Strange issue when changing parts

[Peter_Peak]

Thank you Brandon, I slowed down Cenk’s video and noticed deleting trigs. Also on MK1 volcas video…

Which makes me feeling down again about sophisticatation of Octatrack (despite wonders it already makes). I was let down when I found that sliced sample is not played back without start/end? clicks, even when there was no processing (straight recorded voice, played back through linear slice trigs, sliced by the slice grid which seemed nice continuous in editor). What looked mathematicaly, logically possible in this quite hi-end digital world and would be evidently valued, is non usable without other masking sounds. I would not argue if it occurs when mangling, but clicking in base state at slice trigs which should/could be totally exactly distributed, is dissapointing.
But maybe I’m again missing something, or this extreme, but promising and I think possible Octatrack feature (of lots of slice triggers) would not be needed in real world-but that’s against philosophy of possibilities and experimentation -mangling from calm, realistic state to complex sonical reshaping- for which Octatrack seemed the right tool for me.

However, this mirroring of parameter locks to other values of other machines, based on common place in menu is illogical, almost silly, compromising value of the parts, and probably not usefull for anything-however we can bypass that with more keypresses, more patterns, more workarounds. Maybe I could embrace clicks even there are no workarounds(?), with bad feeling that this is not so ultimate device (and haven’t come across other maybe more common and more standing out issues yet), but this locks seems no quirk/trademark for me-just error.
And I’m wondering, too, why this is not issue for people to which it can save a lot’s of time, when this Thru-Flex mangling has quite big attraction power for buying Octatrack

[Open_Mike]

Whats happening is that plocks are stored in the pattern, not the part. The plock will reference the same knob position if the machine is switched manually, or from changing part while in a pattern.
If you link the next part to another pattern, then you can remove the plocks and add different ones, while the old ones remain in the other pattern/part.

It’s very possible to not have clicks, and the machine indeed works.
For linear slice locks with trigs right in a row to play an entire sample, try setting flex setup Len to slice, and flex main Len to a higher value to avoid clicks…

It’s necessary to really let the machine sink in as most of what people think are bugs at first are intended behaviors that can be utilized once understood…
If you keep thinking it should work one way it will never make sense. You have to just learn how it works, and by practicing and experimenting it clicks and makes sense. The most common OT issue is not understanding the OT and thinking it’s messed because it should work some certain way, but once you understand it it works perfectly the way it is and probably with more options than if it worked the way one thought it should.
It takes some time, try not to get frustrated with it, or maybe plan on getting frustrated for a bit.
if you can put up with learning it, you’ll find the reason lots of us use and love it… -

[Peter_Peak]

Thanks for optimism, confirming existence of problems and workaround suggestions

I’m quite critical on any product I get my hands on, but I also have respect for learning curve of Octatrack. However, making critique of Octatrack is like reviewing art film vs. reviewing nice commercial movie (i.e. beatstep pro). I review art films with different demands than commercial ones. That’s why art film can get lower score than commercial one, even side by side there’s little to compare.

I still don’t know how to change patterns by changing parts (if it’s possible), but that leads me to another possible complication - if you’ll have two patterns in setup, which are simulating/workarounding one pattern without that (i)logical mirror bug, wouldn’t you have every other tracks duplicated too, which you can’t (or can you?) trig and plock change at the same time?

As for clicks in a row played sliced track, I believe I tried every possible parameter combinations, fades, making grid slice end points longer etc…, but I’ll try harder if you could get it clean.

[Open_Mike]

Yes I understand your frustration, when learning the OT I had quite a few jams stopped dead in their tracks and sent me into hours of troubleshooting. Not fun, but I’m entirely glad about keeping at it as now it’s a wonderful hub for my music.

Without upsetting you I just want to tell you that whatever you think right now is most likely extrapolated from not understanding the interactions of layers of the machine, and your jumping straight to its a “bug”. I don’t see the plock/knob/part interaction as a bug, I see it as how the machine is designed, and program things accordingly understanding that that’s the way it works.

Honestly I wouldn’t cast judgment on the machine until you fully grasp it, unless your judgment is simply that it’s hard to grasp.
It’s different than other gear but actually very well thought out and makes a lot of sense once understood. Seriously, for me lots of wtfs about the OT eventually started clicking, and then I’d realize why it’s like that and that it’s actually better that way.
Again I understand it can be frustrating, but I always try to encourage folks to make it through this part, because then it can really get fun!

[Open_Mike]

Jeez you all, I don’t know how I got to be the OT spokesman, I just read the forum a lot and frustrations keep coming up so I try to encourage folks to get though it… But it happens a lot…
I’m not some die hard OT defender, I just want people to have fun making sounds with it like I do…

[barnacles]

It often comes down to “I want the OT to do it like this” vs “the OT works like this”

Learn what it does and use it accordingly!

