Skip to content

Getting Started

Jump to bottom
Wiki Deployment edited this page Aug 3, 2025 · 3 revisions

Getting Started with Qrew

This guide will help you get up and running with Qrew quickly.

Prerequisites

Before using Qrew, ensure you have:

  1. VLC Media Player installed

  2. REW (Room EQ Wizard) with API license

  3. Audio Interface properly configured

    • Multi-channel output capability
    • Calibrated measurement microphone

First Run

1. Start REW

  • Launch REW first
  • Go to Preferences → API
  • Enable "Allow API Access"
  • Note the API port (default: 4735)

Important: Configure Analysis Preferences for Accurate SNR/SDR

For Qrew to calculate Signal-to-Noise Ratio (SNR) and Signal-to-Distortion Ratio (SDR) values that match REW's internal calculations, you must configure REW's impulse response handling:

  1. Go to Preferences → Analysis
  2. In the "Impulse response calculation" section:
    • Set "Keep full IR" (retain entire impulse response)
    • This prevents REW from truncating the impulse response data

Why this matters:

  • REW normally truncates impulse responses to ~1.7-2 seconds after the peak to save storage space
  • SNR and SDR calculations require analysis of the noise floor in regions before and after the main impulse
  • Truncated responses don't provide enough pre-peak data for accurate noise floor analysis
  • "Keep full IR" preserves the complete impulse response, enabling proper SNR/SDR calculation

Note: This will create larger measurement files and slower processing, but is essential for Qrew's quality scoring system to work correctly.

2. Launch Qrew

Installation Options:

From GitHub Releases (Recommended):

# Download the wheel file from GitHub releases
# Example: qrew-1.0.0rc1-py3-none-any.whl
pip install qrew-1.0.0rc1-py3-none-any.whl

From PyPI (when available):

pip install qrew

Running Qrew:

qrew
  • Qrew will automatically detect REW

3. Load Stimulus File

Before starting measurements, you need to prepare sweep files for each channel you want to measure.

Sweep File Nomenclature

Qrew uses a specific naming convention to automatically detect which sweep file corresponds to each speaker channel:

Format: [prefix]-[description]-[CHANNEL].[extension]

Supported Extensions: .mp4, .mlp, .wav

Examples:

  • sweep-log-FL.mp4 (Left channel)
  • measurement-sine-FR.mp4 (Right channel)
  • test-chirp-C.mlp (Center channel)
  • audio-sweep-SLA.wav (Left Surround)
  • my-test-signal-LFE.mp4 (Subwoofer)

Channel Naming Conventions

Common channel identifiers that Qrew recognizes:

Channel Code Example Filename
Left FL sweep-FL.mp4
Right FR sweep-FR.mp4
Center C sweep-C.mp4
Left Surround SLA sweep-SLA.mp4
Right Surround SRA sweep-SRA.mp4
Left Back SBL sweep-SBL.mp4
Right Back SBR sweep-SBR.mp4
LFE/Subwoofer LFE or SW1 sweep-LFE.mp4
Top Front Left TFL sweep-TFL.mp4
Top Front Right TFR sweep-TFR.mp4
Top Rear Left TRL sweep-TRL.mp4
Top Rear Right TRR sweep-TRR.mp4

How It Works

  1. Click "Load Stimulus File" in Qrew
  2. Qrew automatically scans the same directory for all sweep files
  3. Channel detection: Uses regex pattern matching to identify the channel from the filename
  4. Validation: Only files with recognized channel names will be available for measurement

Requirements

  • Channel identifier must be immediately before the file extension

    • ✅ Good: my-sweep-FL.mp4
    • ✅ Good: test-signal-SLR.mlv
    • ❌ Bad: L-sweep-test.mp4 (channel not at end)
    • ❌ Bad: sweep-L-extra.mp4 (text after channel)

  • All sweep files should be in the same directory

  • Use consistent file extensions (.mp4, .mlp, or .wav)

  • Ensure you have sweep files for all channels you plan to measure

Tips

  • Organize by system: Keep sweep files for each speaker system in separate folders
  • Use descriptive prefixes: log-sweep-FL.mp4 vs sine-sweep-FL.mp4
  • Check before measuring: Verify all needed channels appear in Qrew's channel selection
  • Consistent levels: Ensure all sweep files have the same output level

