ONE BACKING TRACK. THREE WORKFLOWS.
Comparing three guitar-solo generation methods
Three ways of generating a guitar solo using the same backing track and ACE-Step model.
Listen first. Make your own comparison.
- Start with the backing. Then listen to a solo with the backing.
- Compare the same moment. Stay in one round and switch methods. Listen for fit, phrasing and joins.
- Make your own judgement. Leave a note if you like, then choose whether to reveal Shaozhi’s reasoning.
NOW SELECTED
Loading audio…
Drag across the waveform to repeat any passage. Click once to seek.
No custom selection. Drag on the waveform or enter start and end times.
Select a version, then press play.
The musical outline
Click a section to jump there. Repeating keeps the same section when you switch methods.
Your listening notes (optional)
If you have a moment, leave a few impressions. Skip anything you are unsure about, or just write a note.
Select “With backing” or “Guitar only” to comment on this solo.
Changes also save automatically as you write or switch takes.
Review before sending
Notes stay in this browser until you choose to send them. You can also export the JSON file and send it to Shaozhi yourself. Listen to as many or as few takes as you like.
Saved notes are specific to this browser and website address. Export a copy before clearing browser data or leaving a private window.
AFTER YOUR OWN LISTEN
Shaozhi’s reasoning
Kept hidden so you can form your own impression first. Opening this is optional.
How the selected method works
THE QUESTION BEHIND THE COMPARISON
Does the way we generate the solo change the result?
We want to compare how the three workflows sound over the same backing track, and how consistently they work across generation rounds.
Listen for how well the guitar fits the backing, whether the phrases and section changes feel natural, and whether the solo follows the musical outline. Also notice the guitar tone, noise, distortion or sudden changes in sound.
Try a few rounds if you have time. Does an impression hold across different takes? We are comparing the complete workflows, without assuming that one will sound better.