git:20260819.9fe9ea3 to git:20260825.ea35391

- description: Upscale and restore video in ComfyUI — both the quick local path (per-frame ESRGAN like 4x_foolhardy_Remacri via ImageUpscaleWithModel + 4x→2x supersample, with its temporal-flicker tradeoff) and temporal-aware super-resolution (SeedVR2, the newer FlashVSR) with the downscale-first restore pipeline; RIFE/FILM frame interpolation via the BUILT-IN ComfyUI 0.26 FrameInterpolate (rife_v4.26 in models/frame_interpolation/) or the ComfyUI-Frame-Interpolation pack; 2x/4x scaling, VRAM tiers, VHS encode. Captures the classic downscale→SeedVR2→RIFE recipe and the current 2026 recommendation.

- 1. **Spatial restore + upscale** — a *temporal-aware* model that increases

- resolution **and** cleans compression blocks, blur, and AI-gen mush while

- (ESRGAN, UltimateSDUpscale per-frame) does badly — per-frame upscalers

- **flicker** because each frame is sharpened independently. Use a video model.

- 2. **Frame interpolation (VFI)** — synthesize in-between frames to raise fps

- (e.g. 24→48/60) for smooth motion. Do this **after** the spatial pass.

- 3. **Encode** — mux frames (+ original audio) back to an MP4.

- The two leading temporal restorers in 2026 are **SeedVR2** (diffusion-transformer

- restorer, the proven workhorse) and **FlashVSR** (newer one-step streaming VSR,

- faster). Frame interpolation is **RIFE** (or FILM) via

- **ComfyUI-Frame-Interpolation**.

- > ⚠️ Verification note: every node/pack/model name below was confirmed against the

- > can't confirm is installed — check with `install_custom_node` (`action: "list"`) / `create_workflow (action:"node_info")`.

- **Downscale → SeedVR2 (temporal restore+upscale) → RIFE (interpolate) → VHS encode.**

- the restore stage (see "FlashVSR" section) — swap it in when speed matters more

- `LoadVideo` / `GetVideoComponents` / `CreateVideo` / `SaveVideo` are **core

- ComfyUI** video nodes (same ones the official comfy.org SeedVR2 template uses).

- `VHS_LoadVideo` / `VHS_VideoCombine` come from **ComfyUI-VideoHelperSuite**

- **audio** through.

- - **The restorer wants a *clean* low-res input, not a big dirty one.** SeedVR2

- to *synthesize* sharp detail rather than faithfully magnifying existing

- compression artifacts/noise. Downscaling first averages away block noise →

- the restorer hallucinates clean, coherent texture.

- - **VRAM + speed headroom.** Cost scales with input pixels × frames. Halving each

- batch (the thing that kills flicker — see below) and a bigger target multiple.

- - **It turns "upscale" into "restore-and-upscale."** A 720p source downscaled to

- 360p then SeedVR2'd to 1080p+ looks dramatically better than 720p→1080p

- - Rule of thumb: downscale to **0.5×** (or to a ~360–480p short side) for messy /

- low-bitrate / AI-gen footage; **skip the downscale** for already-clean,

- high-bitrate sources where you just want more pixels.

- When the user just wants a fast result on **what's already installed** (no SeedVR2 /

- Check first with `list_local_models` (common ones: **`4x_foolhardy_Remacri`** — best

- for realistic footage/water/skin — and **`4x-ClearRealityV1`** for clean/sharp).

- - **Upscale:** `ImageUpscaleWithModel` with a 4× ESRGAN model, then `ImageScale`

- **back down to a clean 2×** (a 4×→2× **supersample**). That downscale-after step is

- the single biggest quality lever here — it averages out per-frame noise.

- - **Interpolate:** the **built-in** `FrameInterpolate` with **RIFE v4.26** (see the

- Frame-interpolation section — no custom node on 0.26+).

- > ⚠️ **Tradeoff — flicker.** ESRGAN upscalers are **per-frame** (no temporal

- > awareness), so they can **shimmer/flicker** on video — most visible on water and

- > signal to switch the upscale stage to a **temporal** model (**SeedVR2** / **FlashVSR**

- > below) — the real fix. So: per-frame ESRGAN = quick & local; SeedVR2 = flicker-free

- > & best. **Order is unchanged:** upscale the real frames first, *then* interpolate.

