inochi2d / inochi-session Goto Github PK
View Code? Open in Web Editor NEWApplication that allows streaming with Inochi2D puppets
Home Page: https://inochi2d.com
License: BSD 2-Clause "Simplified" License
Application that allows streaming with Inochi2D puppets
Home Page: https://inochi2d.com
License: BSD 2-Clause "Simplified" License
Sending an incorrect OSC bundle to the VMC endpoint will crash inochi session with an array out of bounds error. The software will not close correctly and will hang at the stack trace.
/VMC/Ext/Blend/Catte
This should not strictly require Ossia since it looks like the issue is VMC expecting a Name+Float bundle, receiving only a Float, and failing to fail gracefully.v
x86_64
Linux
v0.8.0
This issue is a fork of #41, a bug I discovered while experimenting with that issue.
When trying to run this software using integrated graphics on an AMD CPU, it segfaults (SIGSEGV
) when trying to launch. The hardware in question is AMD Ryzen 7 4700U with Radeon Graphics
, per lscpu
See description.
x86_64
Linux
Nightly (As of 19:11, UTC-5, April 23)
No response
Currently in Session there is no way to manually set a parameter without tracking. This prevents some functionalities commonly used in VTuber livestreaming, like switching on/off accessories, or switching body part poses.
No response
No response
I am trying to open a inx model in inochi-session using the following
$ inochi-session-x86_64.AppImage ./models/Midori.inx
but it's still opening an empty inochi-session
I think adding "$@" in the AppRun may help solving the issue.
ref: https://discourse.appimage.org/t/command-line-parameter-transfer/1537
$ inochi-session-x86_64.AppImage ./path/to/Midori.inx
, is expected to open inochi-session with "Midori" model but is opening empty sessionx86_64
Linux
v0.5.4
No response
No response
When running the program on a monitor with a high DPI it will not respect the system UI scaling options so the UI gets displayed really small
x86_64
Linux
0.8.3
Nothing special really
[INFO] Inochi Session v0.8.3, args=[]
[INFO] Lua support initialized. (Statically linked for now)
[INFO] Scanning plugins at /home/themoep/.config/inochi-session/plugins...
[ERR ] webcam: Unexpected end of input when converting from type string to type uint
[INFO] Found zone webcam
[ERR ] Could not start texture sharing, it will be disabled. Is the library missing?
[INFO] Saving Virtual Space...
[INFO] Saving Zone webcam...
Running on Pop!_OS 22.04 LTS with the default Window Manager (so GNOME 42 based).
Ideally it would use the scaling of the system to set the UIScale
option.
Adding a way to set model coordinates and scale exactly by inputting numbers will immensely help regarding consistency between multiple sessions in Session.
Add a way to input model scale and coordinates. These input fields can be added to a new panel, for example.
No response
No response
If we have multiple models (different costumes for one character, for example) then it will be very helpful if Session can quickly replace the model on stage with another, while retaining the same position and scale.
Add a way to quickly replace a model with another file while maintaining position and scale. Maybe add a right-click menu? Right click the model -> Show context menu, click 'Replace', and then open file browser to choose the replacement model.
No response
No response
Currently, when capturing Session in OBS via Spout2 the results are tied to Session's window size. For livestreaming i think this is more than enough, but for recording full body model in high resolution (in order to save the recording as a video and use the result later for video editing projects, for example) this is not enough unless the user has a high resolution display or using some kind of super resolution provided by GPU drivers.
Add a 'Canvas' or 'Stage' which dimensions are independent from Session's window size. Spout2 output then should render the entire canvas so users can make a high resolution recording without a high resolution desktop.
No response
No response
When running Session for the first time, the 'Virtual Space' settings is empty, and the UI doesn't help user that much without outside documentation or tutorial.
No response
No response
How to reproduce:
Doing this produces the following error message:
inochi-session: /opt/src/bindbc-imgui/deps/cimgui/imgui/imgui.cpp:8925: bool ImGui::ItemAdd(const ImRect&, ImGuiID, const ImRect*, ImGuiItemFlags): Assertion `id != window->ID && "Cannot have an empty ID at the root of a window. If you need an empty label, use ## and read the FAQ about how the ID Stack works!"' failed.
Rebuilding with debug flags and using gdb I found that this is triggered when the tracking system returns a source with an empty name.
I made this patch that stopped the problem, but I don't know if this is an appropriate solution for this problem.
diff --git a/source/session/panels/tracking.d b/source/session/panels/tracking.d
index e0b713f..c3e2e89 100644
--- a/source/session/panels/tracking.d
+++ b/source/session/panels/tracking.d
@@ -137,7 +137,9 @@ private:
uiImEndMenu();
}
} else {
- if (uiImSelectable(source.cName, selected)) {
+ if (uiImSelectable(
+ source.name.length > 0 ? source.cName : " ".ptr,
+ selected)) {
binding.sourceType = SourceType.Blendshape;
binding.sourceName = source.name;
binding.createSourceDisplayName();
This was tested on the binary provided by the 0.0.1 AppImage using a compiled cimage.so and on a Fedora 36 environment.
