Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.Daz 3D is part of
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Okay, but when I edit the speed/volume, how can I hear what the changes sound like?
Unfortunately, there is currently no possibility to do that withoud creating a new sound file, but we will try to implement that feature in future releases.
Okay thanks for clarifying that. I thought I was missing something.
maybe this 64bit binary on github will work
https://github.com/rhdunn/espeak
https://github.com/espeak-ng/espeak-ng/releases/download/1.49.2/espeak-ng-x64.msi
I use Windows 7 Pro, the language is German.
Oh I am very sorry! Excuse me please.
It was only the original shipped Base-Genesis, so Dobit can see that I made no mistake with the Figure (V4 or so).
I would be very proud if Dobit can help me, because this Plugin will be very nice and useful for my animations.
Normally I am using Mimic Live, but I think AniLip is very helpful, too.
The MS speech libraries are known to cause problems when using English on non-English Windows.
"For instance, some examples the words may display incorrectly if the engine language is not the same as the system language. In such cases, this is the expected result and it is not an error of the speech system"
source: https://msdn.microsoft.com/en-us/library/ee125102(v=vs.85).aspx
And while it's a useful plug-in (at least in theory - I haven't actually tried it) what with being able to convert text to an animation without requiring the audio file, the product page must be updated with relevant info regarding speech libraries so potential buyers can decide if it's worth the price for them, otherwise it looks a lot like a bait and switch routine of "buy my product and only later find out you have a limited library unless you spend more money on restrictive content". As well, the statement "Generate visemes from the sound file and write them to the timeline" is misleading since you stated here that it cannot convert audio speech to text, which is exactly what that line states.
As for using your own audio, that's a huge bonus IMO.
As for getting the text and visemes matched up, you have to understand that TTS works best when you spell everything phonetically. "I", "Aye", and "Eye" sound alike when spoken, but you have to know which one you can and cannot use with this sort of system. It takes practice, and spelling counts.
If you saw the promo video, the horse's lines are slurred because "Anilip" was written as one word. The proper way to write this for both the visemes and the TTS system would be "Annie Lip", because "AniLip" is not a common word found in the system's lexicon. To effectively use this type of system for animating character speech, you have to learn how to write for the system's limits. If you've heard a weather radio robot voice reading off the latest weather statements, and heard it mispronounce certain words, it's the same thing here.
It would behoove anyone who wants to use this plugin to go back to their old "The Rain In Spain" speech lessons.
Ok so I bought it. How do you start it? Do you have to import an audio file first?
Nevermind. Found the docs for it (finally).
Is the text window supposed to wipe itself out when you change to a different pane? I'm writing a tutorial on it using the plugin to illustrate certain concepts, and when I swapped from the AniLip pane to the Smart Content and back again, all my text had been wiped out.
Also, Zander, the Genesis 8 figure, is reported as Not Supported by the plugin. As far as I know, he's built from Genesis 8, and the store page says Genesis 8 Male is supported.
How does it work with typos? or misspelled words?
Actually you will have to spell quite a few things phonetically in order to make the engine pronounce it correctly.
For example, to say "Daz Studio", you have to spell it "Daaz" or else it pronounces it "Doz".
"Separate" is pronounced to mean "not together", so to say "separate" as in "separate two eggs", you have to type it as "seppa-rate".
It's also inconsistent. In one test, the male voice read "recreate" in the context of "recreational" when it was written in the context of "re-creating". However, in the very next test, it pronounced both "recreate" and "re-create" identically.
It has trouble with "co-host", reading it as "c-o-host" (I'm trying to make a tutorial video covering at least two figures, male and female, but I will have to load more voices to use more figures).
Also, you cannot assign separate voices in a single animation on the fly. You will have to write the lines for one figure, save that, then do the lines for the other figure and save that, then load them onto the aniMate2 subtracks at appropriate intervals.
I only have two Microsoft voice packs on my system, and they're certainly not ideal.
It does recognize punctuation, like commas and periods, though, so how well you write will determine how well it is spoken.
Hitting a bit of a major snag with multiple speaking figures in a single scene.
You have to build the scenes in order of speaking. If you build multiple scenes with individual figures, and each one starts speaking on frame 1, you cannot move those keyframes, even with KeyMate. You can only delay the start of their motions and the audio track, but they will begin moving their mouths at frame 1.
What this means is that you will have to create successive speakers' scenes so they start talking at the correct frame.
So far I'm up to about 2600 frames for no more than my previous post, and I'm planning to insert a 2nd figure with more dialog, which will start where the first one left off.
If I continue with my plan to include 8 total figures (M5 and V5, M6 and V6, G3M and G3F, and G8M and G8F), and each one has 3000 frames, that's 24,000 frames. Even if I drop the framerate to 24, that's 7 separate files that will require dead space frames that increase exponentially.
Of course that's if I really want to avoid running them through a video editor. Since each figure's segment will be solo closeup shots talking about AniLip and how to use it, that is do-able. However, I despise the concept of post-work of any kind and believe it should be able to be done all at once in one program, or else the program needs to be improved. Crazy, I know, but time marches on and archaic ideas like post-work are simply archaic. Michelangelo did post-work not because he was a genius but because he could not do it all in one go. It's time to progress.
kudos to Dobit for at least bringing this to DAZ studio.
