Signal flow/gain staging diagram for Octatrack MKII

Is there a signal flow or block diagram for the Octatrack MKII? I found lots of text but no diagrams, so here’s my attempt trying to visualise the level & volume controls if it’s any use. The lines show possible connections; they’re not all active at the same time. Am new to the machine so just feeling my way here :slight_smile:

44 Likes

Nice effort. Thanx

1 Like

Thanks for the very clear and helpful sheet! All those hundreds of discussions about ‚gain staging in the OT‘…

Your one-pager should get a sticky, be considered in Merlin‘s guide etc.

That’s an awesome first post in this forum! :+1:

2 Likes

Maybe add ˋOT Gain Staging´ to the thread title…

Thanks for the support guys; have amended title. Will add pick-up machines once I’ve got my head round them. Let me know if there are any errors. I’m not sure about the -12dB stages; they seem strange, especially for the Thru Machine input. If signals routed into tracks on-the-fly are attenuated deliberately to avoid overflows when they’re mixed, it doesn’t explain why audio edit gain defaults to +12dB for record buffers; I would have expected the Thru machine volume to default to +12 as well?

3 Likes

Good job. Seems to correspond to my results, but I wonder if Amp Vol unity is +63, 0 > -12db, -32 > -24 db…Same for Thru VOL.

Level 108 seems to be - 3db, 64 > -12 db

That is to say?

1 Like

@sezare56 thanks for this; the additional dB points would be useful so I can add those.

The unity points are really the main point of the diagram so I’m keen to get them right. If you’re sure about Amp Vol, I’ll do some more tests, but maybe we’re interpreting ‘unity’ differently. I use it as ‘the signal coming out of the enclosing box is the same as that going in’.

So in the Thru Machine and Flex Machine Rec Buff paths, the preceding -12dB boxes lower the signal by a fixed amount, and if the following controls are at unity, the signal remains 12dB lower. My ‘not being sure’ was more about why it was designed this way as it wasn’t consistent… but then I’m still learning the Elektron ‘way’ :slight_smile:

I could simplify the thru machine path by removing the -12dB stage and changing the thru vol arbitray mapping so that +63=unity, but the problem is then what to do with the other path. The audio edit gain is calibrated directly in dB on the machine so there’s no mapping to change.

Comments/suggestions welcome.

1 Like

I’m not sure! :smile:
But it make sense to me.

Where does your [- 12db] box come from?
Why DIR don’t have - 12 db?
Why wouldn’t it be Amp VOL =0 > - 12db, VOL =63 > 0db
Attributes Gain compensate recordings with +12db, maybe because Amp VOL reduce audio…

I’m also not sure about headphones mixer values (50%/50%).

Pickup GAIN (knob E) seems to be for Overdub only. Weird.

1 Like

Yeah intuitively I’d say that if you’ve recorded to a flex machine and want the same volume as you had going in, then the AED gain has to be at +12db, as is default. It’s still a bit of a mystery to me how, where or why this volume drop happens.

Amp VOL? :thinking:

Have just uploaded v3 and amended the link in post 1.

They came from me :slight_smile: I just added them as they reflect how the machine appears to behave by measuring signals in & out. The reduction in level can be represented as either a -12dB stage followed by a volume control where 0 = unity, or leave out the -12dB stage and change the volume control so that +12dB = unity… if that makes sense. It’s just a case of which is easier to understand.

I’ve tried using the latter approach with the new V3 diagram so see what you think. It does look simpler for the src vol control in the Thru machine, but it wasn’t so good for the sample gain control. I had to duplicate it because its unity setting now depends on the source of the sample - if it’s from the static or flex sample slots, 0 = unity, but if it’s from the rec buffs, +12 is unity… but maybe this is clearer after all?

The cue signals from the inputs also have the same issue. There’s a -4dB reduction when you route the inputs to the cue, so I’ve represented this with a -4dB block. It might get messy trying to do the same thing as before with the mixer cue level control. For now, you just have to accept that mixer cue unity position only applies to track cues. If you want to compensate for the -4dB drop from the inputs, you need to raise the mixer cue level accordingly.

I didn’t measure any level reduction here. If you mean why isn’t -12 shown on the control, I’ve only shown the min, max and unity position for all controls to avoid clutter. Have added a generic table on the right showing more intermediate positions.

Again this depends on your scope of ‘unity’. In my definition, it means the box which encloses the control, so if a -12dB signal enters the box, a unity setting means that the same -12dB signal leaves the box. That way, you can track the signal level through all the controls and predict its value at each stage.

I’ll check this.

Not done any measurements yet for pickup machines.

Hopefully the new v3 diagram makes this clearer.

2 Likes

Hi. I was just checking while you were writing. :wink:
As there’s unity with recordings having +12db in Attributes, and VOL = 0, my conclusion was unity is with VOL = +63 if you set Attributes GAIN to 0. Make sense ?

You could have many combination of individual control settings which result in unity gain over the whole chain… but that’s the whole issue (problem!) with gain staging. The key thing imho is to know what unity is at each stage so you can optimise the headroom and noise (for analogue gear certainly).
btw - have just confirmed the headphone mix control and it does seem to work as shown - it adjusts the relative amounts of main and cue signals going to the headphones.

Writing this above, I just thought about Master Track.
My theory doesn’t seem to work with T8 set as Master Track, it would be - 12db lower with VOL = 0. Doesn’t seem the case.

Shouldn’t it be 100%/100% in the middle?

ahh I see what you’re getting at :slight_smile: I’ll go and check, and take a look at the master track.
It’s some beast this machine :slight_smile:

1 Like

Re. headphones, if you listen to a main signal only and move the mix control, it definitely starts loud fully to the left, and reduces in volume as you move to the right, eventually disappearing. So if we say it’s 100/100 in the middle, the ends must be greater than this, in which case what value can we put?

No Amp VOL for Master Track. My theory is still valid. As you said, a 0 db sample don’t clip if you increase VOL to +63. If you add +0.5 in Attributes, it clips.

-3db /-3db or less maybe?

Yes your theory works, but only for signals from the rec buffs; you still need a different unity setting for signals from the sample slots. There’s an old rule of thumb which says gain should be applied as early as possible in a chain. It may not be as relevant with digital or software-based gear, but if we accept that flex & static signals always flow through both the aed gain control and the amp vol control, then the correct place to compensate for the -12dB drop in rec buff signals is at the aed gain stage. Same argument for the thru machine - it’s better to compensate at the src vol control stage.
If you compensate using the amp vol control, you’ll also affect neighbor machine levels as these also pass through it. Just my 2c.

:thinking: For me both (SRC and AMP) affect neighbor…