VisA-HD: 3D Avatars & Adaptive Optics Systems

• ViSA-HD in computer graphics is a two-stage method that couples one-shot 3D reconstruction with few-step autoregressive video diffusion to generate high-fidelity upper-body avatars.
• VisA-HD in astronomical instrumentation integrates a visible-light science arm with an adaptive optics system, enabling diffraction-limited imaging and integral-field spectroscopy.
• The near-identical nomenclature highlights a terminological ambiguity, as the two systems operate in distinct domains with different architectures and objectives.

Searching arXiv for the two cited identifiers and related naming ambiguity around “VisA-HD”. “VisA-HD” is used in the arXiv record for two unrelated technical systems whose names differ only minimally in typography. In computer graphics and generative modeling, ViSA-HD denotes the method described in “ViSA: 3D-Aware Video Shading for Real-Time Upper-Body Avatar Creation,” a two-stage framework that combines one-shot 3D reconstruction with a real-time autoregressive video diffusion shader for upper-body avatar generation (Yang et al., 8 Dec 2025). In astronomical instrumentation, VisA-HD is expanded as “VisAO-High Definition,” a visible-light science arm integrated into the Magellan Clay adaptive optics system, with both diffraction-limited imaging and fiber-bundle integral-field spectroscopy modes (Kopon et al., 2010). The near-identity of the labels suggests a nomenclatural ambiguity rather than a technical relationship.

1. Disambiguation and naming

The two usages can be separated by domain, architecture, and stated objective.

Label in source material	Domain	Core definition
ViSA-HD	3D avatar generation	A two-stage architecture coupling explicit one-shot 3D reconstruction with a real-time few-step autoregressive video diffusion shader
VisA-HD	Visible adaptive optics instrumentation	A visible-light science arm in the Magellan Clay 6.5 m Gregorian AO system, with imaging and IFS modes

In the avatar-generation usage, the central problem is generating high-fidelity upper-body 3D avatars from one-shot input image while avoiding blurry textures, stiff motion, structural errors, and identity drift. In the adaptive-optics usage, the central problem is diffraction-limited visible imaging and spectroscopy over $0.5\text{–}1.0\,\mu\mathrm{m}$, with broadband atmospheric-dispersion correction, high-order wavefront control, and efficient fiber coupling. Because both names appear in the literature with almost identical spelling, treating them as a single system would be incorrect.

2. ViSA-HD as a 3D-aware avatar-generation framework

ViSA-HD is described as a two-stage architecture that tightly couples an explicit one-shot 3D reconstruction module with a real-time few-step autoregressive video diffusion shader (Yang et al., 8 Dec 2025). Stage 1 takes a single RGB image $I$ and its SMPL-X/FLAME fit $\Theta$, and outputs both a deformable 3D Gaussian Splatting avatar in a canonical SMPL-X space and a per-pixel 3D-aware feature map $F_{\mathrm{cond}}(x,y)$. Stage 2 operates per frame $t$ on a noisy latent $x_t$, the previous frame’s output $x_{t-1}$, and the 3D-aware feature $F_{\mathrm{cond},t}$, while also using a static KV-cache of the reference image; it produces refined latents $x_{t-1}\to x_{t-2}\to \dots \to x_0$, which are decoded to a video frame $\hat V_t$.

The geometric reconstruction stage models avatar geometry and appearance by a set of $I$0 anisotropic 3D Gaussians, each attached to a SMPL-X vertex. Each Gaussian $I$1 is parameterized by position offset $I$2, scale $I$3, rotation $I$4, opacity $I$5, and color SH coefficients $I$6. The method fuses three priors: semantic features $I$7 from DINOv2, low-level features $I$8 from multiple VAE encoder layers, and a learnable human prior $I$9. These are lifted into 3D tokens by

$\Theta$0

The reconstruction network uses 5 transformer layers and outputs the Gaussian parameters $\Theta$1 together with the rendered 3D-aware feature map $\Theta$2. The reconstruction loss is

$\Theta$3

The role of priors is explicitly divided. The structural prior is the set of per-vertex learnable embeddings $\Theta$4 on the SMPL-X skeleton and provides canonical shape regularization. The appearance priors are DINOv2 semantics $\Theta$5, which preserve global identity, and VAE visual features $\Theta$6, which preserve fine textures. After transformer processing, these priors are re-projected as dense 3D-aware feature maps $\Theta$7, which condition the diffusion shader by simple channel-concatenation at every denoising step.

3. ViSA-HD diffusion shading, training protocol, and reported performance

The video shading stage is built on a distilled few-step autoregressive video diffusion model, Self-forcing, that at each step models

$\Theta$8

Conditioning is implemented in three ways: channel-wise concatenation of $\Theta$9 with the noisy latent $F_{\mathrm{cond}}(x,y)$0 before each denoiser pass; static identity conditioning via a one-time KV cache computed from the reference frame’s VAE embedding and reused in every cross-attention; and rotary position shifts to align static and dynamic frames (Yang et al., 8 Dec 2025).

