Akai Force Forum: Everything relating to the Akai Force, the new 64 pad, clip-based standalone sampler/groovebox from Akai. While not an MPC, it shares many similar software features to the MPC X/MPC Live including the same underlying code-base.
By darkinprg Tue Nov 02, 2021 8:46 pm
Hello aall.
Iam prety new to Force, but quite experienced with Elektron machines. Force was supposed to replace the HUB which till now was Octatrack.

After switching the units, and putting Octa and other machines on slave - Force midi out - in/thru - in /thru etc. I started to play around and found out this ...

Force internal drums (samples or synth) play sooner, then the other midi clock synced machines...

Means if i run kick on Octatrack - it comes a little later compared to Force internal drums / metronome...

I managed a workaround where I run Force internal synth channels into subgroups, where I put 8ms delay on each of them and Iam fine, synced...

But again when i record external synths into clips, i have to sort this delay again and move each samples start points everytime i record into a clip...

I tried to remove the usb sound card from the path, and record with Force internal soundcard, but i get the same results...

So, are those 8ms related to the the audio engine, and can that be tweaked somehow?


Anyone else noticed this?
I havent compared the dry audio signal from the machines with dry force audio signal...


Cheers peace
Mike
User avatar
By Koekepan Wed Nov 03, 2021 1:35 am
I think I understand what you mean, but if I'm getting something wrong, feel free to correct me.

What I'm hearing is that Force -> internal audio engine gives a certain time

and that Force -> USB MIDI -> external device -> audio in (whether USB or direct) is time + 8ms.

I'm also hearing that all your external devices do the same (although you specifically mention the Octatrack).

What happens if you loop MIDI from the Force MIDI Out to the Force MIDI IN and trigger it?
By darkinprg Wed Nov 03, 2021 7:48 am
Hi, thanks for response.

You probably get it right...
In all my tests the audio from devices synced to force was coming back to force.

I still have to test and see results with when the audio from other gear is not coming back to force but some mixer, that should tell me whether there is latency from midi or happens on audio side, am i right?

Rgs, M
User avatar
By Koekepan Wed Nov 03, 2021 3:04 pm
There's a lot to test here, with different aspects of the system capable of causing latency.

Regular MIDI over a regular interface takes roughly 1ms for a single, three byte message (not all messages are three bytes). USB MIDI is also noted for delay counted around a couple of ms, so while it is fast, it is not instant. This means that anything going out through MIDI and coming back in should be expected to have some delay.

This is why I suggested that you test the Force just talking MIDI to itself, to figure out how much latency results from just that signal transmission and reception.

It is also possible that an audio engine would add latency; after all, there are factors such as buffering involved. However, to test that you would first have to make sure that you had two devices triggered simultaneously, perhaps by the intelligent use of a MIDI splitter.

The short answer is: "Yes, maybe, but ..."

By the way, you'll meet all sorts of people who will angrily pound desks and declare that MIDI doesn't introduce latency, or doesn't introduce significant latency, and it's all audio buffering and so on - or vice versa. But the fact is that there's a world full of equipment out there, all designed with different constraints in mind. Until you've tested it, you just don't know.
By darkinprg Thu Nov 04, 2021 8:01 am
Koekepan wrote:There's a lot to test here, with different aspects of the system capable of causing latency.

Regular MIDI over a regular interface takes roughly 1ms for a single, three byte message (not all messages are three bytes). USB MIDI is also noted for delay counted around a couple of ms, so while it is fast, it is not instant. This means that anything going out through MIDI and coming back in should be expected to have some delay.

This is why I suggested that you test the Force just talking MIDI to itself, to figure out how much latency results from just that signal transmission and reception.

It is also possible that an audio engine would add latency; after all, there are factors such as buffering involved. However, to test that you would first have to make sure that you had two devices triggered simultaneously, perhaps by the intelligent use of a MIDI splitter.

The short answer is: "Yes, maybe, but ..."

By the way, you'll meet all sorts of people who will angrily pound desks and declare that MIDI doesn't introduce latency, or doesn't introduce significant latency, and it's all audio buffering and so on - or vice versa. But the fact is that there's a world full of equipment out there, all designed with different constraints in mind. Until you've tested it, you just don't know.



Cheers,
do you have any tips / suggestion on how to test this midi loop?


I tested this so far:
Force master, Octa slave, both audio into ext mixer = Force drums play slightly later then midi synced devices.

Octa master, Force slave, both audio into ext mixer = timing seemed correct

Having the other device returned on internal/usb audio input introduces again some latency which then only "deepens" the offsyncing...


