Skip to main content

Fat Llama Logo

Fat Llama build - status PyPI PyPI - Downloads

fat_llama is a Python package for upscaling audio files to FLAC or WAV formats using advanced audio processing techniques. It utilizes cpu-accelerated calculations to enhance audio quality by upsampling and adding missing frequencies through FFT (Fast Fourier Transform), resulting in richer and more detailed audio.

Features

  • Upscale MP3 files to high-quality FLAC format.
  • Iterative soft thresholding (IST) for enhanced audio processing.
  • Auto-scaling amplitude adjustment and normalization.
  • Multi-Threaded processing on cpu.

Installation

Install via pip:

pip install fat-llama-fftw

(Note: For CUDA verison please look at https://pypi.org/project/fat-llama/)

Also, requires ffmpeg: https://support.audacityteam.org/basics/installing-ffmpeg

Usage

Example Usage

You can run the example provided in example.py:

from fat_llama_fftw.audio_fattener.feed import upscale

# Example call to the method
upscale(
    input_file_path='input_test.mp3',
    output_file_path='output_test.flac',
    source_format='mp3',
    target_format='flac',
    max_iterations=1000,
    threshold_value=0.6,
    target_bitrate_kbps=1400
)

Function Parameters

  • input_file_path (str): Path to the input audio file. Mandatory.
  • output_file_path (str): Path to the output processed audio file. Mandatory.
  • source_format (str): Format of the input audio file (e.g., 'mp3', 'wav', 'ogg', 'flac').
  • target_format (str): Format of the output audio file (e.g., 'flac', 'wav'). Default is 'flac'.
  • max_iterations (int): Maximum number of iterations for IST. Default is 800.
  • threshold_value (float): Threshold value for IST. Default is 0.6.
  • target_bitrate_kbps (int): Target bitrate in kbps. Default is 1411.

Running the Example

To run the example, execute the following command:

python example.py

This will upscale the MP3 file specified in the example and produce a FLAC file with full processing.

Spectrogram Results

Spectrogram Results

Audio Quality Scores

Generated by the test-fat-llama skill's audio-quality-checker subagent — updated each run, not hand-edited.

Metric Score Notes
Coherence (upscale quality, 0-10) 9.5 every hard check clean and mostly beats the reference (peak exactly 1.0, DC/above-Nyquist/discontinuities/hop-boundary all clean); the interpolation rewrite (zero-order-hold to bandlimited FFT-domain) closed the long-standing "no added detail" gap - the reference's thinnest 5% of in-band bins gain +8.03dB, with clear positive gains from 8kHz to 22.05kHz (up to +5.36dB), confirmed programme-correlated rather than a noise-floor rise
Spectral deviation vs. reference FLAC (0-10) 9.9 convergence=0.9713, correlation=0.9998 against input_test.flac; note input_test.flac is itself a legacy zero-order-hold-era pipeline output, not a clean master, so this should not be read as a quality ceiling

How it works

How it Works

Algorithm Explanation

The upscaling process involves several steps:

  1. Reading Audio File: The audio file is read, and the audio samples are extracted along with the sample rate and bitrate.
  2. Calculating Upscale Factor: The upscale factor is calculated to achieve the target bitrate.
  3. Upscaling Channels: The audio channels are upscaled using an interpolation algorithm. Each sample is repeated multiple times to increase the resolution.
  4. Iterative Soft Thresholding (IST): IST is applied to enhance the audio by adding missing frequencies. This process uses FFT to transform the signal into the frequency domain, apply a threshold to keep significant frequencies, and then inverse transform back to the time domain.
  5. Scaling Amplitude: The amplitude of the upscaled audio is scaled to match the original.
  6. Normalizing Audio: The audio is normalized to the range -1 to 1.
  7. Writing FLAC File: The processed audio is written to a FLAC file.

Why FFT and IST?

FFT (Fast Fourier Transform) is used to transform the audio signal into the frequency domain. This allows for the identification and manipulation of specific frequency components. By applying a threshold in the frequency domain, we can keep significant frequencies and discard noise and add it to our upscaling data to add detail to upscaling frequencies.

The report titled "Fast Sparse Fourier Transformations for NMR Spectroscopy" by Badruddin Kamal, supervised by Thomas Huber and Alastair Rendall, 2015, provides a comprehensive understanding of sparse representations and their applications in signal processing. IST leverages the concepts from this report to add missing frequencies and enhance the audio quality by making it more detailed and rich. This is particularly useful in upscaling audio where some frequencies might be missing or congested.

Test Audio Source

ericzo - beyond link(https://soundcloud.com/ericzomusic/free-electro-trap-anthem-beyond)

Changelog

Changes are now logged in CHANGELOG.md

Release files for fat-llama-fftw 1.4.2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for fat-llama-fftw 1.4.2
File Size Uploaded
fat_llama_fftw-1.4.2.tar.gz 40.9 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for fat-llama-fftw 1.4.2
File Interpreter ABI Platform
fat_llama_fftw-1.4.2-py3-none-any.whl Python 3 none any Details

Total release size: 80.7 kB

Release files / fat_llama_fftw-1.4.2.tar.gz

Download URL fat_llama_fftw-1.4.2.tar.gz
Size 40.9 kB
Tags Source
SHA-256 checksum
How to use checksums
1ca89f2da64549ae679b8805c277904cca06263790691a40e75e7d8237de2673
BLAKE2b-256 checksum
How to use checksums
4d01bdb3ce669601c494304e999a6119678d110caf7175a6a1f4d29a36bb20fc
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.14

Release files / fat_llama_fftw-1.4.2-py3-none-any.whl

Download URL fat_llama_fftw-1.4.2-py3-none-any.whl
Size 39.8 kB
Tags Python 3
SHA-256 checksum
How to use checksums
3d3c8a7795a49972c3975f0639234f30072596bcf8df878c336c78c8f5218bad
BLAKE2b-256 checksum
How to use checksums
7258394e56e35f81a1d296063aab59701025720fe305797e0502acf0c5171dda
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.14

Release history Release notifications | RSS feed

2.0.2

2 release files

1.4.5

2 release files

1.4.3

2 release files

This release

1.4.2 This release

2 release files

1.4.1

2 release files

1.0.4

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page