Old AVI Files: Why They Are Huge and Won’t Play
Ten-year-old camcorder exports, DVR archives, early digital camera clips — all AVI, all refused by phones, all enormous. The problems trace to the format’s age.
What kind of box AVI is
AVI dates to 1992 — nearly a decade before MP4. Its era shows in two structural limits: no modern metadata (no faststart streaming optimization), and a "anything fits" codec policy — DivX, Xvid, MJPEG, even later H.264 can all live inside an AVI shell.
That versatility is exactly what breaks compatibility: a player opening an AVI cannot know which codec the video uses, so it either tries decoders one by one or gives up. Phones and browsers mostly accept "modern container + H.264" only — AVI is excluded by design of the ecosystem.
Why old AVIs are so large: the MJPEG factor
Camcorders, DVRs and early cameras typically wrote MJPEG — every frame stored like an independent JPEG, no inter-frame compression at all. Bitrates routinely exceed 20Mbps: ten minutes can clear 1.5GB while the picture itself stays soft.
That is "huge yet blurry" in one sentence: large because every frame is saved in full (inefficient), blurry because early devices were low-resolution with crude compression. The silver lining: such files have enormous compression headroom — converting to modern H.264 typically leaves 1/5 to 1/10 of the size with essentially identical perceived quality.
Converting to MP4 the right way
AVI to MP4 must re-encode (no stream-copy shell swap like MKV): the legacy codecs are not on MP4’s supported list, and copying them verbatim yields an unplayable file. The standard recipe is H.264 + AAC at around CRF 21 — generous for MJPEG sources, whose inherent softness wastes nothing at higher bitrates.
AVIs already compressed with Xvid/DivX (downloaded legacy media) gain little: they have been lossy-encoded once, and re-encoding maintains or slightly increases size. Quick triage before batch conversion: huge-and-blurry files (MJPEG sources) pay off most; moderate-size legacy-compressed files, barely.