LALAL.AI 2024 Wrapped: A Journey Through Creative Progress

Oksana -

As 2024 comes to a close, we take this opportunity to reflect on the progress and developments at LALAL.AI over the past year. We’ve focused on enhancing our audio processing tools to serve our users better, and we're excited to share the highlights of our journey.

This year brought many new features and improvements designed to make audio editing easier and more effective for everyone, from musicians and karaoke lovers to podcasters and other content creators. Join us as we explore the key developments that have shaped LALAL.AI in 2024.

Major Product Releases

LALAL.AI Voice Changer

In April 2024, we launched the Voice Changer, which allows you to transform songs and vocal recordings by applying various artist voice clones. You can modify your voice to sound like popular artists such as Drake or Taylor Swift, as well as other artists who have submitted their voice samples to LALAL.AI.

The available voices vary from soulful to melodic and include both male and female options. This means you can create unique sounds for your music or audio projects that feel authentic and engaging. Voice Changer supports multiple languages and various audio formats, including MP3 and WAV, making it easy to use in different creative projects.

LALAL.AI Echo & Reverb Remover

In August 2024, we introduced the Echo & Reverb Remover. This tool enhances audio clarity by effectively removing unwanted echoes and reverberations from your recordings. It uses advanced algorithms and machine learning to analyze your audio tracks and eliminate these effects without altering the original sound quality. Echo & Reverb Remover is especially beneficial for the following purposes:

  • Removing unwanted acoustic artifacts from vocal recordings to achieve a polished, studio-quality sound.
  • Enhancing the clarity of your podcast audio to ensure it sounds crisp and engaging for your listeners.
  • Clearing out reverb and echo from dialogue and narration to elevate the production value of your videos.
  • Eliminating echo for professional-quality voiceovers to deliver clear and impactful performances.

Echo & Reverb Remover supports a wide range of audio and video formats, including MP3, WAV, FLAC, AVI, MP4, MKV, AIFF, and AAC. By improving speech intelligibility and overall sound clarity, this tool helps ensure that every word is heard clearly.

LALAL.AI Lead & Back Vocal Splitter

Just a month after the launch of the Echo & Reverb Remover, we presented the Lead & Back Vocal Splitter. This tool was highly requested by many of our users, and we were excited to fulfill that wish in September 2024. By creating this service, we aimed to simplify what was once a complicated process of lead and backing vocal isolation.

With advanced AI algorithms, Lead & Back Vocal Splitter accurately distinguishes between lead and back vocals, enabling you to manipulate these layers independently. This capability opens up new creative possibilities for remixing songs, creating karaoke tracks, or producing educational content such as vocal training exercises, music theory lessons, and practice materials for aspiring singers.

Perseus AI

Launched just two weeks after the Lead & Back Vocal Splitter, Perseus AI debuted on September 25, 2024. This advanced neural network represents a significant leap in vocal extraction technology. It utilizes transformer models similar to those behind OpenAI's ChatGPT, making it one of the first neural networks to leverage transformers for audio processing.

Designed specifically for isolating vocals, Perseus enhances the clarity of extracted sounds while minimizing background noise artifacts. With an investment of 3.5 times more resources than its predecessor, Orion, Perseus achieves approximately a 15% improvement in vocal extraction quality,

Throughout the remainder of 2024, we expanded Perseus' capabilities to include support for extracting Drums and Bass stems in addition to Vocal and Instrumental and Voice and Noise stems. This means users can now isolate not only vocals and remove background noise, but also extract these essential components from their audio and video tracks.

Key Feature Enhancements

🟡Improved Stem Separation

Throughout 2024, we've made significant advancements in stem separation by enhancing the quality of several essential stems, including Bass, Piano, Acoustic Guitar, and Electric Guitar. Initially supported by the Orion neural network, these stems were previously extracted using its predecessor, Phoenix. The Bass stem, in particular, received improvements twice during the year; first with enhancements from Orion and later with support from Perseus, the latest and most advanced neural network.

🟡De-Echo/Reverb Feature

The De-Echo/Reverb function was first implemented into the LALAL.AI Stem Splitter and Voice Cleaner in February 2024, allowing users to enhance audio quality by removing unwanted echo and reverberation from their recordings. This feature uses advanced algorithms and machine learning to effectively isolate and suppress echo and reverb components, resulting in clearer, more natural sound quality. In August 2024, we launched the standalone Echo & Reverb Remover, making this powerful capability even more accessible.

