~5m15:17
yourtunēs AI

ACE Music рвёт все рейтинги! Проверяем, на что она способна новая нейросеть

Nov 17, 2026

Read: ~5m · You save: 10 min

ACE Music Breaks All Ratings! Testing What the New Neural Network Can Do

Testing Ace Music from TuneCore: tracks from melody, prompt and Russian lyrics. Instrumentals impress, but vocals still sound like AI. Watch!

Publication date: 17.11.2026

ACE Music is a neural network product for creating music. According to independent companies that conducted blind listening tests, it holds leading positions in rankings and receives positive reviews. In listener tests, the audience does not know which model created a track and votes blindly.

The author of the review checked how the service understands musical styles, the Russian language and stress placement, and how it performs in practice.

Sign-in and Ways to Create a Track

Sign-in is done through a Google account or by email. The user can start working in one of the following ways:

  • from an idea;
  • from a text description;
  • by recording a melody with a microphone or by performing a melody live;
  • by uploading a musical fragment;
  • using an image or video as a starting point.

To record a melody, select a microphone. The review used a standard Mac microphone. The recording can be canceled and redone.

Track Creation Process

  1. The user chooses a genre, for example rock, jazz, trap, or hip-hop.
  2. The genre, mood, instruments, and tempo are described in the prompt field.
  3. AI Studio creates track variations within a few seconds.
  4. The user selects a suitable variation.
  5. The cover is chosen from ready-made options or created with AI.
  6. The track is published and sent to streaming platforms.

The entire process, from idea to finished release, takes place in a single service. Payment is available in Russia, including payment with Russian cards.

Personal Account

The sidebar menu contains two sections, one of which is not yet working. The section is under development; according to the interface description, it will publish created music and genre ratings. The interface also invites users to join the community on Discord.

The main menu contains two working sections. In them, you can develop an idea based on a recorded melody, an uploaded fragment, an image, or a video. Multiple compositions can be created within a single session.

The advanced mode allows working with song lyrics and the music prompt separately. The duration of the composition is set automatically or can be selected manually in seconds.

Test Results

Test 1: Instrumental hip-hop from a recorded melody

Model: Ace Music V1, which the author called the most powerful version of the service.

  • The melody, hummed into the microphone, was transmitted correctly: the rhythm matches the original.
  • The sample sounds raw and unfinished.
  • Low frequencies are nearly absent, and the bass is poorly developed.
  • The beat (kick and snare) sounds good, but overall the track was rated as rough.
  • The author rated the second variant of this test as unsatisfactory.

Test 2: Generation from a prompt without a melody

The prompt was given without a recorded melody.

  • Two variants with different melodies were produced.
  • The bass, drums, and melody sound coherent.
  • The sound quality was rated as high.
  • The author noted that in Suno, one prompt produces 20 compositions that sound the same. In Ace Music, the variants differ.

Test 3: Russian lyrics, melodic rap

Parameters: song lyrics in Russian, style influence 82%, duration 230 seconds. The lyrics were generated via GPT chat specifically for the test.

  • Pronunciation is normal, and the stress placement is correct.
  • The vocals sound synthetic. The author stated that they contain frequencies that reveal their artificial origin.
  • The recording's sound quality was rated as full.
  • In the author's opinion, the vocals require refinement. Eliminating the synthetic effect would make the result fully satisfactory.

Test 4: Style change to acoustic jazz swing

The same lyrics, style influence of 82%, and duration of 230 seconds were used.

  • The vocals sound more natural than in Test 3.
  • In places, the vocals are too clean.
  • The style matches the specified one, fully according to the author's assessment.

Summary of Test Results

TestTaskAuthor's Assessment
1Beat from a recorded melodyMelody transmitted correctly, weak bass, rough sound
2Generation from a promptHigh quality, variants differ
3Russian lyrics, rapCorrect pronunciation and stress, synthetic vocals
4Jazz swing with the same lyricsStyle match, more natural vocals, but too clean in places

Author's Conclusions

  • The service is at an early stage of development. Two versions of the track are available, with version 1.1 being a preliminary one.
  • According to the author's assessment, the first version is of high quality.
  • The author expects improvements in vocals and a more accurate understanding of prompts.
  • The author suggested that the product may compete with Suno. He also allowed for a possible agreement with rights holders. The author did not confirm either assumption with facts.