From what I can tell, Session currently only searches for the plugins in the user paths. If there's a system path for it, I can't see any in the logs, and a quick search over the source code doesn't show anything I recognize.
Allow setting the path to search for system plugins with at build-time, and maybe have an environment variable to override it for debug purposes.
Session should search the user plugin directories first, then the system paths - if there are any conflicting plugins, prioritize the user plugin.
No response
This is a blocker for the Flatpak version of Session.
Due to a threading issue the threads for virtual space zone items are not appropriately cleared. This causes the garbage collector to attempt to delete stale memory.
See description.
None
None
No response
No response
No response
When post-processing is toggled on (in Session or Creator), the model is covered in bright red artifacts of some kind. The artifacting is random and seems to change whenever the application is restarted--sometimes it's worse, sometimes better.
The artifacts persist in streamed views of the window on Discord and OBS (using Window Capture).
x86_64
Linux
0.8.3
I sent a model to one of my friends, whose laptop is apparently weaker graphically than my desktop, and who also runs on Ubuntu (but not Mint), and she didn't experience the same issue.
We need to get stable, nightly, release candidate and arm64 flatpaks sorted before we release 0.8, will need @orowith2os's help with this.
See above.
No response
No response
Support for the Open Sound Control (OSC) protocol as a tracker source. Possibly as a control plane for the rest of the program as well.
Benefits
An external panel can then be used to change parameters (via OSC) such that the character shows a cartoon sweat effect while a terrified button is held.
Support OSC control messages to control the software, and/or as a tracker input.
No response
No response
importing a pixelmodel only resized by 200% makes it extremely small in session even though it looks fine on creator
None
Windows
8.0
No response
No response
The idea is to add a way to see the list of animations added to a model and be able to bind them to a tracking value (either ratio or expression).
I imagine this could be achieved by either add the animation list to the tracking panel or add it to a new panel that works in a similar way.
The idea would be to bind the animation to a blendshape or expression, trigger the lead in animation it when it goes to a value >= 1 and after that loop until the value goes <= 0, what would trigger the lead out animation and stop.
Values could be configurable though, and
No response
I may make a small video if I have the time...
Currently the viewport follows app's current theme (light or dark). It will be very useful if there is a switch between light/dark viewport background in the toolbar.
No response
When running inochi-session on Wayland if you start moving the window, the model or the opaque things around, they appear to leave a visible trace that doesn't show on screenshots, but also doesn't leave until you refresh the window some way (like minimize + maximize).
Here is a picture that I took on my phone of the issue:
This doesn't happen when using X11 or when running on Wayland using the X11 SDL video driver (setting the envs SDL_VIDEODRIVER=x11 SDL_VIDEO_X11_FORCE_EGL=1
).
This also stopped happening when commenting the line that sets SDL_HINT_VIDEO_EGL_ALLOW_TRANSPARENCY
to 1
on my debug build.
This was tested on the binary provided by the 0.0.1 AppImage using a compiled cimage.so and on a Fedora 36 environment (SDL2-2.0.22).
That is, I hope to say that the dog can support the camera to capture the face and shoulders and arms and so on and apply to Live2D image
Look at the description
No response
No response
I have a puppet which I already asociated tracking parameters before. When I try to add a new tracker on the Virtual Space menu, the app crashes
View > Virtual Spaces
and add a space and a trackerx86_64
Linux
0.8.4
No response
It crashes here:
If a virtual space has two sources, the BlendShapes from the lower source are not read by ratio bindings if the upper source is not currently active.
Expression bindings using the BLEND(...) function work as normal, and the BlendShape tab also shows the values as expected.
In my specific scenario my virtual space has:
I have tested this in both orders and the bug is consistent for BlendShapes (Bones simply did not transmit regardless when OSF was lower). If the source that is higher on the virtual space is not transmitting data, then the lower source BlendShapes will experience the bug.
x86_64
Windows
0.8.3
No response
No response
Hey there. I gave the latest release build a quick test, and ran into the following error:
./inochi-session-x86_64.AppImage
[INFO] Inochi Session v0.5.4, args=[]
[INFO] Lua support initialized. (Statically linked for now)
[INFO] Scanning plugins at /home/ezequiel/.config/inochi-session/plugins...
object.Exception@inui/source/inui/core/window/appwin.d(235): No valid OpenGL 4.2 context was found!