No sorry I won't buy it and regret the awkward Mimic live that suffers similar issues simply because I won't use DAZ studio for such animations for many of the reasons DrNewcestein is only now finding, it is not just lipsync but many other limitations, it is why I use iClone, Carrara and sometimes Poser but each to their own.
DAZ studio certainly does what it is meant to and indeed one can produce awesome work with it but not quickly straight out of the box, post production especially with animations is a must.
Sorry DrN you may need to swallow your pride and grab Hitfilm express or something and just render each of your characters talking as png series with alpha and a matching audio track and assemble it afterwards.
I do have one suggestion though I use with the much maligned Mimic live that may also work with this.
save it as an animated pose or aniblock only ticking the parts affected
find where the audio file is saved and copy it to a handy place
then load it where you want on the timeline like any pose and load the generated audio file at the same point
That may be the way to go, because the problem with multiple figures in a scene is actually deeper than I first realized. No matter what, the speaking animation begins at frame 1, even if you add the leading frames and create a key frame at the last one to start the new scene.
I mean, it's a nice first step for the plugin, no doubt, and it's bringing DS a lot closer to being at least somewhat more viable for more than just making 2D paper dolls of 3D figures, but I guess I expect far too much far too quickly.
I'll try the save-as-pose-and-insert-willish-nillish thing though.
Yeah no it's not saving the facial animations even though those are checked.
I mean, I know it's possible, because the promo video has multiple figures and voices in the same scene. Unless that was composited.
rather hard for me to troubleshoot not having the product but maybe the morphs are under the head rather than the actor tab or maybe they use facial bone movements?
Keymate should yeild some clues, I certainly save 32 bit lipsyncs as poses to reuse
Ok ok I think I got it. With the AniMate2 pane open, right click somewhere near the pane tab (or in the grey area) and check the Studio Keys option. Once the dialog is written to the timeline, you have right-click again and select Create Aniblock from StudioKeyframes. In the ensuing dialog box, select Transforms, and check Morphs/Values.
This will save the mouth animations into the aniblock. Select the aniblock and then save it. If you save it with your existing aniblock collection (as opposed to someplace outside your Content folder), it will show up in the quick list on the Animate pane.
What it doesn't do, is give you more than one voice(Anna), or language(English), like it says in the description.
HOWEVER,
what it DOES do is give you a very good mouth animation, that you can actually lip read.
If I had known that it didn't have, or was able to use, other voices(none of the voice pack downloads show up), I would not have bought this,
but I am glad I did.
It's not any more annoying than what you have to go through to use Mimic, with far better animation results.
It's not the "magic bullet" we all hoped for, but if Dobit can figure out how to get the voice packs to show up, and do an update, this will be the greatest feature of DAZ, BAR NONE.
I've got 3 voices. US male, US female, UK female (not actual names, but you know what I mean). I do not have any special TTS apps or anything, just a basic Win 8.1 Pro 64 install.
I did install some extra voice packs and the Speech API Runtime 11 just a few minutes ago, but apparently it only works up to Win 7, so I'll have to try it there and see if more voices unlock.
I think Win 10 has limited voices already, though, so that may be the issue.
Also, Win7 is limited to Anna.
Hunting down more TTS voices is a royal nightmare. They're available, but I am not paying a subscription fee for software, I'm not paying Amazon for online TTS, and I'm not paying $35 for a single voice file. There are a few free sites with online TTS, and some voices are OK, but they have character limits. Most of the free stuff I've found uses the robotic Wargames voice. I can do that myself in Audition with the built-in effects.
Still waiting for an old Mac program to be converted to Windows that let you write vocals and have the TTS system sing them, but that'll never happen I guess. It's been 20 years.
Hello DrNewcenstein,
I have tried it too and installed MS API extra stuff with Win7, but it was not shown or works?!!!?
In my Systemsettings all the Voices are visible and work proper with other Programs like 'DSpeech', 'Magix', IClone, and so on.
Only AniLip crashes :(
But I don't know why, arghh.....
Do you have a helpful tip, please?!?
Unfortunately no. There are registry hacks online that are supposed to work, but frankly I'd rather not try those at this time.
I'm digging the plugin so far, just that pesky voice actor limitation holding it back from being more useful.
Anyone know if there's a way to keyframe switching to a different camera in an animation?
How viable is this for projects where someone supplies their own actual voice acting, rather that TTS?
That I have not tried yet. My moving adventure is in limbo due to the weather, as is all my recording gear, but that's one of the things I will need to explore once I get moved and unpacked.
I can say from working with it that the timing is based entirely off of the text and the speed settings, so you may end up having to make the DS scene first, record the audio at the same tempo, and edit that audio track to match the animation.
Alternatively, with KeyMate, you should be able to expand specific portions of the animation to better match the spoken track by copy/pasting keyframes.
The promo video has been done with seperate voice recordings from the cloud service acapela-box. We adjusted the voice clips with audacity. It has features to expand or shrink audio and you can easily edit the silence areas between words.
Will this program work with Ivona SAP15 voices used by TextAloud3? I have 6 licenses which I paid a lot of money for. Would be great to use them with this program.