4. Configure Audio

  • Click the gear icon for Settings
  • Verify VLC backend is selected, show gui or not, can use auto(Qrew detects whats installed), libvlc backend(minimal gui/non-gui player), or subprocess(VLC app gui/non-gui player)
  • Configure your speaker layout, or manually select speakers you will measure

4. Select Channels

  • Choose which speaker channels to measure
  • Common configurations:
    • Stereo: FL, FR
    • 5.1: FL, FR, C, SLA, SRA, SW1
    • 7.1.4: Add SBL, SBR, TFL, TFR, TRL, TRR

5. Start Measurement

  • Click "Start Measurement" to begin
  • Qrew will guide you through each position
  • Move the microphone when prompted

Measurement Workflow

  1. Position Dialog: Tells you where to place the microphone
  2. Automatic Sweep: Plays test signal through each speaker
  3. Quality Check: Evaluates measurement quality
  4. Next Position: Guides you to the next location

Post-Measurement Processing

After completing your measurements, Qrew offers powerful post-processing options that leverage REW's advanced analysis capabilities to improve your data quality and prepare measurements for further analysis.

Cross Correlation Alignment

What it does: Cross correlation alignment time-aligns measurements from different microphone positions to compensate for small timing differences that can occur during measurement.

How it works:

  • Qrew automatically uses Position 0 as the reference measurement
  • All other positions (1, 2, 3, etc.) are time-aligned to match Position 0
  • REW performs the cross-correlation analysis to find the optimal time offset
  • Each channel is processed independently

When to use:

  • After measuring multiple microphone positions
  • When you notice timing inconsistencies between positions
  • Before vector averaging to ensure proper phase alignment

Process:

  1. Select the channels you want to align
  2. Choose "Cross Correlation Only" or "Cross Corr + Vector Average"
  3. Qrew processes each channel separately, aligning all positions to Position 0
  4. Results are saved as new aligned measurements in REW

Vector Averaging

What it does: Vector averaging combines measurements from multiple positions into a single averaged response while preserving phase information.

How it works:

  • Takes aligned measurements from all positions for a channel
  • Performs complex averaging (magnitude + phase) rather than simple magnitude averaging
  • Creates a single representative measurement that reduces the effects of room modes and standing waves
  • Maintains phase coherence for accurate impulse response data

Benefits:

  • Reduces measurement noise and room artifacts
  • Creates smoother, more representative frequency responses
  • Provides better data for downstream processing tools
  • Maintains phase information for time-domain analysis

Process:

  1. Ensure measurements are cross-correlation aligned first (recommended)
  2. Select channels for vector averaging
  3. Choose "Vector Average Only" or "Cross Corr + Vector Average"
  4. Qrew creates new vector-averaged measurements in REW

Combined Processing (Cross Corr + Vector Average)

Recommended workflow: This combines both processes for optimal results:

  1. Cross Correlation Alignment: Time-aligns all positions to Position 0
  2. Vector Averaging: Creates averaged responses from aligned measurements

Process flow:

  • Qrew processes each selected channel individually
  • For each channel: Position 0, 1, 2, 3... → Cross Corr Align → Vector Average
  • Results in one high-quality averaged measurement per channel

Using Results with External Tools

The processed measurements are ideal for use with professional analysis tools:

OCA (ObsessiveCompulsiveAudiophile/ A1EvoAcoustica) - GitHub:

  • Import vector-averaged responses for detailed analysis
  • Use aligned measurements for multichannel optimization
  • Audyssey-based Sound Optimization Tool for Denon/Marantz AVRs

REW Advanced Analysis:

  • Use processed measurements for EQ filter design
  • Room correction with higher confidence
  • Improved visualization and analysis

Technical Details

Channel Selection: Qrew automatically identifies which measurements belong to each channel and position based on the naming convention (e.g., FL_pos0, FL_pos1, FL_pos2).

Position Sorting: Measurements are automatically sorted by position number and measurement uuid, ensuring Position 0 is always used as the reference for alignment.

Quality Preservation: Both cross correlation and vector averaging preserve the full frequency and phase information from your original measurements.

Processing Order: When using combined mode, cross correlation alignment is always performed first, followed by vector averaging of the aligned results.

Understanding Measurement Quality Scoring

Qrew automatically evaluates each measurement using a comprehensive 100-point scoring system that combines multiple acoustic metrics. This helps you identify which measurements are reliable for analysis and which may need to be retaken.

