When merging images horizontally, differing native heights create awkward letterboxing unless mathematically scaled. The studio calculates a target baseline height H_target = max(H₁, H₂, ..., H_n). For each individual image i, the scaling multiplier is S_i = H_target / H_i, and the normalized rendered width is W_scaled = round(W_i × S_i). This preserves 100% of optical proportions without squishing or stretching pixels.

Scale to Match expands or contracts each photo so they align flush along the shared baseline. Crop to Fill computes optical center coordinates (W_src - W_crop) / 2 to uniformly fill identical square or 4:3 grid cells—ideal for Instagram carousels. Pad / Letterbox preserves every single pixel of mixed portrait and landscape originals by centering them inside normalized bounds and filling negative gutters with your chosen solid canvas color or transparent alpha.

The canvas rendering engine uses hardware-accelerated 2D context buffers. For ultra-wide panoramas exceeding standard viewport dimensions (up to 8,000+ pixels wide), interactive CSS matrix transforms allow smooth pan-and-zoom inspection without allocating redundant high-resolution raster copies in DOM memory until final rasterization.

Every step of image decoding, matrix transformation, color blending, rounded corner path clipping, and binary ZIP archiving occurs entirely inside your visitor's browser memory. Zero files, photo metadata, or generated graphics are ever uploaded to RiazHub.com servers or external cloud endpoints.