Arturia Forums
Hardware Instruments => MatrixBrute => MatrixBrute - Technical Issues => Topic started by: digidisiggi on January 30, 2017, 06:04:46 pm
-
i'm just asking this because every day in my studio i switch the MB on and the synth is out of tune (e.g. preset A15, the lower sound is about 1-2 semitone higher then the upper sound). The temperature of my studio is 24/7 stable an about 20 degree celsius.
So i have to use the autotune shortcut many times. After about 5min the MB is finally getting in tune. Is it just mine?
I have many analog synth (A6, Matrix12, Moog's etc.) but no such problems with tuning the synth.
Sometimes after autotuning the Brute, the presets sound totally different. So i have to switch off and on the synth again.
Any idea what i can do? How can i check if the MB is not ok?
best, Siggi
-
It's the same here with mine...
-
I will report later this week, when I get mine, but have you tried to let the synth stabilize for about 15-30 min before running any autotune? My Andromeda always requires about 15 min before it is playable and in tune, until then it is really badly out of tune. The good thing about the MatrixBrute is that it is monophonic, so the tuning should be less of an issue than with polyphonic synths. Do the oscillators end up stabilizing to different frequencies at different startups from cold, even when letting the synth to stabilize to normal operating temperature?
Nevertheless, I am also interested in what kind of temperature compensation the MatrixBrute has, or does not have.
-
Mine seems pretty stable even from cold, I haven't felt the need to run run the autotune yet.
-
Ok, that sounds good. One point I did not mention in my previous post is that I would never run the autotune of a synth before the synth has reached normal operating temperature, i.e. before 15-30 min since power on. Doing so right after startup could definitely result in an out of tune synth when the oscillators reach normal operating temperature. I once ran the autotune of my Andromeda very soon after startup and as a result lost one of the voices and it required the full long factory tune to get it back. Since then, I never ran the autotune ever again. Of course, a different synth, but still. Back then, I was not used to analog VCO synths. In any case, I wish the two synths reported above will just need a good warmup period + autotune. Even then, I would expect some drift and variance in sounds. That is part of the VCO magic. Let us know...
-
i'm just asking this because every day in my studio i switch the MB on and the synth is out of tune (e.g. preset A15, the lower sound is about 1-2 semitone higher then the upper sound). The temperature of my studio is 24/7 stable an about 20 degree celsius.
So i have to use the autotune shortcut many times. After about 5min the MB is finally getting in tune. Is it just mine?
I have many analog synth (A6, Matrix12, Moog's etc.) but no such problems with tuning the synth.
Sometimes after autotuning the Brute, the presets sound totally different. So i have to switch off and on the synth again.
Any idea what i can do? How can i check if the MB is not ok?
best, Siggi
I have very similar comments to this. On start up, tuning is a big issue. Interestingly, VCO1 and VCO2 are close (VCO2 slightly sharper), but VCO3 is about a FULL tone out. Auto-tune seems to sort it out pretty quickly. Has been OK after 5 minutes or so.
And yes, the presets do seem erratic too...
-
i will check tomorrow what happens after running MB without autotune. But i guess the MB has no background tuning like the Andromeda. I think my MB is faulty. It canīt be in the year 2017 the analog oscillators are such unstable. My DSI OB-6 and P6 also has no tuning issue at all. I really would love the MB but the tuning issue kills my euphoria. Iīll report.
-
I just checked mine with a tuner.
I used the master tune knob to get the low C on osc 1 in tune and then tested the range and the other oscillators. My master tune is at around 1:30 for this.
Over the full range of the keyboard OSC 1 note C varies by 5 cents.
OSC 2 on LOW C is 17 cents sharp of OSC 1 and varies by 3 cents over range of keyboard.
OSC 3 on LOW C is 35 cents sharp of OSC 1 and varies by 3 cents over range of keyboard.
So there is some tuning difference between the OSCs but they stay in tune extremely well over the keyboard range.
-
@BobTheDog: you are a lucky guy and it seems you received a good copy of a MB. Thanks for your report.
best, Siggi
-
When I turn it on tomorrow I will check from cold and see if it's much different.
