There is a key difference between music created with Mubert and a linear composition created in DAW, where the author decides at what point in time this or that element will be played. For the algorithm to better combine parts, samples should be one-size-fits-all: any sample within the stream, should, at any time fit in with the others in terms of rhythm, harmony and melody.
<aside> đź’ˇ Sample: In this context, it is a digital sound fragment, a loop used by the Mubert algorithm to generate a musical sound stream.
</aside>
All samples are divided into 12 functional groups (INSTRUMENT tag on the platform):
All samples should be 4/4 and and in one selected BPM (per stream).
The duration of loops may be 1/2/4/8/16/32/64 bar(s). The preferred loop duration is 4 to 16 bars.
Rhythm elements may be shorter for better variability.
<aside> 💡 We recommend making samples with reverb or release “ends” and/or a high level of sustain longer, so that the stream plays more smoothly and continuously.
</aside>
<aside> đź“Ž See a configured template project in Ableton Live 11 by CELES7E
</aside>
*A session from the author of many sample packs for Mubert, with conveniently configured fades, instrument groups, and a lot more.*
Each instrument in a stream almost always has several variations. This is due to the fact that a stream can potentially play for hours, unlike a conventional track, which is usually limited to only a few minutes.
Any timbre from the presets should always be “fine-tuned” and reworked in your own way, since most licenses do not allow the creation of sample libraries from ready-made presets.
<aside> đź’ˇ How to create timbre variations
When creating variations, in terms of sound it is important to keep in mind the range assigned to this instrument in the general picture of the stream in terms of mixing. In terms of style, take note of the genre, vibe, atmosphere, mood and other qualities of the stream (and the reference, if any) you want to create.