What The Audio Track Adds: 320 Kbps Over A Two Hour Video

The Week Everyone Blamed The Sound

A coach I work with records two hour podcast-style videos: one camera, a title card on screen, and a conversation. His episodes kept landing above 7 GB, uploads kept failing at his studio, and for a week the working theory in every message was the audio. The microphone was good, the recording sounded rich, and rich sound felt like it should weigh something. I believed the theory too, right up until I priced the audio track on its own.

The answer, up front: audio at 320 kbps over a two hour video weighs 288 MB. The episodes weighed 7,488 MB. The sound was under 4 percent of the file, and the silent-looking video track was the rest. This article does the audio maths rate by rate, so you can price any track in seconds, and then shows the setting change that took the coach's episodes from 7.5 GB to about 1 GB without touching his voice.

The Audio Maths, Rate By Rate

Audio obeys the same rule as video: bitrate times duration in seconds, divided by eight. The only surprise is how small the answers are. Each row in the table prices the sound on its own, with no video counted at all.

I produced that table by running each rate through the video file size calculator with the video bitrate set to zero, which isolates the track completely. Three readings from it guide most of my export decisions now. First, even the top rate, 320 kbps, costs 144 MB an hour, which is real money on a phone plan and a rounding error on a video budget. Second, speech does not need the top rate. Podcast guidance for talk content recommends 96 to 128 kbps in mono, and YouTube's upload table lists 128 kbps for mono audio.

A single voice on a single microphone gains nothing audible from 320 kbps stereo except the weight. Third, mono and stereo at the same bitrate weigh exactly the same, because the rate already counts every channel together. Switching a voice track from stereo to mono does not shrink the file at all; lowering the rate is the only lever that does.

Where does 320 kbps come from in the first place? It is the top of the common MP3 range and a factory default in more than one export preset, which is how it lands on talk content that never needed it. A preset cannot tell a voice from an orchestra, so it insures every file at the music rate.

The same table prices shorter episodes with no new maths: a forty-five minute episode at 192 kbps carries 64.8 MB of audio, and at 128 kbps it carries 43.2 MB. Duration and rate, multiplied and divided by eight, settle every version of the question.

The Video Track Standing Next To It

Now the other track. The coach's export preset had carried a 1080p video rate of 8 Mbps over a frame that showed a title card and changed twice an episode. Eight megabits per second for 7,200 seconds, divided by eight, is 7,200 MB of video. Add the 288 MB of audio and the file lands at 7,488 MB, matching what his uploads kept reporting.

I spent a week blaming the smallest part of the file, because it was the part I understood least.

This pattern is not unique to podcast videos. Any file can hide its weight in the track nobody suspects, which is why I read both rates before diagnosing anything. Meeting recordings have the same shape, with a few hundred megabytes of faces riding on tens of megabytes of voice, as the per-mode table in The Zoom And Teams Recording Guide shows. And the launch-week drive failure on this site started the same way, with a team blaming everything except the export rate, as The 1080p Investigation tells it.

Audio does earn its keep in two places, and I price it carefully there. On very long audio-led videos the track stops being trivial: ten hours at 320 kbps is 1,440 MB of sound alone, before a single frame is counted. And on an audio-only release the rate is the whole file. The coach now publishes an audio-only version of each episode at 128 kbps alongside the video, and it weighs 115.2 MB against the video version's 1,015.2 MB. His listeners on mobile data feel that difference every week, even though his drive never could.

The Fix That Actually Shrank The Files

Step 1: price both tracks separately. Read the file's properties for its total, then run the audio rate through the calculator on its own. Whatever is left belongs to the video track.

Step 2: set audio for speech, not for music. The coach now exports at 128 kbps. Over two hours that is 115.2 MB, and no listener has ever written in about the sound.

Step 3: cut the video rate to match the picture. A title card and a still frame do not need 8 Mbps. At 1 Mbps the video track for two hours weighs 900 MB, and the frame looks identical, because there was never 8 Mbps worth of picture in it.

Step 4: check the new total before rendering. 900 MB of video plus 115.2 MB of audio is 1,015.2 MB, about 1.02 GB. The episodes now upload in a fraction of the old time, and the voice, the part his audience actually comes for, is byte for byte the quality it always was.

The use case generalises to any audio-led video: lectures, sermons, audiobook chapters with a cover image, meeting audio published as video. Price the sound first, because it takes ten seconds. Then spend your effort where the megabytes actually live.