What are “stems”?
Stems are essentially the primary constituents of a song; they’re audio tracks which, when combined, create the final and complete mix. You can think of them as sub-mixes, which can be independently controlled within the context of the entire product. For example, the aggregate of all the drums in a song can be put into one track to create a drum stem.
Now, depending on how a song was created, the original producer(s) could simply go into their music project and render out a clean drum track — that is, the drums without the accompanying piano, bass, synths, etc. However, end users (the listeners) typically do not have this luxury, as the complete mix would be a single file which cannot simply be undone. It would be like trying to un-bake a cake.
Simple Stems seeks to bridge this gap by using AI technology to guess where the stems are in a song. Stagecraft Software did not make this technology; we adopted it from the open-source project Spleeter and created our own application to offer a simple, intuitive interface, local (offline) stem splitting, file dragging/dropping, and more at an affordable price. We cover more about what you can separate and the quality of said separation below.
What exactly can you separate?
Simple Stems can separate audio in three ways:
– vocals and accompaniment (2 stems)
– vocals, drums, bass, and other (4 stems)
– vocals, drums, bass, piano and other (5 stems)
How long does separation take?
Simple Stems processes audio locally on your machine, so the amount of time it takes to separate largely depends on the specs of your machine and the amount of audio being split. Generally, with an average computer, you can expect every recorded minute to take about ten seconds of processing time.
Unfortunately, due to limitations of the Spleeter library, we cannot show a progress bar while Simple Stems is working. This may be implemented at a later point in time.
How well does Simple Stems decompose audio?
The largest factor determining the quality of the separation is the source material. Clean, well-mixed studio recordings will generally separate quite well; however, noisy or crowded audio may be more of a struggle. Whether or not the quality is good enough for you will depend on your individual use case.
We offer a full version of Simple Stems free for 30-days, which we recommend you try before purchasing. You can also check out the video below for a deeper dive into the capabilities (and weaknesses) of the algorithm.
What audio formats are accepted?
Most popular file formats will work, like MP3, AIFF, FLAC, and OGG. You can also put in most MP4 and MKV files, but note that you will only get the audio back.