Barcode Quality Mode

Barcode Quality Mode

BarcodeQualityMode in Aspose.BarCode for Java describes how good or bad you expect barcode quality to be in a given scenario:

  • HIGH – barcodes are expected to be high quality (crisp edges, good contrast, low noise).
  • NORMAL – barcodes are expected to be of typical / medium quality (normal scans, slight imperfections).
  • LOW – barcodes are expected to be low quality (noise, blur, poor contrast), so the engine should be more tolerant.

This enum does not directly specify how “powerful” or “expensive” the recognition pipeline is.
Instead, it works as a hint for the recognition engine and is always used together with a QualitySettings preset.

Typical workflow:

  1. Choose a QualitySettings preset that defines the overall recognition strategy (speed vs robustness).
  2. Set BarcodeQualityMode to describe the expected quality of barcodes in your images.
  3. Optionally, add targeted overrides such as XDimensionMode and MinimalXDimension.

All examples in this article are based on the sample class:

com.aspose.barcode.guide.recognition.performance.QualityModeExample

You can find the full source code on GitHub:

QualityModeExample.java

In the snippets below, variables like imagePath represent paths to barcode images in your application.

1. QualitySettings presets vs BarcodeQualityMode

QualitySettings presets define how strong the recognition pipeline should be in general:

  • QualitySettings.getNormalQuality() – default, suitable for most barcodes.
  • QualitySettings.getHighQuality() – preset for low-quality barcodes, enables more robust processing.
  • QualitySettings.getMaxQuality() – maximal robustness, tries to recognize even very difficult or partially incorrect barcodes.
  • QualitySettings.getHighPerformance() – speed-first preset for clean and predictable inputs.

BarcodeQualityMode is applied after selecting a QualitySettings preset and refines how the reader interprets the expected barcode quality level.

  • BarcodeQualityMode.HIGH – engine expects high-quality symbols, can follow the “lightest” path compatible with the preset.
  • BarcodeQualityMode.NORMAL – engine expects medium quality, uses a balanced path within the chosen preset.
  • BarcodeQualityMode.LOW – engine expects low-quality / damaged symbols, uses a more tolerant path within the chosen preset.

In other words:

  • Presets answer the question: “How strong should the overall pipeline be for this application?”
  • BarcodeQualityMode answers the question: “What quality of barcodes do I expect in this image?”

These two knobs are independent and should be tuned together.

2. Using BarcodeQualityMode on clean images

On clean barcodes, you can combine any preset with a suitable BarcodeQualityMode value.
Below are three typical patterns for a high-quality CODE_128 input.

2.1. Clean image, high-quality symbols (HIGH)

Use a fast preset and mark barcodes as high quality:

String imagePath = "code128_clean.png";

BarCodeReader barCodeReader = new BarCodeReader(imagePath, DecodeType.CODE_128);

QualitySettings qualitySettings = QualitySettings.getHighPerformance();
qualitySettings.setBarcodeQuality(BarcodeQualityMode.HIGH);
barCodeReader.setQualitySettings(qualitySettings);

BarCodeResult[] barCodeResults = barCodeReader.readBarCodes();
if (barCodeResults.length > 0) {
    BarCodeResult barCodeResult = barCodeResults[0];
    System.out.println("HIGH (clean): " + barCodeResult.getCodeText());
}

Here you explicitly say to the engine:

  • “use a performance preset”, and
  • “current barcodes are high quality”.

This pattern is appropriate when:

  • barcodes are generated programmatically or printed with predictable good quality,
  • scanning is done in a controlled environment,
  • and you want maximum throughput.

2.2. Clean image, typical quality (NORMAL)

Use a balanced preset and mark barcodes as normal quality:

String imagePath = "code128_clean.png";

BarCodeReader barCodeReader = new BarCodeReader(imagePath, DecodeType.CODE_128);

QualitySettings qualitySettings = QualitySettings.getNormalQuality();
qualitySettings.setBarcodeQuality(BarcodeQualityMode.NORMAL);
barCodeReader.setQualitySettings(qualitySettings);

BarCodeResult[] barCodeResults = barCodeReader.readBarCodes();
if (barCodeResults.length > 0) {
    BarCodeResult barCodeResult = barCodeResults[0];
    System.out.println("NORMAL (clean): " + barCodeResult.getCodeText());
}

Typical use cases:

  • back-office systems processing documents scanned under normal conditions,
  • services where speed and robustness are equally important,
  • systems with stable but not strictly controlled image quality.

2.3. Clean image, engine tuned for low-quality symbols (LOW)

You can still use a “robust” preset and explicitly say that barcodes might be low quality:

String imagePath = "code128_clean.png";

BarCodeReader barCodeReader = new BarCodeReader(imagePath, DecodeType.CODE_128);

QualitySettings qualitySettings = QualitySettings.getHighQuality(); // preset for low-quality barcodes
qualitySettings.setBarcodeQuality(BarcodeQualityMode.LOW);          // tell engine that barcodes may be low quality
barCodeReader.setQualitySettings(qualitySettings);

BarCodeResult[] barCodeResults = barCodeReader.readBarCodes();
if (barCodeResults.length > 0) {
    BarCodeResult barCodeResult = barCodeResults[0];
    System.out.println("LOW (clean): " + barCodeResult.getCodeText());
}

This combination is useful when:

  • your input pipeline usually contains difficult images (noise, blur, weak contrast),
  • a particular subset of images is clean, but you still want the engine tuned for low-quality barcodes,
  • you prefer to keep the same configuration across different input sources.