I took it for a jam yesterday with 3other guys/setups. There was also MPC One, but the guy had it new, and slaved to the rest, so he didnt know if that happens too with One...
After a while of struggling i had to use force only as master clock for all of us, fancy master clock :)...

I will test it probably today against Ableton link, and how the devices behave then...

I am aware of audio buffers etc, thats why i think, maybe some compensation parameter can be added to input/output (either midi and audio) so these could be tweaked somehow ..

My idea was to replace octatrack, maybe even ableton partially, but with these "features' it kinda doesnt fit that idea...

***It is unbelievable that still today, octatrack wasnt succeded on the realtime sampling field... :worthy:
By andreat668 Thu Nov 04, 2021 9:24 am
darkinprg wrote:Cheers,
do you have any tips / suggestion on how to test this midi loop?


I tested this so far:
Force master, Octa slave, both audio into ext mixer = Force drums play slightly later then midi synced devices.

Octa master, Force slave, both audio into ext mixer = timing seemed correct

Having the other device returned on internal/usb audio input introduces again some latency which then only "deepens" the offsyncing...


I took it for a jam yesterday with 3other guys/setups. There was also MPC One, but the guy had it new, and slaved to the rest, so he didnt know if that happens too with One...
After a while of struggling i had to use force only as master clock for all of us, fancy master clock :)...

I will test it probably today against Ableton link, and how the devices behave then...

I am aware of audio buffers etc, thats why i think, maybe some compensation parameter can be added to input/output (either midi and audio) so these could be tweaked somehow ..

My idea was to replace octatrack, maybe even ableton partially, but with these "features' it kinda doesnt fit that idea...

***It is unbelievable that still today, octatrack wasnt succeded on the realtime sampling field... :worthy:


I am going to share my experience about Force timing issue:

I bought this machine end 2019 with 3.05 os.
At that time I spend all summer in Italy (Rome), the temperature got crazy hot and humid, specially July and August, temperature went up to 36 celsius degrees and 85% of humidity.
Now I am telling you the Force were struggling allot during this period, If temperature reached around 33 celsius degree up, the machine would get scary hot next to the touchscreen on the left side.
Also after 30- 45 minutes of normal operation the timing would get super wild, like sequences starting a bit later, some glitches, etc..
Fortunately I found a workaround for that which was basically producing just at evening/night and switch off the Force after a hour of operation and wait 15 minutes.
After I moved to Germany, I updated to 3.06 and I never had this kind of issues anymore.
From this experience I realised Force may act weird (OS sluggishness, wild timing, instability) when used in an environment with extreme hot weather conditions.
Also this is probably due to the ''passive'' cooling system architecture of the CPU installed into the Force.
Thats why the Force, the Live mk2 and the One got huge open grids on the sides. I guess the MPC X is less prone to overheating due to its ''spacious" layout and bigger form factor.
So keep in mind high temperature is something you want to avoid while using this machines, specially if you want to use them as master clock.
By darkinprg Thu Nov 04, 2021 2:37 pm
andreat668 wrote:
darkinprg wrote:Cheers,
do you have any tips / suggestion on how to test this midi loop?


I tested this so far:
Force master, Octa slave, both audio into ext mixer = Force drums play slightly later then midi synced devices.

Octa master, Force slave, both audio into ext mixer = timing seemed correct

Having the other device returned on internal/usb audio input introduces again some latency which then only "deepens" the offsyncing...


I took it for a jam yesterday with 3other guys/setups. There was also MPC One, but the guy had it new, and slaved to the rest, so he didnt know if that happens too with One...
After a while of struggling i had to use force only as master clock for all of us, fancy master clock :)...

I will test it probably today against Ableton link, and how the devices behave then...

I am aware of audio buffers etc, thats why i think, maybe some compensation parameter can be added to input/output (either midi and audio) so these could be tweaked somehow ..

My idea was to replace octatrack, maybe even ableton partially, but with these "features' it kinda doesnt fit that idea...

