Saving result of Audio2Face using python api

Hi all,

While experimenting with the Audio2Face part of Omniverse, I noticed that the python api does not allow us to save the produced video or model. It looks like the user needs to interact with the UI for this step. Is there a way to programmatically save the results or have I missed a point?

Welcome to the community, @bozkurt15 ! I’m checking with the engineering team…standby! :)

By the way, GTC Spring 2022 sessions are now available for viewing here (requires free registration): GTC 2022: #1 AI Conference


P.S. The Omniverse documentation site has more info and links to all tutorial videos (under the “Learning and Feedback” category on left nav bar): Omniverse Platform Overview — Omniverse Developer documentation

he / him | LinkedIn | Discord ID: Prof E#2041
Omniverse: “A New Era of Collaboration and Simulation” (video)
Omniverse Create’s latest features show & tell: Twitch livestream.
As seen in VentureBeat | Physics Showcase

Omniverse Resources:
Omniverse | Documentation | Tech Requirements | Videos | Roadmap | Twitter | Discord | Twitch

PLEASE NOTE: If you are reporting a bug/issue, please provide OS, GPU, GPU Driver, the version of the app, and full log file (if applicable). For crashes, please zip and provide a link to your logs → C:\Users\ [YOUR NAME] \ .nvidia-omniverse\logs

Hi @bozkurt15

Hope you are well.
There is a programmatic way fortunately. We use this as part of the omni.kit.capture.viewport extension which is generally a good reference and has a few more advanced examples in there but it roughly comes down to the following few lines:

import omni.renderer_capture
import omni.kit.viewport_legacy

renderer = omni.renderer_capture.acquire_renderer_capture_interface()
viewport_interface = omni.kit.viewport_legacy.acquire_viewport_interface()
viewport_window = viewport_interface.get_viewport_window(None)

def _capture_helper(viewport_rp_resource):
        renderer.capture_next_frame_rp_resource(frame_path, viewport_rp_resource)

omni.kit.viewport_legacy.deferred_capture(viewport_window, _capture_helper, is_hdr=False, subscription_name="omni.kit.capture capture helper") # Could be true if you are capturing HDRs

This code does require tracking the amount of accumulated samples if in Pathtraced mode.
There are some examples of that within the omni.kit.capture.viewport extension but let me know if anything comes up and we’d be happy to assist further.