FREE TOOL · DIMENSIONS & LATENTS

AI Video & Image Aspect Ratio & Pixel Dimension Calculator

Calculate exact resolutions for 16:9, 9:16, 2.39:1 CinemaScope, and 4:5—with instant Midjourney --ar flags, video generator presets, and ComfyUI / Flux / Wan latent-aligned multiples (×8, ×16, ×64).

Custom Ratio & Resolution Scale

16:9 1280 × 720

Why Latent Alignment Matters in ComfyUI / Wan / Flux

Diffusion VAEs compress pixel images into latent grids using downscale factors of 8× (SD1.5/SDXL), 16× (Flux / Wan 2.1 video / DiT patch models), or 64× (tile-aligned workflows). Passing unaligned sizes causes subtle stretching, blurred edges, or tensor mismatch errors.

Copy-Ready Outputs & Latent Sizes

Midjourney Flag
--ar 16:9
Decimal ratio 1.778:1
Exact Pixel Target
1280 × 720
0.92 MP · Standard raster
ComfyUI / Wan (×16)
1280 × 720
0.00% drift · Recommended for Flux & Wan
Strict Grid (×64)
1280 × 704
Safe for ControlNet & SDXL tile nodes
VAE Divisible (×8)
1280 × 720
Minimum VAE multiple alignment
Flux ~1MP Sweet Spot
1344 × 768
~1.03 MP · Multiple of 64

AI Video Platform Quick Reference

RatioVeo / Sora / KlingMidjourneyComfyUI (×16)
16:9 Widescreen1280×720 / 1920×1080--ar 16:91280×720 / 832×480
9:16 Vertical720×1280 / 1080×1920--ar 9:16720×1280 / 480×832
2.39:1 Anamorphic1920×804 (crop/scope)--ar 239:1001536×640 / 1920×800
4:5 Portrait1024×1280--ar 4:5896×1120 / 1024×1280
1:1 Square1024×1024--ar 1:11024×1024