***It is unbelievable that still today, octatrack wasnt succeded on the realtime sampling field... :worthy:


I am going to share my experience about Force timing issue:

I bought this machine end 2019 with 3.05 os.
At that time I spend all summer in Italy (Rome), the temperature got crazy hot and humid, specially July and August, temperature went up to 36 celsius degrees and 85% of humidity.
Now I am telling you the Force were struggling allot during this period, If temperature reached around 33 celsius degree up, the machine would get scary hot next to the touchscreen on the left side.
Also after 30- 45 minutes of normal operation the timing would get super wild, like sequences starting a bit later, some glitches, etc..
Fortunately I found a workaround for that which was basically producing just at evening/night and switch off the Force after a hour of operation and wait 15 minutes.
After I moved to Germany, I updated to 3.06 and I never had this kind of issues anymore.
From this experience I realised Force may act weird (OS sluggishness, wild timing, instability) when used in an environment with extreme hot weather conditions.
Also this is probably due to the ''passive'' cooling system architecture of the CPU installed into the Force.
Thats why the Force, the Live mk2 and the One got huge open grids on the sides. I guess the MPC X is less prone to overheating due to its ''spacious" layout and bigger form factor.
So keep in mind high temperature is something you want to avoid while using this machines, specially if you want to use them as master clock.


Hey, and cheers...
Yes, it is a computer no matter what, so "computing issues" may be there, maybe even overheating may have some effect on is processing...
But that is quite not desirable behavior, if you understand what i mean...

And i probably dont have the nerve to spend this coin to buy a device, and wait if the next update sorts this out...

Important is if it fits into someones workflow...
If yes, only good... Others may not "be that lucky" :)...

A friend of mine just confirmed that his device is also having these timing issues I described above... But it fits his workflow as he runs sequences from force into synths, and rarely plays synced to other ppl...
If there isnt anything else around, you probably wont even notice this...

Peace
User avatar
By Koekepan Thu Nov 04, 2021 3:00 pm
OK, for testing the MIDI loop thing:

Make three tracks. Two drums, one MIDI. One drum track driven directly, the other listening for incoming MIDI on channel 1.

On the MIDI channel, send signals on MIDI channel 1.

Connect the MIDI Out and In ports with a single MIDI cable (and the little dongles, of course) and drive the MIDI drum track with the MIDI channel.

Compare the MIDI-cycled track's timing with the directly sequenced track.

Profit!

Wait, no, victory!

Well, at least success!

... or at least, learning.
User avatar
By Koekepan Sat Nov 06, 2021 6:11 pm
Whoa! Slow down there, you're leaping to an unwarranted conclusion.

You're adding a MIDI reception loop, using old-school MIDI which introduces several layers. It was never going to be 0ms! Not within the laws of physics!

The goal is to discover where the latency comes from, not to magic it out of existence. This is no reason to return anything. We're still investigating.
By pitchblender Sun Nov 07, 2021 1:51 pm
Here's a suggestion:

Try doing the same MIDI loop test on the Octatrack, it'll be interesting comparing the two. As Koekepan said there definitely *will* be latency (as we don't have quantum MIDI as yet :lol: ) but it's worth comparing. You could also try it with USB MIDI as I've heard mixed messages about its reliability and speed compared with DIN MIDI.
I haven't experimented with Force latency with other equipment. Was going to but...
I've ended up getting the Holy Grail this week - E-RM Multiclock - for a 'good' price with almost every penny I had. Tried to buy food today and card got rejected! Had to hand back some til it authorized... so I can't eat well til next month but my stuff is going to be in sync!
Will update as and when. Doing that Carl Craig remix competition so no time at the moment...

Cheers
User avatar
By Koekepan Sun Nov 07, 2021 11:36 pm
I did a test today, using a Behringer Neutron as a sound source. I sequenced it over USB from the Force, just to get a series of pings with sharp initial hits for latency measurement.

The "phones" output I sent directly to my Tascam Portastudio. That was my latency baseline.

The "main" output I sent through a Uphoria 1820, over USB, into the Force (obviously through internal buffers and so on, but no USB hub in the way) then back out through the Uphoria 1820, and into the Portastudio.

I recorded some audio, then copied the files and inspected them closely in Audacity.

The measured distance on the waveform display was 16ms.

That's honestly not bad, and it was nice and consistent. This is easy to correct using an in-line delay, and probably wouldn't even be a problem for me in most use cases. If it is a problem in a particular scenario, I just shift one track a few milliseconds back, directly on the portastudio. Done.
By darkinprg Tue Nov 09, 2021 4:43 pm
Koekepan wrote:I did a test today, using a Behringer Neutron as a sound source. I sequenced it over USB from the Force, just to get a series of pings with sharp initial hits for latency measurement.

The "phones" output I sent directly to my Tascam Portastudio. That was my latency baseline.

The "main" output I sent through a Uphoria 1820, over USB, into the Force (obviously through internal buffers and so on, but no USB hub in the way) then back out through the Uphoria 1820, and into the Portastudio.

I recorded some audio, then copied the files and inspected them closely in Audacity.

The measured distance on the waveform display was 16ms.

