Diligence & quality
How to inventory an audio or video dataset
Build a file-level inventory before estimating volume, requesting a quote, or preparing a dataset for review. Count known durations and record the technical properties buyers need.
- Topic
- Diligence & quality
- For
- Buyers and data owners
- Reading time
- 5 min
- Updated
Build a file-level inventory before estimating volume, requesting a quote, or preparing a dataset for review. Count known durations and record the technical properties buyers need.
Use the linked free tool to work through the practical steps. Keep the original files and documentation unchanged, and save your report with the date and scope of the review.
Start with an untouched source copy
Choose a representative folder first. Keep original filenames and folder structure so every inventory row can be traced back to its source. If the collection contains multiple exports of the same recording, preserve that distinction; a larger file count does not necessarily mean more unique material.
The free Media Dataset Inspector reads selected files locally. It reports metadata rather than uploading recordings. Its example contains an original two-second synthetic tone and a deliberately unreadable file, so you can see both successful and failed inspection without using private material.
Record the properties that affect delivery
For audio, record codec, sample rate, channel count, duration and file size. For video, add dimensions, frame rate and audio/video track counts. A MOV container does not tell a buyer which video codec is inside. Likewise, stereo audio channels do not prove two independent speakers were captured.
Export the CSV manifest for filtering and the JSON report for a structured record. The inspector reports the first audio and video track in detail and counts additional tracks. If a project depends on all track-level properties, inspect those streams separately before accepting delivery.
Review exceptions before adding up the archive
A duration that cannot be read is unknown, not zero. Keep exception rows and investigate them. Unsupported, encrypted, empty or truncated files can require a different parser or a fresh export. A readable header is not proof that every frame or sample decodes successfully.
The tool reads common formats through MediaInfo. It does not run a full decode, identify duplicate recordings or measure intelligibility. Follow metadata inventory with listening or viewing checks, representative decoding, duplicate review and the buyer’s written acceptance tests.
Share the inventory with the right context
File paths can reveal names, project titles or private folder structure. Review exported paths before sharing the manifest. Keep rights documents separate and link them through controlled identifiers rather than adding sensitive personal records to a public spreadsheet.
For a seller assessment, describe the source, approximate unique volume and permissions status. For a buyer brief, specify required formats and acceptable variation. The inventory makes those conversations concrete; it does not establish licensing authority or commercial value.
Sources and further reading
Try the workflow on your own data
Open free tool →Related resources
Related reading
Open the free tool
Browser tools
Free dataset tools
Inspect local media, document dataset scope or check an evaluation sample.
Related reading
Dataset quality and delivery
5 min read
Recorded hours vs usable dataset hours: how to count media volume
File-hours, unique activity-hours, speech-hours and accepted hours answer different questions. Define the unit before comparing collections or recording offers.
5 min read
How to write a dataset card buyers can evaluate
Document the dataset’s purpose, origin, collection process, limitations and permissions evidence. A useful card makes unknowns visible instead of hiding them behind marketing language.
5 min read
Dataset manifest vs datasheet vs dataset card
A manifest identifies files. A datasheet records structured answers about a collection. A dataset card explains the collection to users. Most serious deliveries benefit from all three.