Bit allocation is the rate-control decision of how many bits to spend on each frame inside a GOP. Not all frames are equal: an I-frame stores a full picture and gets the most bits; a P-frame describes itself as a small change from the previous frame and needs fewer; a B-frame leans on both past and future frames and needs the fewest. Done well, bit allocation gives clean anchor frames that the rest of the GOP can reference reliably; done poorly, you get visible quality pumping where the picture quality bobs up and down throughout the video.
The typical distribution: out of the bits in a 60-frame GOP, the I-frame alone might use 25–35 %; each P-frame might use 5–8 %; each B-frame might use 1–3 %. The ratios depend on content and rate-control mode. Smart encoders also account for hierarchical referencing — frames that other frames depend on get extra quality budget because their errors propagate. Without this, errors compound through the GOP and quality degrades visibly toward the end.
For a product team, bit allocation is invisible if your encoder is good but glaringly visible if it isn't. Two practical signs of bad bit allocation. Quality pumping: the video appears to "breathe", looking sharp on I-frame moments and slightly soft just before the next I-frame. Visible compression artefacts that worsen mid-GOP: a recompressed clip looks fine at the cut points but gets blocky in the middle. The fix is choosing an encoder with sensible default bit allocation (x264, x265, SVT-AV1 all do) and a sensible crf or capped-CRF setting that doesn't squeeze too hard.