-
Ok just tested from cold.
At power on OSC 1 and 2 are around 1.5 semitones flat while OSC3 is in tune.
After 5 minutes OSC 1 and OSC 2 have come into tune
All oscillators are around 10 cents flat from the readings I was getting at end of day yesterday.
So it looks like the exponential oscillators need at least 5 minutes to warm up.
If you mess around with the autotune in this period you are looking for trouble,
If i was you I would turn it on, wait 30 minutes and then do the autotune. Check its all in tune with a tuner or by ear.
After that the next time you turn it on wait at least 5 to 10 minutes (do not use autotune) and then check the tuning with a tuner or ear, my guess is that it will then be fine.
-
...
If you mess around with the autotune in this period you are looking for trouble,
...
Don't the autotune simply tune the oscillators whenever you hit it, or do things get irrecoverably out of whack if tuned when the oscillators aren't yet stable?
-
Yes it does.
I'm sure it doesn't go irrecoverably out of whack (I haven't tried autotune when it is warming up), I was just trying to get across the fact that if you do the autotune in the warmup period OSCs 1 and 2 are going to go quickly out of tune as they are gradually going sharper, so from a users perspective its "Hey I just tuned it and now its out of tune again".
-
Also on warm up OSCs 1 and 2 stay in tune with each other, the real discordance is noticeable only if you have OSC3 going as well or trying to play in tune with something else :)
-
OK, thanks. Was worried for a moment there :)
Whatever Moog did with the Sub 37 should be the model for VCO stability and calibration... :-p
-
I might have been a little too melodramatic in my wording!
-
Ok just tested from cold.
At power on OSC 1 and 2 are around 1.5 semitones flat while OSC3 is in tune.
After 5 minutes OSC 1 and OSC 2 have come into tune
All oscillators are around 10 cents flat from the readings I was getting at end of day yesterday.
So it looks like the exponential oscillators need at least 5 minutes to warm up.
If you mess around with the autotune in this period you are looking for trouble,
If i was you I would turn it on, wait 30 minutes and then do the autotune. Check its all in tune with a tuner or by ear.
After that the next time you turn it on wait at least 5 to 10 minutes (do not use autotune) and then check the tuning with a tuner or ear, my guess is that it will then be fine.
so, yesterday i started the MB from cold and had the same experience.
- do not use the autotune command in the first 5 min. Things get even more worse.
- also donīt use autotune when MB is in tune. Sure, there is no need for this. Sometimes (not always)after tuning the preset sound doesnīt come back (no sound) and you have push the preset again. Donīt be surprised, all the presets of the MB now sounding totally different. No further autotune action can correct this behavior. You have to switch it off and on. Then everything is normal.
If my MB is not defect, i would say it has the most badest autotune implementation compared with my many other analog synth. Hopefully this can be adressed with a firmware update.
On the other side i have to say (if the Brute is in tune), itīs a fantastic synth and imho the brute has character, is easy to use and very well build.
Siggi
-
After tuning and getting no sound make sure the "Panel" button is not lit, if it is tap it so the light goes out.
Does the sound return then?
-
After tuning and getting no sound make sure the "Panel" button is not lit, if it is tap it so the light goes out.
Does the sound return then?
no. Tweaking knobs doesnīt help either.
-
Have you tried the master volume, I have had it where this can change. Try taking it all the way down and all the way up?
Edit just tried repeatedly autotuneing, after 4 times I got silence. The master volume trick fixed it.
-
Hello,
I guess this topic needs a bit a clarification because it's a bit complex.
VCO 1 and 2 are true exponential VCOs. That means analog circuitry is doing the linear CV (1V/octave) to exponential frequency(double the frequency for each octaves) conversion. This kind of converter always have temperature dependency so it needs to be either compensated, either temperature stabilized. If dependency is not voided properly the slope will be wrong, for example 1 octave will not be 1 octave, but 1 octave and 10 cent.
So if C3 is a perfect C, you will have C4=+10cents, C5=+20cents, C6=+30cents ...
