tonhowtf/omniget · error

não consegui medir a duração

Error message

não consegui medir a duração

What it means

cut_one probes the input with ffprobe and requires a positive duration before doing silence detection. When probe.duration_seconds is <= 0 (missing, unparseable, or zero), it throws 'não consegui medir a duração' (could not measure the duration). Without a reliable duration, keep-range math for silence cutting cannot work.

Solutions

1. Verify the input file plays in a standard player and probe it manually with ffprobe to inspect duration.
2. Remux/repair the file (e.g. ffmpeg -i in.mp4 -c copy fixed.mp4) to regenerate container metadata.
3. Only pass complete, finalized media files to the silence-cut API.
4. Check that the path/URL is correct and fully accessible.

Example fix

// before
run(&opts, "partial-download.mp4").await?;
// after
ffmpeg_repair("partial-download.mp4", "fixed.mp4").await?;
run(&opts, "fixed.mp4").await?;

// Rust
let probe = ffmpeg_probe(path).await?;
anyhow::ensure!(probe.duration_seconds > 0.0, "arquivo sem duracao valida");

fn has_duration(p: &Probe) -> bool { p.duration_seconds > 0.0 }

match cut_one(&ffmpeg, &opts, input).await {
    Err(e) if e.to_string().contains("nao consegui medir") => {
        // remuxar/reparar o arquivo e tentar de novo
    }
    r => r?,
}

When it happens

Trigger: Calling cut_one/run on a file whose probe reports no usable duration: corrupt/unfinalized files, raw streams without container headers, network URLs that ffprobe cannot fully read, or a container ffprobe cannot parse.

Common situations: Feeding a partially downloaded MP4 (moov atom missing); passing an image or unsupported format; interrupted recording; probing a live/stream URL without duration metadata.

Related errors

• {}
• Failed to run ffmpeg
• ffmpeg returned code
• Failed to run ffprobe
• ffprobe failed

AI-assisted analysis of tonhowtf/omniget@8600b91f42 (2026-09-12). Data as JSON: /api/errors/7511df1e2acb66bf. Report an issue: GitHub.

Appendix: source

Thrown at src-tauri/omniget-core/src/core/tools/silence_cut.rs:224

                threshold_db,
                min_silence.max(0.05)
            ),
            "-f",
            "null",
            "-",
        ])
        .output()
        .await
        .map_err(|e| anyhow!("ffmpeg nao iniciou: {}", e))?;
    Ok(String::from_utf8_lossy(&out.stderr).to_string())
}

async fn cut_one(ffmpeg: &Path, opts: &SilenceOptions, input: &str) -> anyhow::Result<SilenceItem> {
    let inp = Path::new(input);
    let probe = crate::core::ffmpeg::probe(inp).await?;
    let total = probe.duration_seconds;
    if total <= 0.0 {
        return Err(anyhow!("não consegui medir a duração"));
    }
    let has_audio = probe.streams.iter().any(|s| s.codec_type == "audio");
    if !has_audio {
        return Err(anyhow!("o arquivo não tem trilha de áudio"));
    }

    let stderr = detect(ffmpeg, inp, opts.threshold_db, opts.min_silence).await?;
    let spans = parse_spans(&stderr, total);
    let (spans, skipped) = cap_spans(spans, MAX_SEGMENTS);
    let keeps = keep_ranges(&spans, total, opts.padding.max(0.0));
    let kept: f64 = keeps.iter().map(|k| k.duration()).sum();

    let mut item = SilenceItem {
        input: input.to_string(),
        output: None,
        duration_before: total,
        duration_after: kept,
        seconds_removed: (total - kept).max(0.0),

View on GitHub (pinned to 8600b91f42)