Introduction and Essence of Ace Music

Ace Music is described as a neural network that leads blind listening tests run by independent companies. The author intends to test how well it understands styles, the Russian language, and stress placement.

  • Ace Music is called a new product in the world of music neural networks that, according to the author, is currently topping all the ratings.
  • Independent companies run blind listening tests: the audience votes without knowing which model created the track.
  • Ace Music stands out in these tests and receives very positive reviews.
  • The author wants to check the model's understanding of styles, the Russian language, and stress placement in practice.

Getting Started and Recording a Melody

A user signs in with Google or email and chooses a starting point: an idea, lyrics, a melody recording, an uploaded fragment, an image, or a video. The author records a melody using a standard Mac microphone and selects the trap style.

  • Sign-in is possible via a Google account or email.
  • You can start from an idea, lyrics, a recording or played melody, an uploaded music fragment, an image, or a video.
  • The sound selection includes rock, jazz, trap, hip-hop, and other genres.
  • The author records a melody using a standard Mac microphone and selects the trap style.
  • The melody recording can be undone (reverse).

Release Creation Process in AI Studio

AI Studio creates tracks from a description of genre, mood, instruments, and tempo, offers cover art options, and publishes to streaming platforms. Payment is available from Russia with Russian cards.

  • Process: track description (genre, mood, instruments, tempo), choice from ready-made options, cover art, publishing (Publish) to streaming platforms.
  • The cover can be chosen from ready-made options or created with AI.
  • The service is fully available in Russia, and payment is possible with Russian cards.
  • The link to AI Studio TuneCore is provided in the video description.

Overview of the personal dashboard

The dashboard has only two tabs in the side menu, one of which does not work yet. The publishing section and genre charts are still under development, and the transition to AI Studio, according to the author, is a separate product.

  • The side menu has two tabs, one of which does not work yet.
  • The publishing section and genre charts are under development.
  • There is an invitation to Discord.
  • The transition to AI Studio is considered a separate product that is not discussed in this video.

Test 1: Melody Recording

First test: instrumental hip-hop with a hummed melody. The author finds the melody copied correctly, but the sound is raw, with weak low frequencies and poor bass. The Ace Music V1 model is named the best according to tests.

  • An instrumental hip-hop track with a hummed melody is being tested.
  • The melody is copied correctly, with the same rhythm.
  • The sample sounds raw and unfinished, low frequencies are almost absent, and the bass is weakly developed.
  • The Ace Music V1 model is used, which, according to the author, ranks at the top in tests.
  • The author rates the result as unsatisfactory.

Test 2: Generation from a Prompt

The text-prompt test without a melody produced a significantly better result: the bass, drums, and melody sound good, and the sound quality is high. Unlike Suno, the service creates two versions with different melodies.

  • The prompt was entered without uploading a melody.
  • Result: the bass, drums, and melody were judged good, and the sound quality is high.
  • The service created two versions with different melodies, rather than many identical ones as Suno does (according to the author).

Test 3: Russian Language and Vocals

Test of Russian vocals: the text was generated via GPT, style influence 82%, duration 230 seconds. The vocals sound synthetic, and the author notes characteristic frequencies that reveal an AI voice. Diction and stress placement were rated as correct.

  • The text was generated via GPT chat specifically for the test.
  • Settings: style influence 82%, duration 230 seconds, track title "Song title test two".
  • Diction is normal, stress is placed correctly.
  • The vocals sound synthetic, with frequencies that reveal an AI voice.
  • Sound mixing is acceptable without complaints, but the vocals need to be improved.

Test 4: Style Change to Jazz

Third test: the same Russian rap verse is reworked into an acoustic jazz swing. The vocals sound more realistic and the style matches precisely, though in places it is a bit too clean.

  • Original style: melodic Russian rap; target: acoustic jazz swing.
  • Settings: style influence 82%, duration 230 seconds.
  • The vocals sound more lively and realistic than in the previous test.
  • The style matches perfectly, but in places the sound is too clean.
  • The test was deemed successful for style matching.

Conclusions on the product's development

The product is at an early stage: only two versions are available, and only the first is fully featured. The author expects the service to develop, improve vocals and prompt understanding.

  • Only two versions are available, with version 1.1 currently just a preview.
  • The first version is described as high quality.
  • The author expects improvements in vocals and prompt understanding.
  • The author believes Ace Music could replace Suno, but this depends on an agreement with the labels.