efm-stacker-f2

EFM F2 Section Multi-Source Stacker

Overview

efm-stacker-f2 combines multiple F2 section files from different captures of the same source material to produce a higher-quality result through majority-vote error correction. This tool is essential for recovering heavily damaged or degraded EFM data by leveraging multiple independent captures.

Usage

efm-stacker-f2 [options] <input1.f2> <input2.f2> [input3.f2 ...] <output.f2>

Options

• -h, --help - Display help information
• -v, --version - Display version information
• -d, --debug - Show debug output and stacking statistics
• -q, --quiet - Suppress info and warning messages

Arguments

• inputs - Two or more input F2 section files from different captures
• output - Output combined F2 section file

Processing Algorithm

Multi-Source Voting

The stacker examines each F2 frame across all input sources and uses a voting algorithm:

1. Alignment: Automatically aligns F2 sections across sources
2. Comparison: Compare each byte position across all sources
3. Majority Vote: Select the most common value at each position
4. Confidence: Higher confidence with more agreeing sources
5. Error Flags: Preserve error correction flags from individual sources

Quality Improvement

• 2 sources: Can correct errors where sources disagree and one is clearly wrong
• 3+ sources: Majority voting provides robust error correction
• 5+ sources: Excellent correction even with heavily damaged individual sources

Alignment and Synchronization

Automatic Alignment

The tool automatically handles:

• Timing drift: Small timing differences between captures
• Section offset: Different starting positions in the data stream
• Metadata matching: Ensures Q-channel data consistency across sources

Sync Requirements

All input sources must represent the same original material:

• Same disc/recording
• Same time period (overlapping capture windows)
• Compatible EFM format (same sampling parameters)

Pipeline Integration

Input Requirements

• Multiple F2 section files from efm-decoder-f2
• All sources should represent the same original material
• Sources can have different error patterns/locations

Recommended Workflow

Multiple Capture Stacking

# Capture same disc multiple times
ld-decode disc_capture1.ldf disc1 --efm
ld-decode disc_capture2.ldf disc2 --efm  
ld-decode disc_capture3.ldf disc3 --efm

# Process each to F2 sections
efm-decoder-f2 disc1.efm disc1.f2
efm-decoder-f2 disc2.efm disc2.f2
efm-decoder-f2 disc3.efm disc3.f2

# Stack for improved quality
efm-stacker-f2 disc1.f2 disc2.f2 disc3.f2 stacked.f2

# Continue with normal pipeline using Unix pipes
efm-decoder-d24 stacked.f2 - | efm-decoder-audio - final.wav

# Or step by step
efm-decoder-d24 stacked.f2 stacked.d24
efm-decoder-audio stacked.d24 final.wav

Domesday Multi-Source Recovery

# Stack multiple Domesday captures
efm-stacker-f2 domesday1.f2 domesday2.f2 domesday3.f2 domesday4.f2 stacked.f2
efm-decoder-d24 stacked.f2 stacked.d24
efm-decoder-data stacked.d24 domesday.bin --output-metadata
vfs-verifier domesday.bin domesday_metadata.txt

Performance and Quality Metrics

Stacking Effectiveness

• Error reduction: Typically 10-100x reduction in uncorrectable errors
• Processing time: ~2-3x single-source processing time per additional source
• Memory usage: Scales linearly with number of sources
• Quality improvement: Diminishing returns after 5-7 sources

Debug Output

When --debug is enabled, shows:

• Alignment statistics between sources
• Voting confidence levels
• Error correction success rates
• Per-section quality improvements

Quality Assessment

Before Stacking

Check individual F2 files for:

• Different error patterns across sources
• Complementary good/bad sections
• Alignment feasibility

After Stacking

The output quality can be verified by:

• Reduced error counts in downstream tools (efm-decoder-d24)
• Improved audio quality in final output
• Better sector recovery rates for data applications

Technical Notes

Section-Level Processing

• Operates on 98-frame F2 sections (1/75 second boundaries)
• Preserves section metadata and Q-channel information
• Maintains ECMA-130 compliance for downstream processing

Error Correction Integration

• Works in conjunction with CIRC correction (in efm-decoder-d24)
• Complementary to Reed-Solomon correction (for data applications)
• Improves input quality for subsequent error correction stages