- **Pack:** `ComfyUI-SeedVR2_VideoUpscaler` (author **numz**) — GitHub

- **Node classes (4):**

- **Models** (auto-download to `models/SEEDVR2/`; 3B = lighter, 7B = best quality):

- **Key params on "SeedVR2 Video Upscaler":**

- **3B vs 7B:** start with **3B fp8** — it's the speed/quality sweet spot for most

- footage. Move to **7B** only when you need maximum reconstruction on faces/text

- **Pack:** `ComfyUI-FlashVSR` (author **1038lab**) — GitHub `1038lab/ComfyUI-FlashVSR`,

- built on **FlashVSR V1.1** (one-step diffusion + locality-constrained sparse

- **Node classes:** **`FlashVSR ⚡`** (preset: Fast / Balanced / High Quality) and

- **`FlashVSR Advanced ⚡`** (`model_version` = Tiny / Tiny Long / Full,

- Supports **2x and 4x** (4x recommended). Needs **≥21 input frames**. Variants:

- **Full** (best, heavy VRAM), **Tiny** (fast), **Tiny Long** (low VRAM).

- SageAttention adds ~20–30% speed.

- **When to prefer FlashVSR over SeedVR2:** real-time / long clips / speed-critical

- jobs, or when SeedVR2 is too slow on your hardware. **Prefer SeedVR2** when you

- Current ComfyUI ships a **core** frame interpolator — nodes

- **`FrameInterpolationModelLoader`** + **`FrameInterpolate`** (RIFE/FILM) — so on

- **0.26+ you do NOT install a custom node**. It auto-detects checkpoints dropped in

- **`models/frame_interpolation/`** (that folder is **empty by default**, which is why

- **`Comfy-Org/frame_interpolation`** (under a `frame_interpolation/` subpath):

- Drop a file into `models/frame_interpolation/` and **restart** so the dropdown

- populates (ComfyUI caches model lists). `download_model` may not target that folder —

- pull it **directly** into `models/frame_interpolation/`. Check what's there with

- **Pack:** `ComfyUI-Frame-Interpolation` (author **Fannovel16**) — GitHub

- **Primary node: `RIFE VFI (4.0 - 4.9)`**

- After interpolation, set the encode node's fps to **source_fps × multiplier** so

- **Alternatives in the same pack:** `FILM VFI` (Google FILM — excellent on large

- two need **≥4 input frames**), `IFRNet`, `M2M`, `AMT`, etc. There is **no

- "GIMM-VFI" node in this pack** — if a workflow asks for GIMM-VFI it's a separate

- custom node; verify it's installed before citing it. Default to **RIFE**; reach

- for **FILM** when RIFE smears fast motion.

- The user's "RIFE 56" is shorthand, **not** a RIFE model version — RIFE in this

- pack tops out at **4.9** (`rife49`). It almost certainly means **RIFE targeting

- ~56 fps** (i.e. a multiplier chosen so the output lands near 56 fps — e.g.

- fps directly. Map it to: **`RIFE VFI (4.0 - 4.9)`, ckpt `rife47`/`rife49`,

- multiplier = round(56 / source_fps)** (multiplier 2 from 24/25/30 fps). Confirm

- - **2x** — safest default; pairs with the downscale-first trick (downscale 0.5×,

- then 2x back recovers original size but *restored*). Lower VRAM, fewer artifacts.

- - **4x** — for genuinely small sources or when you need a big jump; FlashVSR's

- recommended factor. Costs ~4× the pixels — expect VAE tiling / block swap.

- - With SeedVR2 you don't pick a literal "2x/4x"; you set the **target short-edge

- `resolution`** and the effective factor falls out of input vs target size.

- General: **downscale first** to buy a bigger batch; always `clear_vram` before

- - **Temporal flicker** → the #1 video-upscale failure. Cause: per-frame

- (non-temporal) upscaling **or** too-small a SeedVR2 `batch_size`. Fix: use a

- - **Frame-count constraints**: SeedVR2 `batch_size` must be **`4n+1`**;

- FlashVSR needs **≥21 frames**; STMFNet/FLAVR interp need **≥4 frames**. A clip

- **`VHS_VideoCombine`** (VideoHelperSuite) and feed it the `audio` from

- - **Models auto-download on first run** for SeedVR2 and FlashVSR — the first

- The user's older, battle-tested pipeline — still solid — is:

- > **DOWNSCALE the video first → SeedVR2 → RIFE ("RIFE 56").**

- That is exactly the structure the **recommended 2026 pipeline above preserves**:

- downscale-first to give the restorer clean input + VRAM headroom, SeedVR2 for the

- - Use the current **`ComfyUI-SeedVR2_VideoUpscaler`** node pack (4-node:

- DiT loader + VAE loader + [torch compile] + upscaler) with the **3B fp8** model

- - Read **"RIFE 56"** as *RIFE targeting ~56 fps* (a multiplier, typically **2**

- from 24/25/30 fps), using `rife47`/`rife49` — not a literal model version.

- - Consider **FlashVSR** as a faster drop-in for the SeedVR2 stage when speed

- `1038lab/ComfyUI-FlashVSR`. SeedVR2 and FlashVSR weights **auto-download on first

- run**, so `models[]` can be left light — note that in `pack.yaml`. Install nodes

- ad-hoc with **`panel_install_node`** or apply a manifest with

- **`apply_manifest`**. Offer to contribute a finished pack upstream