CAVS to OPUS Converter

CAVS to Opus: What This Page Covers

Opus is the right destination for almost any new audio export — open, royalty-free, and efficient from low-bitrate speech up to full-range music. The catch on this particular page is the source: a genuine .cavs file is a raw Chinese AVS (AVS1) video elementary stream and contains no audio track at all, so there is normally nothing for the encoder to work with. This walk-through shows how the extraction is set up, how to tell whether your file actually holds sound, and exactly which file to convert instead when it does not.

How to Convert CAVS to Opus

1. Upload Your CAVS File: Drag and drop your .cavs onto the page, or click "Add Files" to browse for it. Several files can be queued and run with the same settings.
2. Pick a Compression Mode: The Audio Codec is fixed to Opus. Under File Compression choose Quality Preset, or take manual control with Custom Bitrate (128 Kbps prefilled), Constant Bitrate, Variable Bitrate, or a target file size.
3. Set Audio Channel, Sample Rate or Trim: Audio Channel offers Original, Mono or Stereo, Audio Sample Rate offers Original plus fixed rates, and Trim takes a start time and a duration if you only want one segment.
4. Convert and Download: Click "Convert" and save the .opus file. No sign-up, no watermark.

Walk-through: Choosing the Opus Rate

Opus is standardised by the IETF as RFC 6716, published September 2012. It fuses two engines — a linear-prediction layer derived from SILK for speech and an MDCT layer derived from CELT for music — and, in the specification's own words, "scales from low bitrate narrowband speech at 6 kbit/s to very high quality stereo music at 510 kbit/s". That range is why one codec covers material this different:

Two practical notes. Opus runs at 48 kHz internally, so leave Audio Sample Rate on Original unless you have a specific reason to force one — 44100 Hz in particular is not an Opus rate, and a request for it is snapped up to 48 kHz. And every setting above assumes there is an audio stream to encode; none of them can conjure one out of a video-only file.

Where the Audio Actually Lives

Chinese AVS material is normally multiplexed — coded video and a separate coded audio track interleaved inside one container. A .cavs is what is left when a tool demuxes that container and keeps only the video half. So the file you should point at the converter depends on what you actually have:

Common Errors and How to Fix Them

• "The conversion failed and I got no file" — The expected outcome for a true video-only .cavs. With no audio stream present the encoder has no input to write, so the job errors out rather than handing you a silent placeholder file. Convert the container the video was demuxed from.
• "I picked 44100 Hz and the file came back at 48 kHz" — Expected. The Opus encoder has no 44.1 kHz mode, so the request is snapped to the nearest rate it does support rather than failing the job. Leave Audio Sample Rate on Original, or pick 48000 Hz to make that explicit.
• "I wanted the video, not the audio" — Opus is an audio-only format and discards the picture entirely. To keep the moving footage in a broadly playable file, use CAVS to MP4.
• "My player will not open the .cavs to check it" — Normal. A bare elementary stream has no container header or index, so most players refuse it; that is not evidence of corruption, and it is not evidence of missing audio either.
• "The file is binary garbage in a text editor" — Good sign: that means genuine coded AVS video rather than an AviSynth .avs script. A file that opens as readable text is a script and holds no media at all.

When This Doesn't Work

If the footage was delivered to you as nothing but the video elementary stream, there is no soundtrack to recover — not with different settings, not with a different tool, and not with a different audio format. The data does not exist in the file. The only real fix is to go back a step in the chain to the .ts or .mp4 the stream was pulled out of, where the audio track is still sitting next to the video. If the original recording genuinely never had sound — silent surveillance capture, a test pattern, an encoder output written without an audio input — then no conversion path produces audio, and what you want is a video target such as CAVS to MP4 instead.

Frequently Asked Questions

Why does converting CAVS to Opus fail instead of giving me a silent file?

Because there is no audio stream to encode. The extraction step selects the audio from the input and drops the video; when the input contains only video, the encoder has nothing to write and the job stops with an error instead of producing an empty container. That is more useful than a silent file, which would look like a successful conversion and waste your time later.

My AVS recording definitely had sound — where is it now?

Still in the original container. Broadcast and optical-media AVS content interleaves the coded video with a separate coded audio track inside an MPEG transport stream or an MP4. Demuxing that file down to a bare .cavs keeps the video elementary stream and leaves the audio behind in the source. Point the converter at that original file (TS to Opus or MP4 to Opus) and the audio is read and encoded normally.

How can I tell whether my file has audio before converting?

Two quick checks. First, the extension: a file that a demuxer produced as .cavs is video-only by definition. Second, the size and origin: if the file came straight off a broadcast capture or a disc rip and is still a .ts, .mpg or .mp4, the audio is there. If someone handed you the "video track" from a project, it is not. Note that a container merely renamed to .cavs is detected by its actual contents, so it still converts normally.

Is Opus the right target here, or should I use MP3?

Opus is the stronger codec by a wide margin at any given bitrate, and it is open and royalty-free, so it is the better default for anything modern — messaging apps, browsers, and streaming platforms all handle it. MP3 is worth choosing only for compatibility with old hardware players that predate Opus. Either way the source caveat is identical: CAVS to MP3 also needs an audio track to exist in the file you upload.

What sample rate will the Opus file be?

48 kHz. Opus operates natively at 48 kHz, and in our testing selecting 44100 Hz caused the encoder to refuse the job rather than resample quietly — it reports 48000 Hz as the rate it accepts. That is why leaving Audio Sample Rate on Original is the sensible default. Anything lower that you select is applied as a genuine resample, which reduces bandwidth for speech but is rarely worth doing by hand.

Does Mono halve the file size?

Roughly, for constant-bitrate encoding — but that is not how Opus is usually best used. Opus applies joint stereo coding, so a stereo signal with little separation already costs far less than two independent channels. Setting Audio Channel to Mono is worth doing for genuinely single-source material such as a lecture or an interview; for music it discards the stereo image for a smaller saving than you would expect.

Could my file actually be an AviSynth .avs script instead?

It is worth ruling out, since the extensions look similar. An AviSynth .avs file is a short text script that tells a frameserver how to process other clips — it holds no encoded audio or video whatsoever. Open it in a text editor: readable instructions mean AviSynth, and binary data means a real Chinese AVS video stream, which is still video-only.

What is CAVS, and does the standard include audio at all?

CAVS is first-generation Chinese AVS video: standard GB/T 20090.2-2006, promulgated in February 2006 by the AVS Workgroup (founded June 2002) and later issued by the IEEE as IEEE 1857-2013. The AVS family does specify audio coding, but in separate parts of the standard series — the video part that a .cavs file contains has no audio inside it by design.

How are my files handled, and are they kept private?

Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically a few hours after the conversion finishes — never shared and never made public. There is no sign-up and no watermark.