OpenToonz V1.4 was released!

[Shun Iwasawa]

Dear users,

Thank you for your feedback and patience to the release candidate version.

We just released OpenToonz v1.4. The new version has improved usability and stability, also introduced new features.

Release Highlights

• New Guided Drawing/Autoinbetweening
• New XDTS File Support (Celsys and Toei Animation xsheet file format Import/Export)
• New Cache folder Replacement & Clear Cache Command
• New Camera Column in XSheet
• New Context Aware Toolbar
• New Create Blank Drawing command
• New Fx: Corridor Gradient Fx Iwa
• New Fx: Spin Gradient Fx Iwa
• New Geometry Tool Driven Motion Paths
• New (Initial) Korean Menu Translation
• New Level Settings Adjustment for Multiple Selections in XSheet/Timeline
• New Next/Previous Keyframe Shortcuts
• New Pencil Mode for Toonz Raster Eraser in Freehand and Polyline
• New Swap Colors Button for Gradient Fxs
• New Short Play Feature
• New Fill Tool for Raster Level
• Updated/Reorganized Menus
• New Canon DSLR Camera Capture support (Windows only)
• New Portable Test Builds via zip file download at Github (Windows Only)

..and much much more. (Please see the release notes for details.)

[OpenToonz Website] https://opentoonz.github.io/e/

Click "Download" > "OpenToonz"

Thanks!

Shun

----------

Shun Iwasawa

OpenToonz team

[Pipeline TD]

Congratulations! So many new features!

[Rodney]

Congrats to all involved in the release of v1.4.

It has been a joy seeing everything come together.

There are so many hidden fixes and features that it is hard to list much less explore each one.

The proof however will be in the use of Opentoonz and I'm already seeing new names and faces appear online so that is a good sign.  And to you older (ahem... more seasoned) users... let's work to make v1.5 even better!  :)

[Aaron Wiggan]

Very very good work and much appresiated!

[Isaac Tut]

Is there anyone that I can work with to integrate this with an AI app? Thank you in advance!

[Michael Watkins (ELF Productions)]

Ive seen people take frames that they have animated to do AI inbetweens. What kind of AI did you want? for the most part you will export your animation to then import it to whatever AI you want.

[Isaac Tut]

Is everyone against AI here also? Would anyone be okay with me adding AI tools to OpenToonz? Maybe these models below?

 - https://pkhungurn.github.io/talking-head-anime-4/

- https://app.klingai.com/global/

- https://github.com/Tencent-Hunyuan/HunyuanVideo-Avatar

- Other Models for anime character animation and movements

I want to make sure people actually find these things useful before doing it.

Best, 

-Isaac tut

[Rodney]

Isaac,

There will be many who don't care for ai.

Wha I tell them is that they just haven't found a use for it... yet.

I'm a fan of AI as it is a useful tool.  (People will use ai whether it is in Opentoonz or not)

I'm not sure of the specific use cases you refererence but if you create a Pull Request for any addition to Opentoonz it will certainly be considered.

The potential issues with the references you mention might include:

Talking Head anime  (limited use if the results cannot be rendered or easily setup)

KlingAI (Not sure how useful it will be for those that don't have a paid account)

HunyanVideo-Avatar (Requires Comfy UI?)

Aside:  v1.4 was a very long time ago.

Personally, just thinking about different models/approaches one of the most useful integrations would probably be Hunyan's 3D generator which would take an image by an Opentoonz user and generate a texured 3D (GLB) model from that.

That model could then be turned around in space and captured for use in Opentoonz and/or other program.

That would be no small task.

The idea would be to consider what usages ai can bring to the table leveraging the original content of Opentoonz users.

We do want to steer clear of anything that would come across as a gimmick.

Users will see that as code bloat.  Better to consider how much the feature set will be used.

There is no way Opentoonz can keep up with the various image generators available so my concern would be how quickly any ai feature set can become outdated.

Very interested in hearing more of your thoughts on this.

-Rodney

[Michael Watkins (ELF Productions)]

This tool cuts time spent drawing in-betweens by half

this kind of AI would be most welcome hehe

that first AI of the head could be useful if it can spit out a dopesheet for lip flap keyframes that can be used in opentoonz since opentoonz isnt a realtime software. As for generative AI i dont see a use for it within opentoonz. If you want to generate something outside of opentoonz and then bring it in maybe as a template or background or something sure.

[Isaac Tut]

Oh okay, thanks for the feedback, everyone!

Based on your suggestions, maybe a good place to start would be these AI tools to streamline animator workflows:

• Lip-sync automation: LivePortrait

• In-betweening frames: ToonCrafter

• 2D-to-3D asset generation: Hunyuan3D

• Maybe an additional AI-powered character reposing tool.?