For compensation it is usually a Tempco that is used. A Tempco is a kind of resistor that has a known temperature dift. If you pick a dift value opposite to the VCO dift, it ends up more or less stable. More or less because both have tolerances in their drift value, plus tempco and the VCO components need to be perfectly thermally coupled otherwise they are not at the same temperature.
Using tempco has an advantage and a drawback: they drift very slowly, but drift even when not powered. That means that if a machine stays (even powered off) always at the same room temperature it will be very fast is tune. But if you expose the machine to a temperature change, it will be very long to recover. So it is very difficult to use it in live situations
When using stabilization you use some kind of heater and sensor that ensure the VCO core always runs at a constant temperature much higher that room temperature.
At power up this oven needs few minutes to warm up but when it's locked, it can handle room temperature changes( in a reasonable extend of course).
MatrixBrute VCO1 and VCO2 (and MiniBrute and MicroBrute) are oven based exponential VCOs. That means you have to let them warmup up to 5 minutes. Once locked the slope will be fine (1 octave is a true octave).
But then comes the pitch offset issue. When the octaves are right, you now want the notes to be perfect. The offset is what you basically control with fine tune.
Because it is analog, all the components has a little temperature dependency(much lower than VCO core slope), so all the internal CVs have a bit of temperature dependency too. It's definitely not noticeable to control a VCA, but for a VCO pitch we have around 2 cents per °C .
Usually you want to be in perfect tune with other instruments with fine tune at center position. That is the purpose of "AutoTune" in MatrixBrute. It voids the pitch offset to ensure that absolute pitch is right.
So the good way to use the autotune is : wait at leat 5 minutes for the ovens to warm up(using autotune before that will be point less, because the slope will still evolve).
Once ovens are locked (1 octave = 1octave), then you may use auto tune to void any pitch offset.
At constant room temperate, you may have to autotune again during the first 30minutes (time to get a stable and homogeneous internal temperature).
if room temperature changes, then autotune again...
By the way we have no background autotune, for example when the VCA is closed, because as we've got external CV inputs for the VCA and VCO we don't even know if the VCA is closed, or if the pitch is being controlled by an external source...
OK but what about VCO3?
This one is a linear VCO, that means it does not do the exponential conversion itself, it is fed with a CV that is already exponential (conversion is made by firmware). No analog expo converter means almost not temperature drift, so at startup it is almost immediately in tune. And it does not need to be autotuned... Magic!
But (because their is always a but :), no expo converter means that this kind of VCO can't do exponential frequency modulation based on linear CV, so you cant' directly control it from an external CV. If you feed it with another VCO on the CV input, you will get linear modulation of the pitch instead of exponential (what you get on matrixBrute in the AudioMod section)
You can still use a microcontroller to sample the external CV input, make it exponential and generate a new expo CV, but then bandwitdh is limited by sample rate, so again no full audio range modulation available...
That is why VCO3 is only a source of AudioMods and never a destination.
-
Whatever Moog did with the Sub 37 should be the model for VCO stability and calibration... :-p
I think it's same kind as MatrixBrute VCO3...
-
thanks Bruno for the detailed explanation and clarification to the tuning topic. This helps to treat the MB right.
best, Siggi
-
If my MB is not defect, i would say it has the most badest autotune implementation compared with my many other analog synth. Hopefully this can be adressed with a firmware update.
We are currently investigating but it seems behavior could be random in some case, looks like bug indeed. we will let you know
-
...
OK but what about VCO3?
This one is a linear VCO, that means it does not do the exponential conversion itself, it is fed with a CV that is already exponential (conversion is made by firmware). No analog expo converter means almost not temperature drift, so at startup it is almost immediately in tune. And it does not need to be autotuned... Magic!
...
Can it still be autotuned or calibrated in some other way (other calibration routine, trimmer, ...)? It's analogue so it could be offset relative to the autotuned VCO 1 and 2, or...(?)
Re Sub 37 tuning:
I think it's same kind as MatrixBrute VCO3...
OK :) With regards to the above, the Sub 37 has a calibration routine, and it is needed sometimes.
