SoundMax AI Audio Enhancement & Processing API for Windows, Linux, MAC and DSP

About

Craft a breathtakingly detailed, consistent, and balanced sound for your audio applications using Perceptual SoundMax 24-Band high resolution Audio Processing API developed by ATC Labs. This product brings novel innovations in audio processing & AI/ML based sound-quality enhancements to a variety of applications including Professional Broadcasting, Audio Production, Audio/Video Streaming, Automotive Audio, Consumer Electronics, and more.

SoundMax AI is high resolution Real-time Audio Processing engine which offers control over up to 24 psycho-acoustically tailored bands with high time resolution of 0.5 msec. The multi-band processing in Perceptual SoundMax is totally Linear Phase which eliminates the need for phase-compensators common in conventional audio processing algorithms. It implements all envelope processing (i.e. expansion/AGC and compression/limiting) in each of the 24 bands in a single stage. This avoids a situation where different processing stages are fighting and/or negating each other’s impact. Further, proprietary smooth windowing technology is used to minimize spectral growth (increase in high frequency energy when processing leads to saturation). This is particularly helpful for digital broadcasting.

Perceptual SoundMax utilizes AI-ML based techniques throughout its processing platform. Techniques included sophisticated models for multiband noise analysis, signal tilt analysis, and our patent pending Real-Time Multi-Class Hierarchical Audio Classification Deep ML model that can be used to adaptively adjust processing parameters based on changing audio characteristics. High Resolution with high Accuracy AI model is trained and used in processing.

Audio Processing Tools

  • Multi-band (24-band) high-resolution Dynamic Range Compression (DRC), provides great control in enhancing audio attributes and improving presence and listener impact 10 band dynamics processing.
  • Intelligent Loudness Control (ILC): Intelligent Loudness Control algorithm with enhanced subjective loudness model. It is high time resolution broadcast quality algorithm maximizes loudness and provides the desired sound density.
  • Vocal Enhancement (VE): Vocal, and dialog enhancement .
  • Bass Enhancement: Tunable bass boost module with 2-independent stages.
  • High Frequency enhancement: Tunable treble boost module for high sound quality.
  • Stereo Enhancement: Stereo image stabilization & enhancement module.
  • High Quality Noise Reduction: Adaptive Wide Band Noise Removal (AWNR) that does not distort the audio.
  • Deep AI Technology based Audio classifier. This Patent pending Real-Time Multi-Class Hierarchical Audio Classification Deep ML model accurately Classifies signal into Noise, Speech, Instrumental Music, Vocal Music, genres with high accuracy and high resolution. Useful in adaptively changing processing parameters for best quality.
  • Multi-band Noise controlling based on AI multiband noise analysis and signal tilt analysis. It is useful in loudness enhancement for noisy speech and old recorded signals by gating the noise region perfectly.
  • Innovative Dynamic Listening Fatigue Reduction (DLFR): envelope conditioning for improved listener experience
  • Final Look Ahead Limiting A sophisticated final SmoothClip/SoftClipping algorithm ensures distortion free look ahead limiting.

Features

  • Utilizes deep AI technology for controlling and adapting the audio quality. Multi-Models based on LSTM neural networks are trained on huge audio database comprising Noise, speech, Music, Vocals, different genre of songs to get high accuracy and high-resolution models.
  • Enhanced Consistent Loudness: 24-Band high resolution Audio Processing allows packing in much higher volume in comparison to 5 or 6 band processing and modification of only the desired audio attributes. This results in substantial boost in the overall Sound Density and Loudness, yet the overall sound is distortion free and does not sound over-processed.
  • Crisp & Clear Sound: State of art signal processing ensures that the processing output is free of any phase distortions and doesn’t sound over-processed. Thus, it enhances clarity in addition to loudness and presence.
  • Lively Audio: Vocal Enhancement, Stereo Image Enhancement, Bass Enhancement, and available Sweetening tools make the sound livelier and pleasing to listen to.
  • Maintains Target Level for audio envelope, Improved listening in noisy environment, Eliminates need to constantly adjust volume – from song to song and within song.
  • FM Broadcast Specific Advantage of this Audio Processing Platform: Maximize Modulation Index.
  • Suitable processing to pre-condition audio to be more robust to codec distortions. Codec distortions highly sensitive to non-linearity like clipping, Quantization noise may require creating a headroom – i.e. peak limiting @ -2 to -4 dB Full Scale which is possible with SoundMax processing. Less distortion in AAC/HE-AAC+ 32-128 kbps, MP3 128-256 kbps with processing as compared to no processing.
  • Web based Controller & Profile Generation Software: Web based Controller Software called SoundMax WebConnect running on any browser is available which can fine tune and control the audio quality of SoundMax API and generate new profiles. All modules tunable in real-time without audio jitters.
  • Supports real-time applications and easy to port or integrate in any real-time audio system. The integration of high quality 24-Band audio processing directly into any audio transport products increases ease of deployment and reduces cost.
  • Single API based solution perfect for all type of audio solution even with low computation requirement.
  • Supports Nielsen Watermarking.

Platforms Supported

Perceptual SoundMax Audio processing API is available on following platforms

  • Windows 10/11, both x64 and x86 architecture builds are available.
  • Centos, Ubuntu and RedHat Linux x64 architecture.
  • Yocto and Ubuntu embedded Linux builds on ARM architecture.
  • Universal MAC 64-bit builds for both Intel MACs and ARM based MACs.
  • Cadence Tensilica 3 major architectures which are HiFi 3, HiFi 3z, HiFi.
  • ARM Cortex-M55 architecture builds
Architecture
MIPS
ROM
RAM
HiFi 3 219 255K 225K
HiFi 3z 189 257K 225K
HiFi 4 190 259K 225K

Ready to Create Together?

Get in touch