Would the above tools be a good starting point, or are there other immediate workflow needs you think should take priority?

@Michael.Watkins - for importing things generated outside of OpenToonz, are you referring to images/videos or what do you see as the use case there? 

Appreciate your thoughts!
-Isaac

[Michael Watkins (ELF Productions)]

how does that toon crafter work in opentoonz? or is it just a stand alone

[Isaac Tut]

Hi @Rodney,

I was able to test the tools/workflows I thought would be most helpful - results below.
Also - I was recently talking with a friend in the music industry (he’s written for Drake and others) and he shared his frustration about how AI has been trained on artists’ work without consent or compensation — and how it’s hurting those who’ve worked hard to make it(Animation industry included). That really stuck with me. Just to be clear, I genuinely want to understand what you think would actually help your workflow - not here to exploit or replace anyone.

Here are the tools + Test Results:

• LivePortrait for quick lip-syncing tests — pretty solid for static portraits, but limited for anything other than face motion, not for complex motion. Another example here.

- Hunyuan + Vance + Sam2 --> Options for character reposing (still early stage), Best option for Lipsync when combined with LivePortrait.

• ToonCrafter for in-betweening — Seems to work best when using images from existing anime shows or their provided images- otherwise I wouldnt use it.

         - The test files were not good enough to show 

• Hunyuan3D for generating 3D references from 2D art — early results are promising -- I have not tested this that much - but ive seen good results with it.

I’d love to hear your thoughts:
- Which parts of your workflow actually need support from these tools
- Where could AI reduce grunt work without stepping on creative control?

Best,

-Isaac

[Rodney]

Isaac,

I certainly appreciate your interest.

I get how people are concerned with AI... they should be.

But... AI isn't just coming... it's already here.

It's not going away.

To jump directly to your questions:

- Which parts of your workflow actually need support from these tools

 This is a complex answer perhaps because I am not just a user of Opentoonz but assist with development.  I must assume your focus here is at the user level and not so much the developer-user level athough I see them really as being one and the same just to different degrees.

What I tend to use AI for is probably three fold:  1) Exploration/Research:   Trying to understand what can be done relative to a given task/goal 2) Automation:  I've delved much more deeply into pythin since ChatGPT arrived on the scene and that has been a pretty incredible experience.  3) Organization:  I think we all struggle with organizing our projects 

There are more and I'd say a 4th might simply be Entertainment.  It's quite entertaining to ask ChatGPT for something, get it back it seconds and iterate on the idea.  This can fall into area 1) Exploration in my estimation.

If you run a survey of what software companies are adding via AI to their programs you'll start to see some patterns.

Many are trying to navigate around the concerns of users so the approaches vary but there is a reason why most companies pushing AI are moving toward co-pilot/assistant models; they help you do the things you want to do and not just do it all for you.  There can be some smoke and mirror here but this represents an effort on the part of those who see the bigger picture of where AI is heading and are trying to maintain commerciality so they can get return on investment (and profit).

As a user, finishing is not my strong suit.  It does us little good if we start a project and never finish it.

The part of workflow that actually needed from these tools is support in finishing a project.

 

- Where could AI reduce grunt work without stepping on creative control?

Returning to the 'survey' this seems to be where the focus is for many companies.

Have AI take some raw data and break that down into smaller tasks with defined goals.

Toonboom's entry into AI (Toonboom Ember) is an optional addon that allows for such things as breaking a script down into storyboard panels.

I haven't used this so I assume the panels are empty although I'll guess they won't have to be and likely won't be in the future when the user populates the system with reference images and assets.

Local operation presumably keeps those resources secure.

(See more below)

Other aspects of their implementation include upscaling resolution of images (useful for zooming in closer on an image that otherwise will be too low resolution), inpainting to remove or replace an object and extending an image so it includes additional imagery at the edges needed in a scene.

None of these are original although the script to storyboard panels is probably the most original.  What Toonboom Ember does is just allow users to do that inside the program versus using another program/service to get the job done.

Recently (mere days ago), I decided to try to add a local image to 3D generator to my system and after an initial fail in setup... it works as expected.

That program is called StableProjectorz and in addition to other things generates 3D GLB models from one or more images.

I then had ChatGPT help me create a python script that uses Blender (although Blender is never opened) to render a turnaround of that model with a transparent background.

This was more of a test to see if I could set it up than anything but... it works as expected.

Quality and capability will only get better.

The resulting imagery can be used as reference or actually in a shot.

If this then is my imagery from start to finish... the main critique would be 'I don't like the style/result'.

The speed of the process, yeah, that's going to be hard to match without the AI assist.

This is one of many reasons why I would suggest embracing AI.

