ComfyUI: Generate synced AV with LTX V concat node

Demo summary

The creator demonstrates using the LTX V concat AV latent node to bind video and audio together before sampling to ensure perfect synchronization.

Use the LTX V concat AV latent node to bind the latent video and latent audio together
Send the concatenated latents into the K sampler for stage three
Use a node to split the video and audio latents back apart after sampling
Pass the latent video through the prepped upscale model
Reduce the sampling steps for the second K sampler in stage four
Perform a final VAE decode to generate the high-res video

You must use the LTX V concat AV latent node to ensure synchronization
Sampling steps must be aggressively reduced during the stage four upscale to avoid breaking original motion

Binding video and audio from frame one guarantees they are glued perfectly together
Stage four is where the quality skyrockets by using an upscale model before the second sampling

“you're left with a final razor-sharp high-res video with perfectly synced audio in your hands”

2:590:13Search and load LTX 2.3 templates in ComfyUIThe creator demonstrates how to use the template search feature in ComfyUI to find and load the LTX 2.3 image-to-video workflow.ComfyUI· Image to Video
3:411:25Configure LTX 2.3 foundational parametersA walkthrough of setting dimensions, frame counts, and loading core models including the LTX 2.3 distill model and audio VAE within a ComfyUI workflow.ComfyUI· AI Animation Generator
5:060:37Process visual and audio inputs in ComfyUIThe demo shows how to use the LTX V image-to-video condition node and empty latent audio node to prepare data for the generation engine.ComfyUI· AI Animation Generator
5:431:03Generate synced AV with LTX V concat nodeCurrentThe creator demonstrates using the LTX V concat AV latent node to bind video and audio together before sampling to ensure perfect synchronization.ComfyUI· AI Video Generator
10:250:53Optimize step count and sigma coefficientsThe creator demonstrates adjusting the configuration to 12 steps and tuning sigma coefficients in ComfyUI to improve character consistency.ComfyUI· AI Image Generator
12:200:26Integrate independent video upscale workflowThe demo shows how to replace the standard LTX resampling block with a dedicated video upscale workflow for cleaner high-resolution results.ComfyUI· AI Video Upscaler
Watch “Your LTX 2.3 ComfyUI Workflow Is Probably Wrong… Fix It Like This” →

5:431:03Generate synced AV with LTX V concat nodeCurrentThe creator demonstrates using the LTX V concat AV latent node to bind video and audio together before sampling to ensure perfect synchronization.SOTAI