Quality Ratings

  • PASS (≥ 70 points): Measurement meets professional standards and is suitable for analysis
  • CAUTION (50-69 points): Usable measurement but may need verification or careful consideration
  • RETAKE (< 50 points): Measurement quality insufficient for reliable analysis

Scoring Components

The 100-point score is built from eight key metrics, each weighted according to its importance for loudspeaker analysis:

1. Signal-to-Noise Ratio (20 points max)

  • Range: 20-75 dB
  • Purpose: Ensures measurements aren't noise-limited
  • Good: ≥ 60 dB for most analysis
  • Excellent: ≥ 70 dB for critical work

2. Signal-to-Distortion Ratio (15 points max)

  • Range: 20-55 dB
  • Purpose: Indicates clean signal path and low artifacts
  • Good: ≥ 40 dB for general analysis
  • Excellent: ≥ 50 dB for precision work

3. Mean THD - Broadband (15 points max)

  • Range: 0-2% THD
  • Purpose: Primary distortion metric for loudspeaker linearity
  • Good: ≤ 1% for most speakers
  • Excellent: ≤ 0.5% for high-end drivers

4. Peak THD Spike (10 points max)

  • Range: 0-10% THD
  • Purpose: Identifies resonances, breakup modes, or clipping
  • Good: ≤ 3% narrow-band spikes
  • Watch for: Spikes > 5% indicating driver issues

5. Low-Frequency THD (5 points max)

  • Range: 0-15% THD (below 200 Hz)
  • Purpose: Bass drivers typically have higher distortion
  • Good: ≤ 5% for subwoofers, ≤ 3% for woofers
  • Acceptable: Up to 8% for very low frequencies

6. Harmonic Ratio H3/H2 (5 points max)

  • Range: 0-1.0 ratio
  • Purpose: Odd harmonics (H3) are more audible than even (H2)
  • Good: H3/H2 ≤ 0.5
  • Based on: IEC 60268-21 standards

7. Magnitude-Squared Coherence (15 points max)

  • Range: 0.90-0.99
  • Purpose: Indicates measurement repeatability and system linearity
  • Good: ≥ 0.95 for most analysis
  • Excellent: ≥ 0.98 for precision work

8. Impulse Response Peak-to-Noise (15 points max)

  • Range: 35-55 dB
  • Purpose: Time-domain quality independent of frequency analysis
  • Good: ≥ 45 dB for clean impulse response
  • Excellent: ≥ 50 dB for time-domain analysis

Technical Implementation

Scoring Method: Each metric uses linear scaling within its specified range. For distortion metrics (THD, harmonic ratios), inverse scaling is applied so lower values receive higher scores.

Frequency Bands:

  • Broadband analysis: 20 Hz - 20 kHz
  • Low-frequency analysis: 20 Hz - 200 Hz
  • Coherence: Calculated using Welch's method

THD Calculation: Includes both harmonic distortion and noise floor contributions (THD+N) for REW compatibility.

Interpreting Results

High Scores (80-100):

  • Professional-grade measurements
  • Suitable for critical analysis
  • Excellent for filter design and room correction

Good Scores (70-79):

  • Reliable for most applications
  • Minor issues don't affect overall quality
  • Suitable for general analysis

Caution Scores (50-69):

  • Review specific metrics that scored low
  • May be usable depending on application
  • Consider environmental factors (noise, room)

Low Scores (< 50):

  • Check measurement setup
  • Verify speaker and microphone connections
  • Look for clipping, noise, or room interference
  • Retake measurement after addressing issues

Troubleshooting Low Scores

Low SNR: Check room noise, microphone gain, speaker level High THD: Verify no clipping, check speaker condition, reduce levels Poor Coherence: Ensure stable conditions, check for interference Low Peak-to-Noise: Verify microphone positioning, check for reflections

The scoring system helps ensure your measurements meet the quality standards needed for accurate loudspeaker analysis and room correction work.

Tips for Best Results

  • Ensure room is quiet during measurements
  • Use a calibrated measurement microphone
  • Keep microphone cable away from speakers
  • Allow speakers to warm up before measuring
  • Use consistent microphone height across positions
  • Always use cross correlation alignment before vector averaging
  • Measure at least 3-4 positions per channel for effective averaging
  • Verify Position 0 is your most representative measurement location

Clone this wiki locally