[Peter_Peak]

Guys , I don’t want Octatrack to be something else. I work with difficult audiovisual products and creative concepts, I like it and I’m not afraid of Octa. I just don’t want to call bugs with noble adjectives. Together with mk2, as Elektron planned, here’s new generation of users and it’s also good for experienced ones, if new pressure is made for correcting some issues (I would guess, that conditional trigs are more change of mind after the first reaction on mk2, rather than surprise gift).
I was observing Elektron and Octatrack almost year before I decided to go for it-luckily just at annoucement of mk2. I’m in game industry and I know how difficult is to repair code in relatively small team, when bandwagon of new projects is forcing to abandon not perfect products.

But also I see prestige and uniquity of Elektron-let alone in community, in devotees. I understand that their boxes are little miracles which are pushing big veteran companies into uninspiring recyclers of their own past.

Now, please, could any of you experienced users say, pragmaticaly, if there would be any possible negative effect of making this mirroring bug gone? To be sure: mirroring bug is when plocked value of one machine is mirrored/reflected, changing completely different parameter value of another machine, when changed with part command. Only common attribute is, that they are in the same menu spot.
It happens i.e. when you change parts, flex to thru machine, with trigers plocked in flex (or vice versa in thru, when you plock volume-it will change slice start in flex).
Would it affect negatively any of OT function you use when it will not be mirrored like this?
Is there any reason or advantage of it?
Would it make things better for your work?

I’d write request to electron for this (everybody else is welcome to do the same) and I want to include as much arguments as I can accumulate. Please don’t take me as “disturber of eternal peace” . I just want to take chance of this period, when OT firmware is still open. I don’t want better timestretch or more features-these are true character of instrument, I just see that this could be pain in the axx in what I’m planing to do with OT, also effectively limiting part feature. I’m sure I’ll find another bugs (or whatever we can call it ), but this is actually the most obvious for quite basic role I have found.

Thanks for understanding!

[Open_Mike]

I wouldn’t mind if parts had there own plock relationships to patterns, it just seems rather difficult to implement with the current design of the unit. Go ahead and feature request if you’d like, I’ll take it if they could work it out.

Since plocks are stored in the pattern, it seems you would need something like 4 pattern variations per actual pattern, one for each part.
If one part is selected and the sequencer is running through plocks, somehow the device would need to know when you changed parts to not use those plocks and possibly use other ones. Since plocks run through different values per step, they have to be associated with the moving sequencer pattern and not just a single value in the part itself. So there would need to be 4 available pattern variations within each pattern or a more drastic redesign of how the machine operates.

This stuff has to be thought about since we’re dealing with a unit that’s been out for awhile, coded a certain way, and is probably low on available memory and not in a state to have drastic code revamps done to the firmware.(At least with mk1, and very likely mk2)

Currently we just link a part to another pattern to do this, it does use another pattern and in some cases the patterns must be very similar with only minimal changes to achieve things, but we have 256 patterns in our 16 banks of 16 patterns each with 4 parts, and there’s always more projects. You can set pattern change to change pattern/part down to one step.
The functionality you would like currently exists in a paradigm where you need to copy and adjust patterns, and have your part switch after 1 sequencer step instead of instantly.

I’ll add here that I don’t think many OT users switch parts in the middle of a pattern, I’ve been reading this forum for years and nobody ever mentions this. All my part switches are quantized with pattern change…

Most folks set the OT up so a bank has patterns 1-4 set to part one, 5-9 set to part two, etc.
You could set the OT up so a bank has pattern 1=part 1, pattern 2=part 2, pattern 3=part 3, pattern 4=part 4,and then repeat with pattern 5=part 1, pattern 6=part 2, etc…
This would organize your banks into 4 main patterns(1,5,9,13) each with 4 part variations with their own plocks(2-4, 6-8, 10-12, 14-16).

Another thing one can do is have a scene or two pre-selected with the part to be loaded that overrides the plocks affecting the first part. It can only be set to one value per parameter and not move with the sequencer, but you can override the parameters you had plocked to a new stationary value that works with the part.

The last thing is that some users have probably programmed parts in ways that when you change them mid pattern, the plocks affect other machines in ways they like, and this will no longer work for them. Doesn’t seem like a common thing to do but over seven years of OT programming I’m sure some folks have projects like this…

Honestly it seems like a pretty deep thing for them to change and I seriously wouldn’t get your hopes up, but I support your request and good luck!

[barnacles]

Plocks are pattern data. It tells the sequencer to set parameter slot “X” with value “Y”. That’s how the octatrack is designed and implemented. You want plocks to be part data, but they are not, because machine and parameters are not part of the sequencer. They don’t contain the pattern data.

The plocks are not mirrored, it’s not a bug. When you change part without changing pattern, the pattern does not change, so obviously your plocks will not change either. They remain in place with the pattern. But you changed the part so the machines and part data will be different.