The paired-supervision loss for timestep $F_{\mathrm{cond}}(x,y)$1 is

$F_{\mathrm{cond}}(x,y)$2

where $F_{\mathrm{cond}}(x,y)$3 is the tiny VAE latent. Distributional realism is further enforced by adversarial distribution preservation (ADP), in which the discriminator $F_{\mathrm{cond}}(x,y)$4 uses a frozen WAN-diffusion backbone plus a trainable head, with relativistic GAN objectives

$F_{\mathrm{cond}}(x,y)$5

Training is progressive. The two stages are trained end-to-end for short clips, and then the video shader is fine-tuned alone on longer sequences for temporal stability. Runtime considerations are explicit: the Self-forcing backbone has 1.3 B parameters; only the input convolution for $F_{\mathrm{cond}}(x,y)$6 and the self-attention weights are finetuned, while all other weights remain frozen. On an A100, the reported breakdown is 2 ms/frame for 3DGS render, no VAE-encode step because it is bypassed, 62 ms/frame for video-diffusion denoising, and 2 ms/frame for VAE decode, for a total of 66 ms/frame, approximately 15 FPS. RGB-conditioned variants are reported to spend 36 ms on encoding, totalling 101 ms, approximately 10 FPS.

On 100 held-out speakers in self- and cross-reenactment, ViSA-HD reports PSNR $F_{\mathrm{cond}}(x,y)$7, SSIM $F_{\mathrm{cond}}(x,y)$8, LPIPS $F_{\mathrm{cond}}(x,y)$9, IPS-self $t$0, IPS-cross $t$1, and speed $t$2 FPS. In the same table, GUAVA reports $t$3, $t$4, $t$5, $t$6, $t$7, and $t$8 FPS; MimicMotion reports $t$9, $x_t$0, $x_t$1, $x_t$2, $x_t$3, and $x_t$4 FPS; Champ reports $x_t$5, $x_t$6, $x_t$7, $x_t$8, $x_t$9, and $x_{t-1}$0 FPS; and VACE reports $x_{t-1}$1, $x_{t-1}$2, $x_{t-1}$3, $x_{t-1}$4, $x_{t-1}$5, and $x_{t-1}$6 FPS. Ablations state that 2D keypoints produce severe finger artifacts, 3D-aware RGB render improves structure but remains blurry, and 3D-aware feature maps yield the highest fidelity. The ADP loss is reported to improve sharpness and high-frequency detail dramatically. Qualitatively, the system is described as producing highly detailed textures such as hair strands and cloth wrinkles, fluid natural motion from 3-step autoregressive diffusion, structural consistency from SMPL-X anchoring and 3D-aware features, and neck-artifact removal through pruning of FLAME neck vertices during head-body fusion.

4. VisA-HD as a visible-light adaptive-optics instrument

In the astronomical usage, VisA-HD is a visible-light science arm built into the Magellan Clay 6.5 m Gregorian adaptive optics system, as described by Kopon et al. (Kopon et al., 2010). The two principal subsystems are the Adaptive Secondary Mirror (ASM) and the W-Unit, which houses both the Pyramid Wavefront Sensor (PWFS) and the VisAO science channels for imaging and integral-field spectroscopy.

The ASM is an 85 cm concave mirror with an $x_{t-1}$7 Gregorian prescription identical to the LBT ASM. It uses 585 voice-coil actuators with less than 1 ms response time. Its low emissivity and dual-use design allow it to feed both mid-IR science through BLINC/MIRAC4 and visible science through VisAO simultaneously via a dichroic. At the Nasmyth focus, the W-Unit accepts the incoming $x_{t-1}$8 beam, passes it through a telecentric triplet lens that converts $x_{t-1}$9, then through the advanced triplet ADC and a beamsplitter wheel. The transmitted beam goes to the Pyramid WFS through a fast steering mirror, K-mirror, double pyramid, and reimaging optics to CCD39; the reflected beam goes to the VisAO science arm.

The imaging path comprises a pick-off from the W-Unit beamsplitter, a $F_{\mathrm{cond},t}$0 silver tip/tilt mirror, a filter wheel, a $F_{\mathrm{cond},t}$1 glass plate with central coronagraphic spot, a fast shutter, and an E2V CCD47. The pixel scale is $F_{\mathrm{cond},t}$2 mas/pixel, corresponding to an $F_{\mathrm{cond},t}$3 field of view. A tip/tilt loop picks off light to a LUCA CCD at 2 kHz. In the IFS path, a beamsplitter in place of the coronagraphic plate sends light into a custom $F_{\mathrm{cond},t}$4 fiber bundle at the $F_{\mathrm{cond},t}$5 focal plane. Fine-scale mode inserts two exchangeable triplets in the W-Unit to slow the beam to $F_{\mathrm{cond},t}$6, giving a 20 mas/pixel plate scale, while coarse mode remains at $F_{\mathrm{cond},t}$7 and 105 mas/pixel. The fibers have 160 $F_{\mathrm{cond},t}$8m pitch and are fed through a microlenslet array with approximately 99% fill factor; output coupling uses a custom fused-silica aspheric lenslet array matching fiber $F_{\mathrm{cond},t}$9 to the LDSS3 $x_{t-1}\to x_{t-2}\to \dots \to x_0$0 slit.