----------------
??:? [0x558e2abc8096]
??:? [0x558e2abc9ba6]
??:? [0x558e2abab1af]
??:? [0x558e2aab35ea]
??:? [0x558e2aa4cfa5]
??:? [0x558e2aa633f9]
??:? [0x558e2abaae9b]
??:? [0x558e2abaad95]
??:? [0x558e2abaabed]
??:? [0x7f4d21a3554f]
??:? __libc_start_main [0x7f4d21a35608]
??:? [0x558e2aa491e9]
fish: Job 1, './inochi-session-x86_64.AppImage' terminated by signal SIGSEGV (Address boundary error)
Running the command in the bash shell (fish is my usual) adds this line to the end of the error message:
Segmentation fault (core dumped)
My system information:
3D and OpenGL apps in general seem to work, so I really don't think the problem is with my system specifically. I should probably also mention that I get this error with the zipped version of the release as well.
When deleting a Virtual Space, it crashes if the IP cannot be resolved.
1 Create Virtual Space
2 name it (main in this case)
3 Select VTubeStudio
4 phoneIP set to 0.0.0.0:8001
5 save changes
6 save (bottom)
7 reopen menu
8 try delete it (will crash)
x86_64
Windows
v0.5.4
0x00007FF70D50910D in fprintf
0x00007FF70D508E76 in fprintf
0x00007FF70D4FCCFF in fprintf
0x00007FF70D4E68CA in fprintf
0x00007FF70D4C7668 in fprintf
0x00007FF70D43794E
0x00007FF70D4EE008 in fprintf
0x00007FF8018D9363 in recalloc
0x00007FF802B326BD in BaseThreadInitThunk
0x00007FF8040CDFB8 in RtlUserThreadStart
No response
After loading a model, the entire scene becomes invisible, leaving only a black background and inochi-session's UI. This issue has been observed on Intel GPUs and software rendering; Nvidia GPUs are not affected. At least versions 0.8 up to the latest nightly (a462922) are affected.
Using an Intel GPU or software rendering:
x86_64
Linux
$ export LIBGL_ALWAYS_SOFTWARE=1 MESA_DEBUG=1 GALLIUM_DRIVER=llvmpipe
$ ./inochi-session
[INFO] Inochi Session v0.8.0, args=[]
[INFO] Lua support initialized. (Statically linked for now)
[INFO] Scanning plugins at /home/user/.config/inochi-session/plugins...
Mesa: User error: GL_INVALID_OPERATION in glFramebufferTexture2D(window-system framebuffer)
Mesa: 7 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glFramebufferTexture2D(non-existent texture 4)
[ERR ] Could not start texture sharing, it will be disabled. Is the library missing?
Mesa: 7 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
[INFO] Created expression p3304717045_0...
[INFO] Created expression p3304717478_0...
Mesa: 374 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 4 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_OPERATION in glDrawElements
Mesa: 2 similar GL_INVALID_OPERATION errors
Mesa: User error: GL_INVALID_OPERATION in glDrawBuffers(unsupported buffer GL_COLOR_ATTACHMENT0)
Mesa: User error: GL_INVALID_OPERATION in glDrawArrays
Mesa: User error: GL_INVALID_ENUM in glBlendEquationSeparateEXT(modeRGB)
[INFO] Saving Virtual Space...
I created this to formally document a string of issues reported on the #support
Discord channel.
So far, this issue has been observed on:
This issue does not affect Nvidia GPUs or Mesa softpipe.
Drag-and-dropping model file from file manager doesn't come naturally for a lot of users, so a clear and obvious File->Open menu will help a lot.
Add 'File' menu with 'Open' function to the existing top menu.
No response
No response
Adding background image transform will grant more options to the scene, and being able to bind those transform to expression bindings will be a nice added option. This for example can be used for some kind of background parallax effect combined with the character moving left and right, which is already common in terms of rigging features.
No response
No response
This is old issue but since it's not listed here yet I thought I'll add the entry. As said in the title, enabling post-processing causes trailing images to appear.
1, Open Scene Settings
2. Enable Post Processing
3. Move loaded model around.
x86_64
Windows
62a328
No response
No response
At the moment, running Session with no Virtual Spaces, hidden UI, and no model loaded takes 33% CPU vs 13% CPU took by VTube Studio when no model is loaded and webcam tracking is disabled (i3 9100F).
x86_64
Windows
0.8.3
No response
No response
Now that animation system is implemented in Creator, adding animation recording feature in Session will be very nice (similar to what VTS already have).
Add animation recording feature to Session. Even a 'dumb' one which just records parameter values at every set amount of time without any smoothing or predictive keyframe placement will be very useful already.
I'm not sure what's the best way to do it if Session is now tied to .inp files only, but if we have the recording feature plus animation import feature in Creator (as per my previous suggestion) then user can record via Session -> Save to file (.inp) -> Open their .inx file in Creator -> Import animation from the .inp file.
No response
No response
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.