whats best for ya, you best RUN" but theres nothing selectable in his scene
Aye. Managed to make a SFM video now. :)
The engineer doesn't have the face!
right click anywhere in the animation set window and select show hidden controls........then select the face section and right click, now make sure there is a check beside selectable
I'm so excited about this whole thing! Watch the portal 2 boots trailer to get even more excited about this!
It works with the TF2 characters but not with the ones available in the Steam Workshop
Nice....now if only SFM would work for me. :( Whenever I try to play it now, it just says: "Not Responding", and never starts working. Even 5 hours later, it still does that.
Last tutorial-scout is red
This tutorial-scout is blue
I think he left something out...
Automated lipsync? Good god, that's badass
YAY I know how to play the piano, it will be so much easier
I just noticed, the Scout model is actually HWM Model, where can I find one? My SFM doesn't show any of HWM Models.
SFM May have not loaded them yet,Just keep the model browser open then you should see it (player\hwm\scout for getting a scout)
Well, technically it was an inflatable balloonicorn. But your point is still valid.
I found the solution. Check with his "sound" tutorial. What I did wrong ( and I assume you too) was that I zoomed into the actual video clip, and put the sound there. Now, go check out the sound tutorial (#3 or 4), and see that he opens a separate worktable for sounds. Put dialogue in the "Dialogue" bit, show the waveform, and it should work! If you still cant get it to work, i reccomend re-doing the scene, following the sound totorial first, that's how I got it working. Hope it helps.
Might wanna watch the Graph Editor tutorial, it's the same for props and models.
Wow alot of people are skipping video because they "dont need" to watch them I suggest watching all of them then reveiw them and even save them for future referance like I am doing.
First Click Scout's face and then right click, click "extract Phonemes" And Click extract and Press Y and say SomeThings and Press Y again and review it.
He sounds so much like Scout.