I do understand your thought, it just doesn’t match up with the way OT works. Hopefully I was able to explain properly.

[tengig]

I do understand your thought, it just doesn’t match up with the way OT works.

I guess @Peter_Peak is suggesting that perhaps it is worth Elektron re-examining this and I wholeheartedly agree, in that I adore this machine that I am learning, but I would never want Elektron to not consider making it even smarter.

My understanding of Parts is still absolute noob-level, so I have no view on what would be a good idea. I do want to lend my support though to creative and supportive criticism of the device, since, as @Peter_Peak pointed out, it isn’t a frozen OS.

I am very inspired to fully understand Parts now, and where they fit into the architecture!

This discussion is interesting and it is helping me understand the machine; thank you everyone

[tengig]

Most folks set the OT up so a bank has patterns 1-4 set to part one, 5-9 set to part two, etc.

You could set the OT up so a bank has pattern 1=part 1, pattern 2=part 2, pattern 3=part 3, pattern 4=part 4,and then repeat with pattern 5=part 1, pattern 6=part 2, etc…

This would organize your banks into 4 main patterns(1,5,9,13) each with 4 part variations with their own plocks(2-4, 6-8, 10-12, 14-16).

wide eyes ooh… right!

I’m gonna go continue my Octatrack training right now

[Open_Mike]

Just a reminder you can do it anyway you want, I don’t actually recommend the second part/pattern approach I mentioned, it was just an example of how you could achieve the results we were discussing.
I’d recommend the first approach for starters, and then deviating your own way eventually…

[Peter_Peak]

However I think of this mirroring, I think it’s not …wise. I checked every machine, it’s mirroring everything, except pickup machine, where are no playback trigs. In AMP and LFO, there are the same parameters, which can be logical and useful, but the rest is incompatible and probably unwanted.

I’ll suggest three variations, what can be done, with considerable possibility of switch it on/off/cycle variations in project menu/personalisation:

1. in program, put flag to plock, which type of machine it’s plocking and when it’s changed by part to other machine, ignore that type of plock-don’t mirror it on parameter of other machine. It can still blink to state it’s plocked, but it will not highlight anything from other machine’s track parameters (maybe just category button in which it’s changed). I think this is not in violation of pattern vs parts competence.
2. allow mirroring only for common values (flex+static src, amp and lofs). this can be useful, I can’t tell how much it’s used in real use. use of mirroring different effects are even harder to guess, if that would be used (even that OT has reputation, that everything can be used ).
3. if anything programming is impossible, easiest, the most economical fix could be to move VOL parameter of thru machine to the right slot, creating empty middle slot, so maybe the most important parameter in important method of OT use - STRT in flex/static - would be mirroring to nothing when parts will change. At least this could be done, I hope.

Thank you everyone for input, suggestions and workflows!

[Open_Mike]

You have some good ideas but there are complications that would arise from at least the first one, but I don’t necessarily want to go into details…

One thing I want to mention is that nothing is being “mirrored”. The plocks simply hold a value that targets a knob on an edit page, they don’t know what machine or effect it is, they are just placeholders.
From the patterns view it’s always telling the part, “plock the lower left knob of the playback page”, for example…
When you change parts or machines the pattern doesn’t know…

Just saying this for clarity because although it can be as plocks being mirrored, its actually always a direct connection from the pattern to the current part…

[Open_Mike]

Imagine a bike where the gears attached to the pedals are the patterns, and the gears on the back wheel are the parts.
When you shift the back gears(parts) it’s still the exact same front gear(pattern) driving it.
The front gear doesn’t know you switched the back gear, it still has a direct connection and drives it the same as any other back gear…

[Peter_Peak]

Thanks Mike for creative analogy. For me, mirroring is wheel with broken teeth-it works, but sometimes makes cycling unpleasable. I couln’t find it in the manual, and if Elektron ever count with it, they should use this behaviour with more creative way as most other functions, finding and placing more rational/interesting relations between machines, effects.

On the other hand, I found another two workarounds-but they are still workarounds, only confirming that Octa is at the same time sophisticated and underdeveloped/underfixed.

Workaround 1 - changing parts in one pattern, but with two tracks - one track is for THRU live+recording in part1, in part2 it’s FLEXwith associated recording track of the other track. This other track is in part1 also THRU, without rec trig, with all prepared plocks and in part2 they are both FLEX with associated recording tracks of each other tracks. With this, mirroring is not at work, at the price of one additional track.
But, while I was playing with this, I came to conclusion, that manual changing parts is really bad. Not entirely because of thru<>flex tracks-I wouldn’t mind some gap in this track, when not so precise timing change at the end of pattern. But because the other drum FLEX track I had, with plocked sliced, which often runs without them in cycle right after parts change.