🟡Reworked "Enhanced Processing"

At the beginning of 2024, we improved our audio processing capabilities by introducing the Enhanced Processing feature, which replaced the previous Processing Level setting. This update allowed users to refine not only vocals but also additional stems, including Bass, Drums, Piano, and Voice. Enhanced Processing utilizes a companion neural network that runs alongside the main stem separation network (Phoenix or Orion) to improve extraction quality. When enabled, it helped reduce the bleeding of auxiliary stems into the main stem and enhanced the quality of auxiliary stems during silent periods in a track.

In December 2024, we further upgraded this feature by introducing a two-settings mode: Clear Cut and Deep Extraction. Clear Cut is designed for straightforward extractions where users want clean isolation with minimal artifacts. At the same time, Deep Extraction caters to more complex audio needs, preserving intricate details even in challenging mixes. This dual approach ensures that users can select the most suitable option based on their specific requirements, enhancing overall audio quality and providing greater flexibility in audio editing.

🟡Noise Canceling Level

The Noise Canceling Level feature was introduced in March 2024, giving users more control over background noise reduction during audio processing. This feature offers three distinct levels of noise reduction to cater to a variety of audio clean-up needs: Mild, which applies no additional processing and retains the audio's original characteristics; Normal, which activates soft compression to subtly reduce background noise while preserving the integrity of the main audio; and Aggressive, which implements more forceful compression to significantly reduce noise, making it perfect for tracks recorded in environments with high background noise levels.

With Noise Canceling Level, users can quickly enhance their audio projects, whether they're cleaning up voice or vocal recordings, podcasts, or any other audio content. The feature can be accessed through the upload section of the LALAL.AI Stem Splitter by selecting the Voice and Noise stem and adjusting the settings according to their specific needs.

🟡Output Format Selection

The Output Format Selection possibility was introduced on August 12, 2024, allowing users to choose from a variety of output formats when downloading their processed audio and video files. This addition provides greater flexibility, enabling users to select formats that best suit their projects. Supported audio formats include MP3, OGG, AAC, AIFF, WAV, and FLAC, while video files can remain in their original format or be converted to audio formats.

Impressive Usage Statistics

As our platform continually evolved in 2024, engagement levels have soared, reflecting the growing trust and reliance on our audio processing tools. The following statistics illustrate the remarkable scale of activity and highlight how our features have made a significant impact on users' projects throughout the year.

  • 5,599,384 stems split, demonstrating our ability to handle complex audio tasks with ease.
  • 1,132,634,499 seconds of content processed this year alone, which equates to approximately 13,036 days, or 35 years, 9 months, 334 days, 4 hours, 41 minutes, and 39 seconds.
  • 3.8 million users on the platform, which includes 1.3 million new users compared to last year.
  • 9.7 million files uploaded, with MP3 being the most popular format at 3,083,658 uploads, followed by WAV (1,212,159), MP4 (528,018), M4A (384,043), FLAC (120,328), and MOV (56,836).

Top Stems and Formats

The top three most popular stems extracted with LALAL.AI this year were:

  1. Vocal and Instrumental
  2. Voice and Noise
  3. Drums

The top three most popular formats were:

  1. MP3
  2. WAV
  3. MP4

Unsurprisingly, MP3 remains the dominant audio format due to its storage efficiency, compatibility, and widespread accessibility.

What's Next

As we look forward to 2025, we're committed to further improving the capabilities of the LALAL.AI products. One of our primary goals is to improve the remaining stems by integrating Orion support, which will elevate the quality and precision of audio separation across all instrument categories.

Additionally, we're excited to announce plans for a new product that will complement our Voice Changer service. Curious about what it might be? Get ready for this major release, it will be available soon! For now, let’s shift our focus to the New Year celebrations. We look forward to seeing you in 2025!

We wish you all a joyful and prosperous New Year filled with creativity and inspiration!🎉

Follow LALAL.AI on Instagram, Facebook, Twitter, TikTok, Reddit, and YouTube to keep up with all our updates.