That and perhaps "Know thy enemy!"  ;)

In terms of 'grunt work' I'd say this was exploring how I might approach using AI to fill the gaps in what my current workflow and Opentoonz capability can achieve.

We'll have to examine what 'creative control' is more to dig much further.

- Rodney

[Rodney]

Isaac,

You mentioned your friend with interest in music and I don't want to neglect that.

Opentoonz needs development on that front and I suspect AI might be a golden opportunity to not only bring audio features up to date but bring them into the modern age.

My one suggestion related to these concerns would relate to 'sourcing'.

Computers are really good at this kind of thing and there is no reason why a production shouldn't be able to source to such a degree that we tire of reading the credits/sources.

It's like sitting through the credits of a movie where we likely only want to if we know someone we expect to see in those credits (or patiently wait through them to get to the extra scenes.

Things will get dicey at the point where we really can't correctly attribute a credit/source.

It can be really hard to source an inspiration and I don't think we want to live in a world that expects every thing that inspires us to be documented fully and proper monetization attributed.

It can be done but will never be perfect.   Someone will get the macro dollars.  Someone will get the micro dollars.  Someone will feel left out and/or even cheated.

But again, computers are pretty good at this stuff and AI can certainly provide sourcing.

There will surely be some friction in the future over who owns this and that and what constitues derivative work.

Ultimately, where there is money people will gravitate.

'The worker is worthy of his hire' and should be reasonably compensated.

Proper credit where credit is due is important and assigning our names to someone else's work is obviously wrong.

None of this is mysterious however.

It should be common sense.

A key component of AI is the training and the better the training material the better the results.

The better the sourcing of that training data the more likely the originator will recieve the credit due.

It is my current view however that if someone uses my drawings and derives something original from that I'd love to get a tip of the hat but I don't feel it to be a requirement.

There are some finer points to consider when specific styles are being used but I question just how original any style truly is in light of properly sourcing how those style came to be in the first place.

Ultimately, where things get dicey is when artists are trying to pay the bills.

It's always been tough to be an artist and that doesn't look to be changing any time soon.

I am confident that the problematic areas of AI can be resolved.. and they must... because AI isn't going away any time soon.

- Rodney

 

[Isaac Tut]

Hey Rodney,

Thanks for the thoughtful reply — a lot of what you said really resonates. I agree completely: AI isn’t going anywhere, so it’s on us to be at the forefront to help shape how it actually supports artists instead of sidelining them.

I’ve been messing around with some automation tools to see what might actually reduce friction for creators, I even created a platform for creating anime via workflow automation(script-to-story pipeline). But lately I’ve been more focused on motion mimicking — trying to handle repetitive tasks without touching the creative core(getting consistent good scenes is VERY hard with AI). Hence why im interested in OpenToonz.

I think AI will make it easier for independent artists to get their work out without needing huge teams or budgets, and I think there's a real opportunity here, especially for those who are super talented but get held back by production overhead or lack of access. Even if they don’t want to do it all themselves, they should at least have the option — and get credit and fair compensation when they do. Over the past 4yrs, the more I learn about the anime industry the more I feel like it might implode if things don't change(>$200k/episode, overwork, poor pay, etc)

Appreciate you sharing your experiments too — the Blender/GLB setup sounds awesome - I'd love to see more on what you're testing. Ill also talk to my friend in the music to see what he thinks(I know nothing about music production). I’d love to keep learning from what you’re testing.

p.s - let me know if you have any animation projects going on - I would love to help or collaborate.

– Isaac

[Rodney]

Isaac,

My main (animation) project these days usually consists of creating some short doodle while testing a change of code submitted for Opentoonz.

I figure I have to test something so it might as well be an excuse to draw and animate.

These tend to be more still drawings that are moved about or short sequences of two or three frames than fuller animation.

I do need to lean more into full animation but that takes focus and time that I don't always feel I have when testing.

I do a weekly drawing session with my niece where I tend to take 30 minutes or so out of an hour to work from some initial concept to (hopefully) something worth calling 'done'.

In other words, no real animation projects in the pipeline.

Just lots and lots of testing and short form animation.

I share a lot of what I do draw in the Opentoonz discord where we troubleshoot users problems and explore.

Regarding those that work in anime/animation... change is in the wind.

It was rough moving to digital... rougher navigating the onslaught of 3D... now... the next opportunity(TM) is arriving in the form of automation/ai.

It isn't just animators that are going to feel the effects of this one.

You hit on one of the main issues with AI... the consistency across scenes... and they are rapidly dealing with that.

This is probably not the ideal place to share info on all of this (although I'm certainly not opposed to seeing more discussions in this forum).

 As my focus is on Opentoonz I try to relate most things in some way to working in OT.

And even after 10 years of OT... still learning!

(and we've just scratched the surface)

- Rodney