I did not see any mention of this, but i tried something in version 12 with tags.
You can add tags yourself as everybody knows, but you can not add an subtag to an already excising group tag. There is no button for.
these are standard groups with there already created tags.

but there is a way around this to still ad an own tag under that standard group. do the following.
like you mis FX/Sweep up or FX/Sweep low or in this example under standard group " Type" the tag [STEMS]

1. take a sample folder.
2. in Live 12, select that folder and select all samples with stems and select as tag loop
3. go to that sample folder and locate the ...xmp file from the folder where the stems samples are.
4. open this in Note++
5. add your own subtag under that standard group tag.
6. Copy the same rule from type|Loop and replace this with Stems.
7. Reload in Live. and there it is...an extra subtag under an existing group tag " Type"



<rdf:li rdf:parseType="Resource">
<ablFR:filePath>01_160BPM_F#minor_Open Hat 2.wav</ablFR:filePath>
<ablFR:keywords>
<rdf:Bag>
<rdf:li>SampleCreator|Flame Audio</rdf:li>
<rdf:li>Type|Loop</rdf:li>.....Type is the group
<rdf:li>Type|Stems</rdf:li>.....Stems is own created subtag under that group.
</rdf:Bag>
</ablFR:keywords>
</rdf:li>

 

You can add sub tags by using '|' for example tag|sub tag.

 

They had like 10 years to add even simplest features and decent sequencer so it stopped interesting me

 

How stable is the beta?

 

I’m using it every day for personal projects, the only real issue I’ve noticed is crackling audio while monitoring is active during recording. For this reason I’m still using 11 for client work, but as soon as the beta fixes that issue I will completely switch over.

 

12 beta works flawless here

 

I have tried this but this does not work.

 

You guys don't have to wait here it is (I'm part of the Beta testing team since v12.1): Its over at centercode, just google it.

Build can be downloaded in the Beta program.

Expressive Chords:
you can get it also in the Beta program!

 

Last edited by a moderator: Sep 2, 2025 at 7:03 PM

Main v12.3 integrations:

SPLICE included in the "Packs" section
If you have a Splice account you can use directly inside Ableton the built in AI sound matching system (you drag & drop your loop or drum kick anything and it matches the key & tempo it's matching from Splice database what would sound good with your sound/loop inside your Ableton Project)

Inside arrangement view you can now do instant STEM SEPARATION
So Ableton now has it's own Stem Separation function!!!

 

Its def chugging along. Seems to have some CPU lags. Might just be some new library things happening. Ill have turn off my lib and do some more testing.

Stems are cool something Ive wanted in daw, its probably just Demucs 4 stem under the hood. Suno Studio / stems gets you up to 12 and sounds cleaner IMO






One thing the beta is doing for now is once its stemmed it doesn't keep the same warp points and timing. Im sure this will be fixed by public release. Right now its just like stemming with another tool and importing them into ableton. Kind of a bummer on first use.



NOTE: Demucs, MVSEP, and a lot of these AI stem algos that FL studio and others use. Are trained on open internet data. So they use copy written music to train the models that know what each stem is. I dont mind, but food for thought to all the anti AI people out there.

 

Ah just a quick update from SonicBloom's review:

If you put this inside your options.txt it will give you better quality (HQ mode sort of) stem separation I have no idea if it's better or not
Maybe in some cases it works better here's the line simply copy/paste to options.txt
-_Feature.StemSeparation.HighQuality=True

 

welcome features! In my opinion, they should have said bye splice, and hello suno. With suno studio's evolution, and the continued evolution, its only a matter of time before it or another tool replaces splice. Although, Splice has mentioned they are delving into AI, so maybe that's the long term goal. They would get a ton of hate, but its only a matter of time before suno or similar tools are daw native, may as well jump the gun.

 

Fully agree with this. All daws will have gen AI features in them. As well as prompt based editing.

Funny thing is suno studio is starting from a full track and then stripping it away and adding new elements. This is the reverse for ableton and other DAWs.

They are starting at different places will all find a middle ground somewhere.

 

¿Does anyone know if the steam sepparation runs on CPU/GPU?

There is a lot of QoL in this update, the boucing group is a godsend in terms of reducing tedious tasks.

 

its pretty quick on my M1 might be using the MPS chip. Ill try on my old intel later and see

 

Re this:

https://www.instagram.com/reel/DOH0nUwCR7I/?igsh=NTc4MTIwNjQ2YQ==

 

They should sort out the indexing chug.. so many people reported it and loathe it.

 

I don’t need stem separation for music production. I need ARA2 for production.
They added the attack parameter to the tremolo. WOW!!! I remember I used this parameter in Roland D10 ages ago.