State the stored sample format first
Uncompressed PCM audio has a storage relationship based on samples, channels and the bytes used for each sample. Microsoft documents these relationships in WAVEFORMATEX and PCM format guidance. This is different from estimating an AAC or other compressed stream from its encoded bitrate. Do not apply a PCM formula to every file merely because it contains audio.
Check the actual stored format, including whether a nominal sample precision is packed or held in a larger container. A 24-bit value stored in a four-byte slot has a different payload size from a packed three-byte sample. Duration and channel count also need to describe the recording being prepared, not an advertised device maximum.
Work through a hypothetical ten-minute recording
Assume an authorised recording uses 48000 samples per second per channel, packed 24-bit samples and two channels. Each sample value occupies three bytes, so each audio frame across both channels occupies six bytes. The payload rate is 48000 multiplied by six, or 288000 bytes per second. These are explicit assumptions for the calculation, not a quality recommendation.
For ten minutes, multiply by 600 seconds to obtain 172800000 payload bytes. That is 172.8 decimal megabytes, or about 164.8 mebibytes. File headers, metadata and any additional content add overhead. Reserve working space for the recording and intended outputs rather than filling a drive to the last calculated payload byte.
Keep storage arithmetic separate from listening quality
A larger sample rate, word length or channel count increases the calculated PCM payload, but those numbers alone do not establish that an audience will hear a better programme. The recording source, processing and delivery specification matter. Choose the format required by the authorised production workflow instead of maximising every setting solely because storage is available.
If the final deliverable is compressed, use its actual encoding specification or measured output for the delivery budget. The PCM source size is useful for capture and editing storage, while the published stream can have a different rate. Keep both estimates labelled so a producer does not confuse local master storage with a viewer's internet data consumption.
Verify the real file and working copies
After a short capture, inspect the actual format and file size to confirm that the assumptions match the tool's output. A recording application may use floating-point samples, another channel layout or compression. If the measured result differs, identify the format before revising the formula. Do not silently treat the largest discrepancy as unexplained overhead.
Include backups, intermediate renders and export copies in the project plan. One source recording can require several files during editing, even though the sample formula describes one payload. The useful outcome is a transparent storage estimate with declared units and assumptions, checked against an authorised sample, rather than an unsupported promise based on the recording's duration alone.
Questions answered
Can I use this formula for an AAC stream?
No. The worked formula describes explicitly assumed uncompressed PCM storage. Use the compressed encoding's actual bitrate or measured output.
Why can two 24-bit recordings have different storage rates?
The stored sample container, channels, sample rate and other format choices can differ. Check the actual representation, including packing.
Sources & further reading
Official references for the subjects discussed. Service terms and device software can change.