That's honestly not bad, and it was nice and consistent. This is easy to correct using an in-line delay, and probably wouldn't even be a problem for me in most use cases. If it is a problem in a particular scenario, I just shift one track a few milliseconds back, directly on the portastudio. Done.


Honestly not bad? You dont mean that right?
Do you produce music? :)) (Joke)
Can you imagine hihats which are shuffled, effected with delay playing together with other rhytmic stuff? Its like the sound of breaking glass...

You guys seem like youre not understanding its offsync in 9of10 scenarios of how you could use the force...

And then you're telling me I should not return it, well i would definitely not retun it if it worked as expected/advertised...

I have returned the force already yesterday..
Bought digitakt, everything is synced as is expected ;)

I work in a music shop, and I will be informing and openly NOT RECOMMENDING Akai devices to our customers because of this...

Peace
User avatar
By Koekepan Tue Nov 09, 2021 5:27 pm
You seem to have missed the parts about the laws of physics.

But you know, what, you're the ultra-genius here, why don't you replicate that loop comparison on the Digitakt? And the Octatrack? See how stable it is, and see what the total latency is? Report back to us about how reliably instant it is.

Go on. Show us the truth.

Edited to add: now that I think about it, give us both the audio loop latency that I demonstrated above, and the MIDI loop latency that I proposed.

For those following along at home: the Force does 44.1KHz audio, which means that an audio buffer of 256 samples (fairly middle-of-the-road setting) will get you an immediate 6ms regardless of anything else.

Also, just for yuks, I did the MIDI loop test and my result was just about spot on 6.5ms. For a MIDI send, transmission, reception, interpretation and application? That's not crazy bad. Jitter was way lower; this was pretty consistent.

I really think darkinprg doesn't actually understand this stuff.
By darkinprg Wed Nov 10, 2021 10:12 am
[quote="Koekepan"]You seem to have missed the parts about the laws of physics.

But you know, what, you're the ultra-genius here, why don't you replicate that loop comparison on the Digitakt? And the Octatrack? See how stable it is, and see what the total latency is? Report back to us about how reliably instant it is.

Go on. Show us the truth.

Edited to add: now that I think about it, give us both the audio loop latency that I demonstrated above, and the MIDI loop latency that I proposed.

For those following along at home: the Force does 44.1KHz audio, which means that an audio buffer of 256 samples (fairly middle-of-the-road setting) will get you an immediate 6ms regardless of anything else.

Also, just for yuks, I did the MIDI loop test and my result was just about spot on 6.5ms. For a MIDI send, transmission, reception, interpretation and application? That's not crazy bad. Jitter was way lower; this was pretty consistent.

I really think darkinprg doesn't actually understand this stuff.[/quote]

Look maybe i will do that octa digi loop to show you ...because i dont have any syncing issues now, WHEN NOT USING FORCE...

Im almost 40, an IT guy, i work with daw and synths for over 20yrs now, and im called in when theres deeper questions from customers of our shop. So my with technical knowledge on this topic i feel pretty confident.
Lets put all that aside...
Explain to me what i seem not to understand
- i use some devices, i replace one with Force and all my synths are out of sync.
When i jam with other setups, i take force with me, im out of sync...
So what am i getting worng here?

I use ableton as main hub, all my devices sit on grid, because ableton lets you to tweak midi latency for each midi output, or there is an external instrument device, which lets you tweak audio signal.
When i use octa as mixer/hub i dont have any out of sync issue, i can jam with others with no syncing issues ..

Theres only one thing at fault, wnd that is me using Force.



I understand there is some audio buffer required, im a computer guy, those numbers you state, yes impressive, it is less then my computers latency is. But i had other devices, which did not had any issue...
VMaschine, Octatrack, Toraiz squid, Circuit, Electribe2.. Non of these had any issue, i place Force on desk aaaaaand its off sync...

I may be nobody to you, to everyone, but as a customer i dont need to understand and accept someting because of some inner architecture it has... AS A CUSTOMER I AM INTERESTED IN PLACING THE DEVICE ON MY DESK AND HAVE FUN, and nothing you will write or say will change that Force timing is off, unless youre a force firmware programmer and will actually write a fix for this...

By now based on your replie, youre just a nickname (worse if akai moderator) on a forum. Biased about a product, which has its flaws...

"For me force did not work because i have more devices on my desk", and i will make sure ppl around me interested in akai mpc/force products will know about these flaws...

Based strictly on your response (no matter if your akai or not) i will not be buying Akai midi mix, but will go for novation controller instead...

I had my time on product forums, this was the worse experience, and you on behalf of Akai showed me to go away from Akai products...

Thank you, have a great offsync day... :roll: