Home · Guides · updated 3 September 2026

Voice notes, audio and transcripts

A voice note cannot be printed, cannot be skimmed, and is the one part of a conversation where a person is directly identifiable — which is exactly why the identification has to be proved rather than assumed.

What comes out of the export

Ask for media and the archive contains the voice notes as files, ordinarily .opus, named in the WhatsApp convention and referenced from the transcript at the point they were sent. Without media the transcript records only that audio was sent at that moment — audio omitted — and the recording is not in the record at all. See what “media omitted” means.

Hash the file, not only the archive

The hash in Part A fixes the whole export, which is necessary and coarse: it cannot identify one recording. A voice note that will be played in court, put to a witness, or transcribed, needs its own hash so that the file played is demonstrably the file certified. The mechanics are the same as for any attachment — see photographs, voice notes and documents.

The transcript is not the recording

A transcript of a voice note is a document about a document. It is useful, often necessary, and it is somebody’s account of what they heard. Treat it as that:

• Name the person who prepared it, and how. A transcript that appears from nowhere is a transcript nobody can be cross-examined on.
• Keep it verbatim. Hesitations, interruptions and unclear passages marked as unclear, rather than smoothed into readable sentences.
• Number the passages against the entry number of the note in the transcript of the conversation, so a line can be found in the audio.
• Produce the audio with it. A dispute about what was said is resolved by listening, not by comparing two transcripts.

Translation

Most voice notes in Indian proceedings are not in English, and the court works in the language of the court. That makes two documents, not one: a transcription in the language actually spoken, and a translation of it. Merging them into a single English text hides the step where the disagreement will happen.

Identify the translator, state the language and the dialect where it matters, and keep the original transcription on the file. Where a word or an idiom is doing real work in the case, that is precisely the word the other side will want to hear for themselves.

Identifying the speaker

The account tells you which handset the note was sent from. It does not tell you whose voice is on it, and in a family or a business those are frequently different questions.

• Recognition evidence comes from a person who knows the voice and can say so on oath, and who can be asked how well and from where.
• A forwarded note was recorded by somebody else. Forwarding is common, and the transcript does not always make it obvious.
• Content can corroborate: things said in the recording that only one person knew, or that fit conduct proved elsewhere.
• Comparison of voices by an expert is a separate exercise with its own requirements, and it is not something a certificate or a conversion tool supplies.

The general problem of tying an account to a person is the same one that arises across a conversation, and it is set out in group chats, attribution and third parties.

Duration, and the part that is not said

State the duration of each note and produce it whole. An excerpt is an editorial decision, and a recording produced in fragments raises the same objection as a conversation produced in screenshots: a reader cannot tell what fell between them. If only part is relied on, cite the timings within the complete file rather than filing a trimmed one.

Delivering audio to the court

• Produce the original file as it came out of the export, with its filename and hash.
• Where the court needs a more portable format, deliver a converted copy in addition, labelled as a conversion, with its own hash stated and the original still on the file.
• Tabulate every audio file in the exhibit — entry number, filename, duration, size, hash — so the set can be checked without playing it.
• If the exhibit says the files are delivered with it, deliver them. A document that says so and does not is a false statement on its own face.

Section63 lists every voice note in the exhibit with its entry number, duration, size and its own SHA-256, and delivers the audio alongside the document rather than pretending a recording can be printed.

---

This guide explains procedure and states the law as we understand it. It is not legal advice, and Aarohan Enterprises is not a law firm. Whether a court admits a particular record, and what weight it gives it, is for that court to decide. Have an advocate settle anything you intend to file.