-
Thanks Bruno for explaining the details of how the tuning control is implemented in the MatrixBrute. As I mentioned earlier, I have got used to this with the Andromeda that out of tune is expected for the first 15 min. Also my MFB Dominion 1 is badly out of tune for the first 15 min too, but gets in tune eventually. Will get my Brute today, so I can report in the following days, how that unit gets / stays in tune. I have no problem of slight variation +/- 3 cents of the tuning. I buy this synth for character. Let's see...
-
so it has no background tuning ? pretty strange, as i've noticed some funny things too about tuning, i'll write today a detailed post about all things that i've noticed on my unit
-
Ok, I have now a couple of hours experience with the MatrixBrute. I must mention that I have not yet updated the firmware. It seems that when the synth has been on for 1-2 hours (the next day after bringing the synth in the room, not from cold), and running auto tune, the Osc 1 and 2 are a few cents off from each other, when the tune knobs are in the detented mid positions. I need to adjust Osc 1 down by a few cents, maybe 5-10 cents, by ear to get the beating minimized. The beating is slightly random in nature, different from other synths that I have used. I still need to install the latest firmware, but I wish there is possibility to adjust the tuning so that after the unit has fully warmed up, the auto tune would result in same pitch for Osc 1 and 2, when the knobs are in their detented positions.
The whole autotune process is really fast, like 1 second. I am wondering, what does it check when doing it, and does it override the current knob positions for coarse and fine tune, while doing it, and then return back to the current knob or pitch values of the program loaded? It would be useful to have a calibration procedure for the coarse and fine tune knob center positions, where the user should first turn the coarse and fine tune knobs to the detented center positions, and then save these in the operating system and the auto tune function would use these values as target values when operating the auto tune. Now, I suspect that my MatrixBrute thinks that the pots are not in the center positions, even if they are, and apply tuning based on some preset values for the tune potentiometers, during which the synth checks some control voltages based on preset voltage or digital values, and not the actual voltages set by the tune potentiometers? I would expect that while having the coarse and fine tune knobs in the detented center positions, running the auto tune would result in pitch within +/- 1 cent between Oscs 1 and 2. Now they are 5-10 cents off right after running auto tune.
Bruno@Arturia, is it possible to calibrate the actual coarse and fine tune knob detented center positions by the user, so that the autotune would result in equal pitch for both exponential oscillators 1 and 2? If not, please implement this in a future firmware. I believe this is an issue with some of the MatrixBrutes at the moment. I will install the new firmware tomorrow and report, if anything changes.
Edit. Nevermind, I never actually entered the auto tune function properly, please see explanation below. All work ok now with latest firmware and succesfully running the auto tune!
-
Ok, a quick but important update about how the MatrixBrute auto tune function should be used properly. I just installed the new firmware. My unit shipped with a firmware with something like 1.0.096 or so. Now it has the 1.1.108 or so. I ran the autotune procedure again now with the new firmware... Or so I thought!! The same result, Oscillators 1 and 2 not in tune with each other. However, I found a solution that might be of help to many others as well!
The instructions in the manual to launch the auto tune procedure are not precise enough, or they kind of are, but depending on how you press the two buttons (Kbd Track + Panel), you may not even get to the auto tune process at all! I was already wondering, why did Arturia design the auto tune function so that every time you use it, you also switch between the Kbd Track and Panel modes. That must be a bug I thought, but the latest firmware does it as well! If you found the same issue when trying to run the auto tune, you did not actually auto tune the synth at all!
At first, I tried to launch the auto tune process by keeping the Kbd Track-button pushed down and shortly after that pressing the Panel button too. Something happened and the audio muted and unmuted, so I thought it ran the auto tune, nice. But at the same time the Kbd Track and Panel buttons switched states too. This happens with the original + updated firmware. This way the synth does not go into the auto tune mode, so no wonder why the "tuning" did not have any effect at all! I tried several different ways of pushing the two buttons, and found how you get to the auto tune mode every time, and do not unnecessarily switch between Kbd Track and Panel modes.