3. Working with noisy images

With noisy or degraded images, BarcodeQualityMode becomes more important.
Higher presets (HighQuality, MaxQuality) already target difficult barcodes.
BarcodeQualityMode tells the engine how bad the current data is expected to be.

3.1. Noisy image with a speed-first preset (HIGH)

In some cases you may still treat a noisy image as “high-quality” if the noise level is controlled and barcodes remain legible:

String imagePath = "code128_noisy.png";

BarCodeReader barCodeReader = new BarCodeReader(imagePath, DecodeType.CODE_128);

QualitySettings qualitySettings = QualitySettings.getHighPerformance();
qualitySettings.setBarcodeQuality(BarcodeQualityMode.HIGH);
barCodeReader.setQualitySettings(qualitySettings);

BarCodeResult[] barCodeResults = barCodeReader.readBarCodes();
System.out.println("Noisy + HighPerformance + HIGH mode: results = " + barCodeResults.length);

Use this configuration when:

  • noise is moderate and barcodes are still close to high quality,
  • you prefer to keep latency as low as possible,
  • and some read failures are acceptable.

3.2. Noisy image with a robust preset (LOW)

If the image is clearly degraded (strong noise, weak contrast), it is better to use a robust preset and explicitly mark symbols as low quality:

String imagePath = "code128_noisy.png";

BarCodeReader barCodeReader = new BarCodeReader(imagePath, DecodeType.CODE_128);

QualitySettings qualitySettings = QualitySettings.getHighQuality(); // preset for low-quality barcodes
qualitySettings.setBarcodeQuality(BarcodeQualityMode.LOW);          // engine expects low-quality symbols
barCodeReader.setQualitySettings(qualitySettings);

BarCodeResult[] barCodeResults = barCodeReader.readBarCodes();
System.out.println("Noisy + HighQuality + LOW mode: results = " + barCodeResults.length);

This pattern is appropriate when:

  • images come from mobile cameras or cheap scanners,
  • you often see motion blur, compression artifacts, or strong noise,
  • and throughput is less important than maximum recognition probability.

4. Combining presets, BarcodeQualityMode, and X-dimension hints

The real power of QualitySettings comes from combining:

  • a preset (HighPerformance, NormalQuality, HighQuality, MaxQuality),
  • a BarcodeQualityMode value (HIGH, NORMAL, LOW),
  • additional overrides such as XDimensionMode and MinimalXDimension.

Example: clean CODE_128 with very small modules and a speed-first preset.

String imagePath = "code128_clean.png";

BarCodeReader barCodeReader = new BarCodeReader(imagePath, DecodeType.CODE_128);

QualitySettings qualitySettings = QualitySettings.getHighPerformance();
qualitySettings.setXDimension(XDimensionMode.USE_MINIMAL_X_DIMENSION);
qualitySettings.setMinimalXDimension(1.0f);
qualitySettings.setBarcodeQuality(BarcodeQualityMode.LOW);
barCodeReader.setQualitySettings(qualitySettings);

BarCodeResult[] barCodeResults = barCodeReader.readBarCodes();
if (barCodeResults.length > 0) {
    BarCodeResult barCodeResult = barCodeResults[0];
    System.out.println("Preset + overrides: " + barCodeResult.getCodeText());
}

Here:

  • getHighPerformance() provides a fast baseline,
  • setXDimension(XDimensionMode.USE_MINIMAL_X_DIMENSION) and setMinimalXDimension(1.0f) tell the engine that bar widths may be very small,
  • BarcodeQualityMode.LOW says that barcodes may be low quality (weak / noisy), so the recognition path should be tolerant even though the preset is speed-oriented.

Use this approach when:

  • you know something about expected module size (print resolution, target device, etc.),
  • you need to support tiny barcodes rendered at the edge of resolution,
  • you want to stay close to performance presets but still help the engine with geometry hints.

5. Practical recommendations

When configuring BarcodeQualityMode and QualitySettings in a real system:

  • Start from a preset that matches your scenario:
    • getNormalQuality() for general-purpose applications.
    • getHighPerformance() for clean, predictable barcodes and maximum throughput.
    • getHighQuality() when barcodes are often low quality or damaged.
    • getMaxQuality() when you need the most robust behavior, even for difficult or partially incorrect symbols.
  • Set BarcodeQualityMode according to expected quality of current inputs:
    • HIGH – clean synthetic data, good printers, controlled environment.
    • NORMAL – ordinary scans, moderate imperfections.
    • LOW – noisy/blurred barcodes, poor lighting, low-contrast prints.
  • For noisy or degraded images, prefer combinations like:
    • HighQuality + LOW,
    • MaxQuality + LOW.
  • For clean high-throughput scenarios, start from:
    • HighPerformance + HIGH,
    • or NormalQuality + NORMAL for a balanced profile.
  • Combine presets with targeted overrides (XDimensionMode, MinimalXDimension, etc.) when you know details about symbol geometry or device resolution.

Notes about the example class

The example class QualityModeExample generates its own test images:

  • code128_clean.png – a clean synthetic CODE_128 image,
  • code128_noisy.png – the same image with Gaussian noise added by ExampleAssist.addGaussianNoise(...).

These fixtures help demonstrate how different combinations of presets and BarcodeQualityMode behave on clean vs noisy data.
In your own projects you can reuse the patterns but work with your real images instead of generated ones.

Full class source:

QualityModeExample.java

Minimal XDimension Deconvolution Mode