1. Intro / Motivation

After experimenting with simple audio filters (bass, treble, telephone), I wanted to try something more dynamic: a compressor.
Compression is one of the most important tools in audio production. It helps balance quiet and loud parts of a recording, making the overall sound more consistent and polished.

This project is part of my ongoing journey of exploring audio processing in Python, and of course, building experience that I hope will help me one day work at Dolby.

2. Project Overview

The compressor script works as a command-line tool that:

• Reads a WAV file.
• Applies dynamic range compression with adjustable parameters.
• Outputs a new, compressed WAV file.
• Optionally generates a visualization showing both waveforms and the gain reduction applied.

The adjustable parameters include:

• Threshold (when compression kicks in)
• Ratio (how strongly loud parts are reduced)
• Attack / Release (how fast compression reacts)
• Makeup gain (boost after compression to restore volume)

3. Technical Approach / Key Concepts

Like my filtering project, this script was built to use only a small number of dependencies:

• numpy for numeric processing
• scipy.io.wavfile for reading and writing WAV files
• argparse for command-line arguments
• matplotlib for optional visualization

The key steps are:

1. Convert audio to float32 and normalize (so it works between -1 and 1).
2. Calculate input levels in dB for each sample.
3. Compare to threshold:
   • If louder, reduce by the chosen ratio.
   • If softer, leave unchanged.
4. Smooth transitions using attack and release.
5. Apply makeup gain to bring overall volume back up.
6. Write the result back to a 16-bit WAV file.

Here’s a graph showing what it looks like in action (example using a copy of AJR's The Big Goodbye):

In the first graph, you can see the original waveform (blue) vs. the compressed waveform (orange). In the second, the gain reduction applied to each stereo channel.

4. Challenges / Lessons Learned

• Getting the attack and release smoothing right was tricky. Without smoothing, the sound was unnatural and choppy. With it, the compressor reacts more musically.
• Stereo audio had to be handled carefully — each channel is processed independently, so the compressor works correctly on both left and right.
• The visualization was harder than expected, but ultimately very useful. Seeing gain reduction over time made it clear how the compressor behaved.

5. Next Steps / Future Experiments

Some improvements I’d like to explore:

• Add a sidechain input (classic “ducking” effect).
• Try different compression curves (soft knee, multiband).
• Add support for real-time processing, instead of just offline WAV files.
• Combine this with my filters into a larger “audio lab” toolkit.

6. Conclusion / Takeaways

This project gave me a much better understanding of how compression works under the hood, and why it’s such a powerful tool.
Even though it’s a basic script, it’s exciting to see how much the sound changes — and to visualize it clearly.

I plan to keep building on this, moving toward more advanced effects and possibly chaining them together into a more complete processor.

Thanks for reading!
- Heath / MinoDab492

(Note: the code used in this project is not open-source or available on GitHub or other platforms. If you'd like to request access to the code for whatever reason, you can email me at heath.garvin@minodabproductions.dev, or contact me on Discord @minodab492.)