Proper way to start the MatrixBrute Auto Tune proceduce:
Method 1: Press both the Kbd Track and the Panel buttons EXACTLY at the same time and keep them pushed for 1-2 seconds. Then the display starts scrolling a text "Tune..." If the text does not appear, and instead your Kbd Track and Panel buttons switch states and audio mutes and unmutes (due to panel mode change), you did not enter the auto tune mode. Try again.
Method 2 (even better): Press and hold the panel button + then press the Kbd Track button too. Thanks BobTheDog for this method!
Now, my oscillators 1 and 2 are in the same frequency, when all coarse and fine tune knobs are in the center detented positions, nice! I will give an update later about how the tuning stays like that in my room. :)
-
I think the trick for any of the PANEL shortcuts is to press and hold PANEL then press the shortcut key, so for autotune press and hold PANEL then press "Kbd Track".
-
BobTheDog, you are right. I just tried this as well. You can enter the tuning mode either by pressing the two buttons exactly at the same time, or first holding panel and then pressing Kbd Track! In any case, I think it is good that we found this issue and now we have proper way to launch the auto tune mode. I am sure many MatrixBrute owners who face the synth being out of tune and think it is possibly faulty did not actually enter the auto tune mode nor saw the Tune text, but thought it was fine because the manual does not mention the text appearing. Arturia should add to the manual that you should see a scrolling "Tune" text on the display, when the auto tune does its thing. It takes about 2-3 seconds to complete, at least the text disappears then.
-
Related to tuning. There is slight variance in the result of the auto tune after running it multiple times, maybe +/- 2 cents between the two oscillators. If you program a typical lead sound with Osc 1 and 2 both having just a saw wave and levels almost equal, you get the typical phasing (or beating), because the two oscs cancel each other with slow beating. However, programming such a sound is sometimes difficult right after running the autotune, because to get such slow beating, you need to tune both oscs really close to each other. It seems that the detents in the fine pots make this slightly more difficult, because you easily get too far from the tune of the other oscillator. However, you can go past this issue by tuning both oscillators off the center, where the detent is not a problem. Then, you just use the master tune pot of the synth to set the correct tuning. However, this does not necessarily sound in tune with Osc 3, if you want to use it.
Another interesting thing I found is that especially after running the auto tune, and you are trying to find the very slow beating between the two oscillators, very often the synth stops the beating completely and kind of "sucks" like a magnet the frequency of the other oscillator to match perfectly the tune, like using sync even if it is not in use! You need to tune the oscillator again off and try to come back closer to the frequency of the other to get the slow beating. I find this behaviour peculiar. I am actually wondering, does the auto tune process end immediately after the text disappears, or does it continue on the background for a while, because there is often very strange random variation in the beating, especially after running the auto tune. The longer you have played since auto tuning, the easier it is to find the slow beating without the two oscillators getting to the exact same frequency like magnet.
Btw, I must mention that even though some of my posts may sound negative, I like the synth very much and am here just to solve these issues / understand the features for the best of all users, and some of the issues found so far are not even issues, but need to be better commented in the manual. This Brute is really special, has put smile on my face many times while discovering something new. It is definitely a character machine with all these little features.
-
...
Another interesting thing I found is that especially after running the auto tune, and you are trying to find the very slow beating between the two oscillators, very often the synth stops the beating completely and kind of "sucks" like a magnet the frequency of the other oscillator to match perfectly the tune, like using sync even if it is not in use!
...
Could be sympathetic synchronization (injection locking (https://en.wikipedia.org/wiki/Injection_locking)). Known to happen when several oscillators can minutely affect each other through adjacency and common circuitry.
-
I think the snap of the tuning is not an oscillator type thing, it is caused by the fine tune knobs snapping back to the dead zone.
You can show this by mapping the mod wheel or macro knobs say +3 to OSC 1 pitch and -3 to OSC 2 pitch, now use the mod control to set the beating between oscillators.
First thing you will notice is that the resolution is much higher than the fine tune knobs so you can get much slower beatings going on, also there is no snap back in tuning like you get with the fine tune knob.
