1
0
Fork 0
Commit graph

40 commits

Author SHA1 Message Date
Robbert van der Helm 68cf0455ee Rework and optimize block smoothing API
You now need to bring your own buffer instead of the smoother having a
built in vector you would need to pre-allocate. This makes the API
simpler, and also much more flexible when doing polyphonic modulation.

In addition, the new API is much more efficient when there is no
smoothing going on anymore.
2022-07-06 14:29:29 +02:00
Robbert van der Helm 1424b98e38 Break PolyModulation into that and MonoAutomation
This is needed to properly support polyphonic modulation, since the
modulated value may still be automated in the meantime and the
polyphonic modulation must act as an offset for that value. This does
mean that the plugin must add the normalized value and normaliezd offset
by itself. The `PolyModulation` event now also contains a description of
how this can be used. It would have been nicer to be able to send
polyphonic automation-style events instead (like the PolyModulation was
doing before), but that's sadly not feasible without NIH-plug being
involved in the voice management.
2022-07-06 13:40:12 +02:00
Robbert van der Helm 45c141c394 Make smoothing styles Debug, Clone and Copy 2022-05-29 14:21:41 +02:00
Robbert van der Helm e8764e7327 Use double precision for exponential step sizes
Otherwise you will get issues with rounding errors if someone ever tries
to have smoothing times over a second or so.
2022-05-12 13:50:45 +02:00
Robbert van der Helm 6a40cc9916 Remove debug print 2022-05-12 13:44:22 +02:00
Robbert van der Helm f3d02f6210 Fix updated exponential smoothing 2022-05-12 13:44:01 +02:00
Robbert van der Helm 1fea9f94df Deduplicate the smoother implementation
By introducing a new trait for conversion to and from floating point
numbers.
2022-05-12 13:34:02 +02:00
Robbert van der Helm dc5dfb06c6 Reach 99.99% of target for exponential smoothing
Instead of 99.97%. Both numbers are not based on anything, but this
calculation makes more sense.
2022-05-12 13:21:32 +02:00
Robbert van der Helm 28c7ea4843 Rename ExponentialIIR -> Exponential
The suffix was added to denote that this could do some funny things
because of the snapping to the target value, but in practice that
snapping distance will likely be so tiny that it is likely not
noticeable.
2022-05-12 02:22:12 +02:00
Robbert van der Helm 222036b335 Get rid of unnecessary loop in ExponentialIIR 2022-05-12 01:56:16 +02:00
Robbert van der Helm 116d245149 Add an exponential smoothing style
Based on a one-pole IIR low pass filter that reaches 99.97% of the
target value in the designated time before snapping to that value. This
will be less efficient than the linear FIR smoothing style.
2022-05-12 01:35:01 +02:00
Robbert van der Helm b89b4dfbb2 Add a .previous_value() method to the smoothers 2022-04-27 17:43:40 +02:00
Robbert van der Helm 98d725d24f Add an Iterator implementation for smoothers 2022-04-04 17:58:25 +02:00
Robbert van der Helm 6f01097b03 Add mapped next_block() functions for smoothers 2022-04-04 17:44:49 +02:00
Robbert van der Helm 3d7a23c812 Expose steps_left on the smoothers 2022-03-23 16:50:09 +01:00
Robbert van der Helm 80457ac0f9 Fix doc links after prelude migration 2022-03-03 23:30:29 +01:00
Robbert van der Helm f581294d7b Update rustdoc formatting for links
Apparently it showed this text verbatim, and not in monospace.
2022-03-03 23:05:12 +01:00
Robbert van der Helm fdda32696a Remove unused import 2022-03-01 23:50:41 +01:00
Robbert van der Helm 8eafcebe62 Mention the panicking behavior of block smoothing 2022-03-01 20:23:04 +01:00
Robbert van der Helm 3e8b2b3759 Use AtomicRef instead of Mutexes for block smooth 2022-03-01 19:52:56 +01:00
Robbert van der Helm 3122c0cb41 Pass block references to the block smoother
So you can't mess this up by passing the maximum block size instead.
2022-03-01 17:33:22 +01:00
Robbert van der Helm dffddfaf04 Implement the block smoothers 2022-03-01 17:29:09 +01:00
Robbert van der Helm d470a1db5a Add functions for allocating block smoothers 2022-03-01 17:07:03 +01:00
Robbert van der Helm 892aef7e37 Mark smoother functions as inline 2022-03-01 16:55:30 +01:00
Robbert van der Helm fd3f4c2c48 Allow skipping steps in the smoothers
I'll need this for Diopser since it's going to have a parameter to
control the automation's granularity.
2022-02-13 18:33:05 +01:00
Robbert van der Helm c99877ec05 Fix Smoother::is_smoothing() 2022-02-12 20:54:03 +01:00
Robbert van der Helm f645b6a92f Relicense under the ISC license 2022-02-12 16:04:46 +01:00
Robbert van der Helm 379c74e874 Remove spurious dbg!() calls 2022-02-10 23:41:04 +01:00
Robbert van der Helm 2486368d08 Use AtomicF32 for the smoothing
This should at least not have a big performance impact since we can use
relaxed memory ordering here. AtomicCell always uses acquire/release
ordering.
2022-02-06 03:43:59 +01:00
Robbert van der Helm bf070fce5a Use atomics in the smoother
This is needed so we can share the params with the editor, but it isn't
great, is there a better alternative?
2022-02-06 01:33:19 +01:00
Robbert van der Helm b901dac012 Disambiguate rustdoc link references 2022-02-05 12:56:03 +01:00
Robbert van der Helm 63d30d33e4 Fix rustdoc links 2022-02-05 12:52:34 +01:00
Robbert van der Helm 073fe743c2 Split up set_target and reset in smoothers
This makes it a bit less awkward to use in user code, if you need a
separate smoother there.
2022-02-04 15:17:42 +01:00
Robbert van der Helm 30609c3ecd Add an is_smoothing() check 2022-02-03 16:51:41 +01:00
Robbert van der Helm 2ca54d220d Add logarithmic smoothing 2022-02-02 23:00:17 +01:00
Robbert van der Helm 1ad477ee4f Add smoothing tests 2022-02-02 22:34:29 +01:00
Robbert van der Helm da291175d1 Rename SmoothLinear to Linear
We'll rethink this when we get to sample accurate automation.
2022-02-02 22:00:00 +01:00
Robbert van der Helm 0ecec104fb Completely reset smoothing on init or restore 2022-02-02 21:41:20 +01:00
Robbert van der Helm ea7dc2ffa3 Implement smoothing for integers 2022-02-02 21:26:20 +01:00
Robbert van der Helm 8f89754ba5 Add simple linear parameter smoothing 2022-02-02 21:26:13 +01:00