View Full Version : The maker of Akeytsu would love to answer questions from the LW Community (Anim Prog)

02-24-2015, 08:52 AM
Hey guys Rob here,
I do hope this is kosher with the moderators of the forum. If not feel free to close this thread. No harm no foul.

I am friends with Aurélien Charrier who happens to be the maker of Akeytsu a cool new independent character animation add on application.
It is a neat program that re-thinks character animation from the ground up.

So as a favor to me he is interested in answering questions of the LW community and maybe getting to know you
(If any of you is interested)

Here is a cool video of the program:

So if you guys have question for him just pot them on this thread.

02-24-2015, 09:30 AM
You might want to point him to this very opinionated thread at the Foundry

02-24-2015, 10:50 AM
I'm curious about how it can be used with Lightwave. I noticed FBX export is coming and I'm guessing that's the planned method of file exchange. Will there also be an MDD or Point Cache export option too?

If it's an 'add on', does this mean he is considering making it a Lightwave plugin? I think I might be more interested in that but it depends on how smooth the workflow is.

Thanks for any answers.

BTW, I wonder if this thread should be moved to third-party.


02-24-2015, 11:15 AM
I have no problem if the thread is moved...
I sent him your question Greenlaw.

02-24-2015, 12:06 PM
Thanks. Will look forward to seeing what Akeytsu has to offer. :)


02-24-2015, 12:12 PM
From the demo that I saw it looks to be rigid bone deformations. Not sure if it can do secondary stuff like cloth. I'm definitely interested in seeing where this program goes.

02-24-2015, 09:18 PM
I'd love to see him address and answer the questions and issues I mentioned in this (http://community.thefoundry.co.uk/discussion/post.aspx?f=4&t=91092&p=905404) and this (http://community.thefoundry.co.uk/discussion/post.aspx?f=4&t=91092&p=905420) posts over in Foundry forums (and enumerated below).

These were just what came to mind watching their brief demo -- unordered, more to come:

UI elements like listviews and treeview all lacked visible means of filtering/searching, what is plan for addressing when hundreds of entries or more present?

Related to #1, what list ordering/grouping/management functionality is present? Context menus didn't seem to contain much in that regard, yet critical in complex scenes.

Discoverability of what entries represent in lists and tree requires manual iteration replacing main workspace. Any plan for previews other than manual iteration?

RGB "quick access" widget for axis access only seems to support constraining to a single axis, is two-axis constraining for movement etc. possible?

RGB "quick access" widget also seems to lack means of adjusting "sensitivity" of operations (f.e. moving cursor further from/closer to center when angle rotating in modo). Is there any provision for adjusting speed/sensitivity of actions using that widget, and if so, are the adjustments contextual or global? Where do users look to see value (and/or sensitivity factor)?

UI panels like the weighting panel seem to have lots of pre-populated "common value" buttons. Are such things (there are other instances) easily customizable by users for their values/needs?

Are there any provisions for procedurally- or formulaically-driven animation, or must it all be keyframed/IK? What about dynamics-driven rigs or objects? All are commonly used today in animation, so replacing existing systems without these seems quite problematic.

Somewhat related to prior entry, handling of "parenting"/attachments seem non-existent other than static rigging-based approach shown, which seems problematic as approach shown won't work well (if at all) in MANY common scenarios.

IK demo excluded foot rigging. Is IK system able to deal with modern foot rigs properly, when IK/FK foot blending needed?

Graph editor seemed to lack provisions for "curve-wide"/multi-selected/grouped management of tangents/interpolation/etc. How do users manage per-frame interpolation, generally flatten tangents across groups/periods of keyframes, and so forth?

Again related to #1, filtering/searching/selection controls in graph editor seemed minimal to outright lacking, which seems disasterous for dealing with complex scenes. What provisions exist for parametric keyframe and/or channel selections, or even channel view control?

Please demonstrate and enumerate what export workflow is expected, and how well import is handled by target package? How much control do users have over contents and structure of exported data?

Are there mechanisms/APIs for arbitrary user data being attached to any and all objects/structures/entities, and reliably saved/loaded/exported/etc?


