Describe the sound in words
Type a Prompt that describes the sound to isolate and get it back on its own.
Split a song into parts
Sam Audio Text-guided
Audio separation with SAM Audio. Isolate any sound using natural language—professional-grade audio editing made simple for creators, researchers, and accessibility applications.
Made by Open Source500 creditsper run Pro
Opens the audio studio with Sam Audio Text-guided ready to use. You can look around as a guest. Running it takes the Pro plan or higher.
What Sam Audio Text-guided makes, what it needs and what you get
What it makes
Split a song into parts
Splits a song into separate tracks such as vocals, drums and bass.
What you need
- The song to split
- What to change
What you get
Separate audio tracks
- WAV and MP3 files
About Sam Audio Text-guided
Sam Audio Text-guided pulls one sound out of an audio clip. You describe the sound in plain words, and it returns audio with that sound isolated. In Text-guided mode it accepts WAV, MP3 and FLAC files.
You provide the Audio and a Prompt that describes the sound. Predict Spans lets the model work out when the sound occurs. Reranking Candidates, from 1 to 7, sets how many candidates are made and ranked. Higher values improve quality but add waiting time and cost. Max Chunk Duration, from 10 to 60 seconds, is the longest stretch processed at once. Longer audio is split into overlapping chunks and blended. Chunk Overlap sets the overlap, from 0 to 30 seconds.
A second mode, Timespan-guided, asks for Spans, the time ranges where the sound occurs. Its Prompt is optional, and Trim To Span cuts the result to those ranges. A run costs 100 credits, the same as Sam Audio Visual-guided.
Other models for this job
Same job: split a song into parts. Compare before you choose.
What it does well
Find when the sound occurs
Turn on Predict Spans to let the model work out the times when the sound appears.
Long audio handled in chunks
Max Chunk Duration (10 to 60 seconds) and Chunk Overlap (0 to 30 seconds) control how longer audio is split and blended.
Timespan-guided mode
Give Spans, the time ranges where the sound occurs, and add a Prompt if you like.
How to use Sam Audio Text-guided
How It Works
How It Works
01
Add your audio
In Text-guided mode, add a WAV, MP3 or FLAC clip.
02
Describe the sound
Write a Prompt that says which sound to isolate.
03
Generate
Start the run, which costs 100 credits, and play the isolated sound.
Tip Use Timespan-guided when you already know when the sound occurs. It takes those time ranges as Spans.
Good for
Isolate one sound from a recording
Describe the sound and get audio with that sound on its own.
Isolate a sound from part of a recording
In Timespan-guided mode, give the time ranges where the sound occurs and use Trim To Span to keep only those ranges.
Questions
Quick answers to help you get the most out of Sam Audio Text-guided in FunAI Studio.
1. Which file types can I add?
2. What do I get back?
3. How much does it cost?