MIDI to automation lane

[ljekio]

Hi
I noticed a nasty feature - automation from midi source recorded very coarse and smoothy.
As example I take thor sequence with rectangle curve and send it to EMI.
LoopMIDI returned this signal in Reason through standard (other) midicodec.
After it I received signal by second Thor's rotary.

2016-05-31 00-39-18 Document 1 .jpg (105.54 KiB) Viewed 1579 times

First scope showed signal before midi sending.
Second scope after EMI -> loopMIDI -> Thor (picked directly from rotary).
Very light smooth but rectangle saved.
Lane after track record on bottom.

I tried change option "automation cleanup level" at the preferences but unsuccessfully:

2016-05-31 00-54-16 Document 1 .jpg (45.11 KiB) Viewed 1579 times

But if I record handy automation, turning the knob manually, Reason will precise record it.

Is there any solution for more sharp midi record of automation?

[orthodox]

Your automation events are quantized to 1/16 during recording, maybe it's something with that?

[ljekio]

No, snap off not affected too.
And rate of sequencer curve is 1/16.

UPD: 1/8 rate of sequencer wrote correctly and rectangle.

[ljekio]

OMG, this sine LFO with 1/16 of frequency

2016-05-31 11-20-29 Document 1 .jpg (41.2 KiB) Viewed 1548 times

Possibly it old bug?
Or ancient feature for economy resources from Reason 1?

[ljekio]

Bump this thread. Waiting for normal midi-recording.

[selig]

I'm not seeing that at all on my end. Here's a sine waved recorded via MIDI Out Device over IAC back into Reason, showing what happens at different audio buffer settings (buffer settings as clip label above clip:


Note how everything falls apart at higher buffer settings!
Here's what the data looks like for the lowest buffer setting (64):

[ljekio]

Another thread, where is some screen about it.

viewtopic.php?t=7504648

[selig]

Another thread, where is some screen about it.

viewtopic.php?t=7504648

Totally not seeing this at all, looks normal with the exception that on bar one there's a slight glitch (but I never use bar one so I'd never see it myself).

Here's the matrix test, using MIDI Output Device (MOD) to record via IAC bus. As you can see there is exactly the number of data points you would expect, and while they are about 10 ticks behind they are all within 4 ticks of each other (about 5ms of delay/slop, which if you could quantize automation data points would be a moot point.):

[ljekio]

Keyword here is frequency of LFO.
Just try it with 1/16 (with default 120 bpm)

[selig]

Hold the presses, at higher rates/tempos I DO see what you were seeing, so it's still not fixed yet. Workaround is super simple, record automation at a slower tempo if it's important to track accurately.

[selig]

Keyword here is frequency of LFO.
Just try it with 1/16 (with default 120 bpm)

Got it - 1/16ths DO work at around 70 BPM and lower, so at least the workaround is simple.

[ljekio]

Yes, but if I play with breathcontroller, many of nuances will spoiled because it.

[selig]

Yes, but if I play with breathcontroller, many of nuances will spoiled because it.

If you can create 16th note full on full off with a breath controller, then maybe - however, I just did a mod wheel test where I'm moving it up/down in 8ths and it's recording just fine. I'm not sure how breath could move much faster, if at all, than slamming the mod wheel up/down as fast as I can.
I feel there may be a difference between using a loop back/IAC bus and simply recording controller data.
Do you have examples of recorded controller data that's nothing like the source, and how are you doing the tests if so? Not sure what the best test would be, other than to record the audio of a MIDI performance and compare it to the audio generated by playback of the same MIDI data (make sense?). Would be an interesting experiment anyway!

[ljekio]

Do you have examples of recorded controller data that's nothing like the source, and how are you doing the tests if so? Not sure what the best test would be, other than to record the audio of a MIDI performance and compare it to the audio generated by playback of the same MIDI data (make sense?). Would be an interesting experiment anyway!

I did it today. The controller recorded in Reaper and Reason. In Reason in two ways - midi and audio through Thor's audio out. Thor received breath controller and converted to audio out.

2020-06-12_15-37-33.jpg (37.4 KiB) Viewed 778 times

Top curve is import from midi from Reaper.
You can see ladder of midi quantization but not smoothness of angles of curve.
Second - automation that record Reason.
Bottom curve - audio track from Thor.
Smoothness of angles gives a slightly different character to the resulting tremolo (so-called flutter tremolo).

Import from Reaper

2020-06-12_15-55-29.jpg (21.73 KiB) Viewed 778 times

Recorded in Reason

2020-06-12_15-55-58.jpg (19.37 KiB) Viewed 778 times

[selig]

I did it today. The controller recorded in Reaper and Reason. In Reason in two ways - midi and audio through Thor's audio out. Thor received breath controller and converted to audio out.

Curious which better matches the original, meaning if you record the performance as audio and MIDI, then use the MIDI to re-record the audio (from both sources), which of the "re-recorded" audio sounds closest to the originally recorded audio? Any differences would be attributed to how MIDI was recorded or played back, right?

[ljekio]

Curious which better matches the original, meaning if you record the performance as audio and MIDI, then use the MIDI to re-record the audio (from both sources), which of the "re-recorded" audio sounds closest to the originally recorded audio? Any differences would be attributed to how MIDI was recorded or played back, right?

We can see that it is useless to record audio in this case, the curve is the same.
Reason filters the frequency of MIDI controller signals, making it lower than the MIDI sample rate.