Workaround 2 - So I was finding how to do the trick without parts - by simple mute/unmute combo. One track is playing THRU, other muted is playing FLEX plocks of recording track of the other track. Instead of part change, I just switch muting of the tracks. But it still costs two tracks per one channel live+mangle option.

Changing parts with patterns is great, solid, but there’s doubling of whole patterns.
So, if parts was ever meant to do the thing in realtime by themselves, in one pattern (in style of thru<>flex mangling), they’d need to get rid of mirroring and also to have a switch to quantize at the end of pattern.

I’d add other analogy . There’s saying, that when you are learning to drive on bad/simple/old/car with problems, you’ll know much more, when you’ll finally sat in the normal/good one. So, we are learning, but we are still in car which needs repairs.

[oldgearguy]

a few comments on this discussion, probably not directly solving the issue, but possibly a bit helpful (or just random text that doesn’t help anyone).

Most people don’t fully understand Parts and tend to be very cautious with them. As a result, they either mainly use one Part or have a carefully set up linking of Patterns and Parts (as outlined in an earlier post). Parts are powerful once you grasp what they control and how to use them, but it’s not intuitive at first (or second).

Patterns are generally easier to understand and manage and folks are comfortable with switching to a new Pattern on the fly, chaining Patterns, or using the Arranger to sequence them.

On the surface, it appears there are 2 ways to change a parameter value on the OT. One is by P-Locking the parameter to a set number in a particular Pattern and the other is by changing the value and saving it as a different Part. For that particular parameter (all else staying the same), the net result is that either way causes that value to change, but the way it’s done is actually very different.

Saving it as another Part makes a copy of all the data associated with the state of the OT whereas the p-Lock simply says - at this particular location in memory, jam this value there. No matter what the state of the OT (as described in a saved Part), that memory location will be getting that saved P-Locked value.

To suggest to Elektron they need to change the behavior would likely result in a ‘will not implement’ comment because you’re asking to fundamentally change the way that P-Locks are implemented. Right now, they are lightweight, fast, and easily modified. They take up very little memory and processing power and since memory is already an issue for some folks, asking Elektron to quadruple the storage needed to manage P-Locks in a different way would certainly impact the available memory for sampling/playback, impact the speed and flexibility of the P-Lock system, and possibly impact the CPU’s ability to manage 8 tracks of samples + 2 effects per track + MIDI.

[Open_Mike]

I hear ya, but the OT is an older car with a new shinny coat…
It could be drastically different and improved in a myriad of ways, I just don’t realistically think much will happen besides minor bugfixes. If major reworking were to be done I think we would have seen an full on OT dps-2 instead of a mk2 with minor hardware changes running basically the same os…

It’s been an interesting conversation, but I must admit the OT can do so much stuff, if I were you I would forget about this and just explore and have fun with all the other stuff the OT can do.
I honestly think your digging into a very specific area too deeply too quickly and its more of a distraction towards making music with the OT and its keeping you focused on one little corner of a vast machine.

My honest advice is to accept the part structure for now and focus on other areas but get back to it later, use pattern change for part change. If you really can’t accept the part structure honestly I’d sell the OT because I feel it’s so unlikely to change as it’s been this way for seven years, but hey keep at it if you want…

In a nutshell if you fight the OT and want it to be different, you’ll probably not like it, not get much done with it, and eventually sell it…
If you accept the way it is and go with it and it’s quirky ways, it’s quite capable of doing extraordinary things…

You can get the OT to do stupid crazy never heard before next level audio experiments, before even considering changing part. It’s a lot more fun!

I will leave you on your own now if you still want to hope for major improvements, but I’m seriously looking out for you by advising to move on.

[Open_Mike]

Funny side notes:
-I’ve referred to the OT as a car multiple times and always refer to it as manual transmission vs automatic.
-Here’s were I refer to it as a spy car:

Should I choose OT or MPC Live?

The OT however is like some James Bond spy car. Harder to handle, manual transmission, but turns tight corners and passes others when the drivers good. You don’t even know it but every once in awhile you get activation codes from headquarters that open secret spy gadgets you didn’t even know you had. There’s water over there, well you’re spy car is amphibious and can cross it, but the controls are difficult so that only you, the spy, knows how and a regular person can’t control it. Wait a minute, how is my car flying, better consult headquarters!

Should I choose OT or MPC Live?

Oh, for sure. A lot of folks would hate to have to deal with driving a spy car and all its quirks, just to get to town…

Food for thought:
-I could have easily threw my OT out the window in my first three months with it but now I will never sell it.

[avantronica]

just placeholders

and a very handy exploit route too … pum ‘rate’ hack

Scene P-lock rate on a Flex to -64 (or whatever) and then switch machine to Pickup

Now the x-fader can play back your pum loop as Normal through to reverse and everything slow in between as per Flex

But - yes, the locks can get messy when machines or effects are changed