02-24-2015, 09:41 PM
And I thought robertoortiz was against rental? :D

Ryan Roye
02-25-2015, 04:24 AM
Akeytsu needs the following:

1) Relative motion loading/motion editing. People need to be able to apply animations relative to the character's current position/orientation/scale. The ability to steer a character around as they walk or run and maintain forward momentum is extremely important... it should never be assumed that people make cycled animations just to move things in a straight line.

2) Solid retargeting. Adding to the emphasis on ability to use pre-made motions, the user needs to be able to apply these motions to any character with minimal correction needed.

3) Multi-axis editing. Akeytsu suffers from the same problem as 99% of animation packages/workflows have in the fact that its manipulations only affect 1 axis at a time... how the hell did this become a thing!? If they want to get serious with this project, they need to innovate and do away inefficiencies such at this.

4) VISUAL editing workflow. In the video, I noticed that the author was punching in numbers and tweaking text fields. Please eliminate the need to ever do this and keep the workflow in the 3d viewport and your users will thank you.

Believe it or not, Lightwave has all of the above (if you include 3rd party stuff)... but in varying degrees of complexity which can be a turnoff to users. Akeytsu needs to be *substantially* better if it ever hopes to become anything more than a gimmick, especially if they are going with a rental model.

02-25-2015, 11:27 AM
And I thought robertoortiz was against rental? :D Well I am in favor of helping out the little guy. No need to bunch up on a small developer trying to talk to us. Besides he talked to me about his flexibility on the subscription issue.

02-26-2015, 10:10 AM
Here you go guys...Ill send him more questions later today.


Even if it is a little embarrassing for me to speak directly on another software forum, I’d like to thank Roberto for having organised this discussion and you guys for your messages.
Many questions relate to the availability, or lack thereof, of certain features. To address these, there are a few main points I’d like to clarify about akeytsu :

The main point is the product’s philosophy.

akeytsu is completely different from other existing packages. It has its own identity, which some will love and some of you will probably hate. And therefore it also has its own workflow. This means that when you find certain things that are missing compared to other software, it may well be deliberate : akeytsu’s workflow expects the same result to be achieved through another way.

The second point is akeytsu v1.0 ’s target audience.

It is a huge task building an animation package from scratch, that would address the needs of all animators out there : TV/FX productions, video games, hobbyists.

This is why akeytsu v1.0 will focus mainly the video game (and especially Unity) community. I met a good number of those users, and they told me about their needs : they don’t need an animation package that can deal with complex scenes. They want to create character animations with a straightforward, easy-to-learn tool.

This is why akeytsu v1.0 won’t address the needs of each and any animator out there. However, it is actually our intent to address them in the long run.

So, bear in mind that features you may find lacking, may be lacking because of the two reasons above. Some may also be lacking because we are only approaching a Beta version. Of course, a few may be lacking because we simply missed them. Beta testing will be an opportunity to add those.

Even now, I think many of you could be surprised by everything you can already achieve with it !

Now here are answers to some of your questions :

- Rates and licensing conditions are not final yet. We are still giving some thoughts about offering perpetual licenses as well.

- @Greenlaw, our choice for v1.0 is to focus on a strong support of FBX. This way, akeytsu will be able to interact with a big majority of studios and pipelines. Import/export formats will be added along the way, according to priority wishes of our users (many already pointed us at « point cache » export). About the add-on/plug-in part of your question, it can’t apply as akeytsu is a standalone software.

- Ernpchan, we (of course) handle smooth skinning. In the demo video, I made rigid attachments only to quickly fix some issues on my props. But if you look closely values aren’t set directly to 1, they go through intermediate values. I did not take time to show the actual smoothing process. You may skin anything you want : clothes, hairs, etc.

-Jwiede, you won't have to manage (at least today) hundreds of entries via listviews and treeviews in akeytsu, so no need to have a full range of filtering features. We will only add the essential ones in the Beta version. Same thing for very complex scenes, procedural animation or dynamic driven rigs, we will take care of this later, this is not our current goal.

2 axis-constrained translation is accessible directly via the 3D standard manipulator. Both spinner and 3D manipulator work in a complementary way. Multi-axis rotate editing will come quickly too on 3D manip (cf Chazriker). No spinner sensitivity settings for the moment.

