SubtitleEdit/subtitleedit · error · InvalidOperationException

Tesseract exited with code {proc.ExitCode}: {err}

Error message

Tesseract exited with code {proc.ExitCode}: {err}

What it means

Thrown after the Tesseract process ran but exited with a non-zero code. The captured stderr (drained concurrently to avoid pipe deadlock) is appended to the message, so it usually names the real cause (missing language data, corrupt image, bad --psm, model errors).

Source

Thrown at src/seconv/Core/TesseractOcrEngine.cs:100

                ArgumentList = { pngPath, "stdout", "-l", Language, "--psm", "6" },
                RedirectStandardOutput = true,
                RedirectStandardError = true,
                StandardOutputEncoding = System.Text.Encoding.UTF8,
                StandardErrorEncoding = System.Text.Encoding.UTF8,
                UseShellExecute = false,
                CreateNoWindow = true,
            };
            using var proc = Process.Start(psi)
                ?? throw new InvalidOperationException("Failed to start tesseract process.");
            // Drain stderr concurrently — Tesseract emits warnings to stderr, and reading
            // stdout to completion while stderr fills the pipe buffer would deadlock.
            var stderrTask = proc.StandardError.ReadToEndAsync();
            var stdout = proc.StandardOutput.ReadToEnd();
            proc.WaitForExit();
            if (proc.ExitCode != 0)
            {
                var err = stderrTask.GetAwaiter().GetResult();
                throw new InvalidOperationException($"Tesseract exited with code {proc.ExitCode}: {err}");
            }
            return stdout.Trim();
        }
        finally
        {
            try { File.Delete(pngPath); } catch { /* best-effort */ }
            prepped.Dispose();
        }
    }

    private static SKBitmap Preprocess(SKBitmap source)
    {
        const int scale = 2;
        var w = source.Width * scale;
        var h = source.Height * scale;
        var prepped = new SKBitmap(new SKImageInfo(w, h, SKColorType.Rgba8888, SKAlphaType.Opaque));
        using var canvas = new SKCanvas(prepped);
        canvas.Clear(SKColors.White);

View on GitHub (pinned to 17a9f07487)

Solutions

1. Install the requested language pack (apt install tesseract-ocr-fra / brew install tesseract-lang, or drop the .traineddata into the tessdata folder).
2. Set TESSDATA_PREFIX to the directory holding the .traineddata files and retry.
3. Read the stderr in the message — it states the exact Tesseract error.
4. Verify the Tesseract version matches the model format.

Example fix

// before: -l fra but only eng installed
engine = TesseractOcrEngine.Create("fra");
// after: install pack + keep language
$ apt install tesseract-ocr-fra
engine = TesseractOcrEngine.Create("fra");

// Verify the requested language's traineddata exists
var tessdata = Environment.GetEnvironmentVariable("TESSDATA_PREFIX") ?? Path.Combine(Path.GetDirectoryName(ExecutablePath) ?? ".", "tessdata");
if (!File.Exists(Path.Combine(tessdata, Language + ".traineddata")))
    throw new InvalidOperationException($"Missing {Language}.traineddata in {tessdata}");

null

try { return engine.Recognize(bmp); }
catch (InvalidOperationException ex) when (ex.Message.StartsWith("Tesseract exited")) {
    logger.Error("Tesseract stderr: " + ex.Message);
    return string.Empty;
}

When it happens

Trigger: proc.ExitCode != 0 after Process.Start of tesseract with arguments [pngPath, stdout, -l Language, --psm 6]. stderr is awaited and folded into the message.

Common situations: Requested a language whose .traineddata is not installed (e.g. -l fra with only eng present); Tesseract version mismatch with the model; malformed/expired PNG written to the temp work dir; insufficient tessdata dir / TESSDATA_PREFIX unset.

Related errors

• Tesseract not found on PATH. Install it from https://tessera
• Failed to start tesseract process.
• No image handler for format '{options.Format}'
• API key invalid (or perhaps billing/API is not enabled)?
• "Perhaps billing is not enabled (or API not enabled or API k

AI-assisted analysis of SubtitleEdit/subtitleedit@17a9f07487 (2026-08-13). Data as JSON: /api/errors/908f44a940bb23b8. Report an issue: GitHub.