P.S. The mod knobs values are saved with the preset so I am using this technique to set slow beatings between the oscillators and leaving the fine tune knob alone. Hopefully Arturia can add some kind of scaling to the fine tune knob to make it more "Fine".
-
Concerning autotune on my MB it doesn't seem too accurate, on the middle C I get:
Osc1 = -9.8 cents
Osc2 = -9.4 cents
Osc3 = 1.3 cents
So Osc1 and Osc2 are basically in tune with each other but out of tune with reality and Osc3.
I'd be interested to see other peoples results.
-
I managed to achieve the magnet like snap to the same frequency with the oscillators 1 and 2 also, when both were tuned far away from the detented center, so it must be something in the circuitry, not just the mechanical detents of the knobs. However, it does not happen always, but more often soon after running the auto tune. I also find that the tuning drifts after a while, so that it is actually easier to set this kind of slow beating sound with oscs 1 and 2, letting the fine tune of one of the oscs be in the detent spot, while the correct tuning of the other can be found outside of the detented spot. Run the auto tune and you are dealing with the detents again.
-
Thats interesting, I must admit I haven't heard the snap using the mod matrix approach, maybe I have been lucky. Why don;t you give a try and see.
One thing I have noticed is that an oscillator can change tuning as if they have a square wave modulating them, haven't got to the bottom of this and it only seems to happen in the first 30 minutes after power on.
-
One thing I have noticed is that an oscillator can change tuning as if they have a square wave modulating them, haven't got to the bottom of this and it only seems to happen in the first 30 minutes after power on.
This could be also the same as I found interesting random pitch variation of the two oscillators that can be easily spotted by listening to the beating / phasing of two saw waves, but not by listening to 1 oscillator alone. It indeed sounds a little bit like square wave modulating, trying to compensate some voltage during warmup period. I believe the same can kick in after running the auto tune. Because I can hear this behaviour more often soon after running the auto tune, I think it still keeps working on the background even after the Tune text has gone away, and actually stops doing its thing after a few minutes.
Would be interesting to hear from Arturia about how long does the auto tune actually keep working, or should it be over after 3 seconds, when the text goes away too?
-
I Just did a test, repeatedly autotuneing and checking the tune of OSC1, there are the tuning I got in cents middle C
-13
-7
-10
-9
-22
-9
-16
-15
As you can see a fair amount of variation, there must be something going wrong here.
-
Managed to get the square wave type modulation tuning error using the matrix approach when the control is at 100% and the mod depth is set to 3, changing the mod depth to 5 and lowering the control value made it go away.
Using one OSC: https://soundcloud.com/bobthedog/mb-strange-osc
-
Did you check how the two oscillators tune in the repeated test in addition to just 1 osc? I have not measured the tunings of the oscillators other than listening the beating by ear. There are slight variations in tunings also between the two oscillators after each autotune, but if they both go such a large amount in the same direction, it is more difficult to hear.
-
Osc 1 and osc 2 stay roughly in tune with the varying results, so both are off by nearly the same amount.
-
BobTheDog, your idea of using one of the macro knobs for osc fine tune is quite useful. I will probably set VCO2 fine for M1 always, if not needed for something else. Then, it can be quickly tweaked to get suitable beating, if the tuning tries to do its own thing. Btw, I am fine with the synth even if the tuning is a little bit wild as it is now. There are workarounds to deal with some of it, and it makes the synth sound more unique and organic. However, I still want to see what Arturia can do to improve the consistency of the auto tune and possibly explain why the oscillator frequencies sometimes just snap together when they are close, and if that could be disabled. I would much prefer freely drifting oscillator frequencies in that case, not trying to force sync them.
-
YES !!! Injection locking! :o
As a child in the 1970's I had several discrete NE 555's flashing their own separate LED's. Each NE555 had different value peripheral
timing components but all the LED's ran in sync!
I wondered what the mechanism was back then , no one was interested enough to answer my question.
This sync was absent when the chips had their power supply isolated from one another .
I should have tried inserting a blocking diode at various points. :-\