Dynamic parenting and reverse foot are planned developments. You may already tweak curve interpolation inside the curveboard using key selections. Some curve filtering options are already available (and demonstrated in the demo video).

-Chazriker, of course trajectory motion is planned (you may already keyframe a trajectory for your character’s root node as a separate layer). Concerning "full" retargeting features, the 1.0 version will include essential features, and more will be added along the way.

Thank you and best regards to all
Aurélien Charrier

02-26-2015, 02:17 PM
-Jwiede, you won't have to manage (at least today) hundreds of entries via listviews and treeviews in akeytsu, so no need to have a full range of filtering features. We will only add the essential ones in the Beta version. Same thing for very complex scenes, procedural animation or dynamic driven rigs, we will take care of this later, this is not our current goal.


I can easily see how even dozens of entries (something you showed in your demo) will be difficult to manage cases without filtering/searching, batch renaming, and so forth, but if such functionality isn't in the plan, so be it. What do you consider "essential for v1" in that regard?

Also, I note you didn't mention formulaic/expression-driven animation (useful for gears, machinery, etc.). As these kinds of anim are quite common in games, will there be some kind of support in v1?

Finally, is there any support for changing "fixed value" buttons (such as in weighting panel) to meet user needs? And/or is there any scripting+GUI support so users can easily create their own customized versions of panels' "shortcut buttons" like those in the weighting panel, custom pickers, etc.?

Thanks for answering our questions!

02-26-2015, 02:38 PM
I simply don't understand why I should use this tool other than out of sheer curiosity on my own. Other than the nice looking UI and selection disk thing, I'm not sure what kind of features this system will support. There are so many of them, so I suppose my ultimate question is:

Which features will have the greatest emphasis in this software and what general functions will be supported?

02-26-2015, 03:42 PM
Well I am in favor of helping out the little guy. No need to bunch up on a small developer trying to talk to us. Besides he talked to me about his flexibility on the subscription issue.

There is no reason he can't offer a subscription at low price and a higher priced permanent license - that way those who don't want to shell out allot of cash and who would rather pay incrementally can do so, and those who want to own it outright can as well.

How does the developer define "flexibility?"

02-28-2015, 08:42 AM
First I want thank Aurélien Charrier for taking time to answer our questions...

Here you go:

Hi Roberto,


Here are some answers below. It is quite difficult to answer all questions in details, first because unfortunately i don’t have enough time and then, because akeytsu being so different that some questions are not appropriate. I did my best.

Excellent WE to you and your little family ;-)

Hi guys,

Even though I love to answer questions, I will try to do it quickly as unfortunately time is running and I am sure you prefer I invest my time in the Beta than messages.

One more time, everything is a matter of choices and we made strong choices. In other software, you have to build everything : your interface, your features, your scripts, etc. because nothing was designed. akeytsu is the opposite of a "Do-it-yourself" product. It is a designed product. Therefore meaning less possibility for customization.

And the emphasis was put on a vital trinomial: 1) character creation/modification with ultralight rig system, 2) keyframing management and 3) character manipulation. Whatever the production type, from Iphone app to feature movie, an animator will always spend most of his day making posings and timings.

No formulaic/expression driven animation in V1. It will arrive later. Offering a generic light rig system working for any kind of character (or prop) was our priority. You may already make a machine in akeytsu. Indeed, you may spend a little more time to animate gearwheels, but less time to prepare the machine and far less time to prepare / animate all the characters around (which surely represents 80% of the production). Matter of choices !

Concerning the tree view, if you look closely at the demo, I did not use it during animation and only three times during rigging. And it is actually quite representative of its use in akeytsu's real life workflow. Again matter of choices, in akeytsu tree is not a "main" window like in other software. Picker should be preferred by far (with some shortcuts coming).

About licensing, flexibility means that we are aware of your feedback and we will do our best to offer the good balance between your wishes and our possibilities.

Thanks again for your questions and best regards to all.

Aurélien Charrier

02-28-2015, 09:49 AM
ok sounds like a "standardized" strictly CA pre-built system? Like a version of "character studio" or "CAT" if you will?

03-02-2015, 10:31 AM
Yes in effect it is. It would be a dedicated solution for JUST for animation.