5. Optical performance, diffraction limit, and atmospheric-dispersion correction

The system is specified to achieve diffraction-limited performance over $x_{t-1}\to x_{t-2}\to \dots \to x_0$1 (Kopon et al., 2010). The theoretical angular limit is

$x_{t-1}\to x_{t-2}\to \dots \to x_0$2

In angular units, the summary gives $x_{t-1}\to x_{t-2}\to \dots \to x_0$3 mas at $x_{t-1}\to x_{t-2}\to \dots \to x_0$4 and $x_{t-1}\to x_{t-2}\to \dots \to x_0$5 mas at $x_{t-1}\to x_{t-2}\to \dots \to x_0$6.

On bright stars and under good seeing, with $x_{t-1}\to x_{t-2}\to \dots \to x_0$7 cm at $x_{t-1}\to x_{t-2}\to \dots \to x_0$8, the reported mean Strehl over $x_{t-1}\to x_{t-2}\to \dots \to x_0$9 is approximately 20%, using

$\hat V_t$0

with laboratory and end-to-end simulations indicating $\hat V_t$1 at $\hat V_t$2. In IFS mode over $\hat V_t$3, feeding LDSS3 yields a spectral resolving power of $\hat V_t$4.

Broadband diffraction-limited imaging in the visible requires lateral chromatism of less than 10 $\hat V_t$5m at the $\hat V_t$6 focal plane, against approximately 2000 $\hat V_t$7m of atmospheric dispersion at $\hat V_t$8. The advanced ADC therefore uses two counter-rotating cemented triplets, each consisting of crown glass, flint glass, and anomalous-dispersion glass identified as Schott N-KZFS4. The wedge angles are optimized to correct both primary and secondary chromatism out to $\hat V_t$9, using the Zemax “Atmospheric” model with Magellan site parameters. The reported result is a 58% reduction in residual rms lateral color across $I$00 relative to a conventional two-doublet ADC. Laboratory verification with a white-light point source and narrowband filters at 532 nm, 850 nm, and 905 nm found measured and predicted dispersion to agree to within 0.3–0.6%, limited by lab alignment.

6. Imaging suite, integral-field spectroscopy, and operational constraints

The imaging-mode filter suite includes SDSS $I$01 at approximately 625 nm, $I$02 at approximately 770 nm, $I$03 at approximately 910 nm, and a long-pass $I$04 filter for $I$05 nm, matching the CCD47 quantum-efficiency cutoff at approximately $I$06 (Kopon et al., 2010). Neutral-density wheels are specified in 1″, 3″, and 6″ diameters with ND $I$07. Coronagraphic plates carry centrally etched chrome spots with diameters $I$08 and various apodizations, serving both as anti-blooming protection on CCD47 and as high-contrast focal-plane masks.

Spectral Differential Imaging is implemented with a Wollaston prism having $I$09 beam split, producing two polarized images on CCD47. The coronagraph wheel holds side-by-side narrowband filters, one centered on H$I$10 at 656.3 nm, [OI] at 630.0 nm, or [SII] at 673.1 nm, and the other on adjacent continuum. The summary states that line-minus-continuum subtraction suppresses speckles to $I$11 contrast at separations $I$12.

For the fiber-bundle IFS, the expected on-sky performance includes $I$13 spectra over $I$14 and 20 mas spatial resolution. Simulations for a $I$15 mag point source predict signal-to-noise of approximately 20 per $I$16 nm bin in a 1 hr integration, with seeing-limited throughput of approximately 30% including fibers, spectrograph, and detector QE. The system summary also states the principal trade-offs and enabling technologies: effective operation in visible AO requires bright guide stars of roughly $I$17 mag in median seeing; the corrected field is limited to $I$18 by isoplanatism at $I$19; user-selectable beamsplitters or dichroics trade wavefront-sensor photons against science photons; and further Strehl gains would require frame rates above 1 kHz and even higher actuator order, while IFS sensitivity remains limited by fiber losses and spectrograph throughput.

Taken together, the two meanings of “VisA-HD” occupy distinct technical domains. One combines SMPL-X-conditioned 3D Gaussian Splatting, DINOv2 and VAE priors, and few-step autoregressive video diffusion for one-shot upper-body avatar creation. The other combines an 85 cm adaptive secondary mirror, a 585-mode pyramid AO loop, an advanced triplet ADC, and interchangeable imaging and IFS channels for diffraction-limited visible astronomy. The similarity of names is therefore terminological rather than methodological.