Latest MegaDrum firmware version

[dmitri]

All-in-one archive of MegaDrum firmware:

megadrum_20100827.zip

Don't try to load this firmware into MegaDrum with a non-protected bootloader, it won't work!!
See viewtopic.php?f=3&t=1251 for details.


All firmware files in the archive have the following naming scheme:
megadrumppp_hh_yyyymmdd.bin for 22 inputs versions
megadrumppp_hh_32_yyyymmdd.bin for 32 inputs versions
megadrumppp_hh_48_yyyymmdd.bin for 48 inputs versions
megadrumppp_hh_56_yyyymmdd.bin for 56 inputs versions
where
ppp - AVR type. 32 - for Atmega32, 324 - for Atmega324(p), 644 - for Atmega644
hh - crystal clock in MHz. Atmega32 versions with a clock above 16MHz are overclocked, Atmega324(p)/644 versions with a clock above 20MHz are overclocked.
yyyymmdd - date(version) of the firmware.

Use mdfwupdate utility to load the firmware into MegaDrum.

See changes before 2010 in History of MegaDrum firmware versions

Changes in version 20100827:

1. Fixed a serious bug (which was there since 20100126) which, for no apparent reason, could cause some random inputs to get stuck and not responding to hits and requiring power cycling.

Changes in version 20100801:

1. Added support for adjustable half open pedal levels.

Changes in version 20100702:

1. Ooops, in previous version positional sensing was not in fact working on CrushB and Tom1H. In this version positional sensing works on all 4 inputs (SnareH, RideB, CrushB and Tom1H).

Changes in version 20100701:

1. Made positional sensing available on 4 inputs: SnareH, RideH, CrushH and Tom1H.
2. Had to remove firmware versions for 40 inputs on Atmega32/324p. Not enough RAM on these MCUs for positional sensing on 4 inputs.

Changes in version 20100623:

1. Better handling of rotary encoders.

Changes in version 20100618:

1. Enable/disable HiHat pedal "soft chicks" now really works.

Changes in version 20100613:

1. Added support for two additional HiHat pedal position levels with dedicated Notes. The levels between open and closed now go like this:

Open - notes are set on the HiHat cymbal input.
SemiOpen - notes are set in BowSO/EdgeSO/BellSO.
HalfOpen - notes are set in BowHO/EdgeHO/BellHO.
SemiClosed - notes are set in BowSCL/EdgeSCL/BellSCL.
Closed - notes are set in BowCL/EdgeCL/BellCL.

See viewtopic.php?f=3&t=1192 for discussion of this feature.

Changes in version 20100603:

1. Added an option to enable/disable HiHat pedal "soft chicks". See viewtopic.php?f=3&t=972&p=14589 and viewtopic.php?f=3&t=716 for discussions of pros/cons of "soft chicks".
2. All firmware files are now encrypted.

Changes in version 20100516:

1. Added support for Positional Sensing on a SnareH input. It is still experimental, see viewtopic.php?f=3&t=1215 for details.
2. Some signal sampling improvements.

Changes in version 20100402:

1. More tuning to AltFalseTrSupp algorithm to work better with mesh head pads which have long after hit ringing.

Changes in version 20100324:

1. Improvement to AltFalseTrSupp algorithm for mesh head pads. It is not needed now to set MinScan high for such pads.
2. Improvement to head/rim separation for some types of dual piezo/piezo pads, e.g. Roland PD-125X.

Changes in version 20100313:

1. Added options for fine tuning Head/Rim/Crosstick separation: DualMidPoint and DualMidWidth. Both can be set between 0 and 15. The higher DualMidPoint the easier rim hits and harder head hits are and vice versa. The higher DualMidWidth the easier rim shots are and harder head/rim hits.
2. Fixed a bug where with QuickAccess enabled some combination of jumping to a pad and then navigating in the menu could lead to resets or hangs.

Changes in version 20100307:

1. Major re-work of the alternative false triggering suppression (AltFalseTrSupp=Yes) algorithm.

Changes in version 20100301:

1. Fixed a bug with 'pre-burst' which might have caused intermittent false triggering with maximum velocity. The bug was more evident with stm32.
2. Some adjustment to dual zone piezo/piezo pads handling to better separate zones for some types/brands of pads.
3. More modifications to AltFalseTrSupp algorithm.

Changes in version 20100212:

1. Added support for '2 pulses per detent' rotary encoders (AltEncoders setting).
2. Modified algorithm for the alternative false triggering suppression (AltFalseTrSupp)

Changes in version 20100126:

1. Small adjustment to Dynamic Threshold algorithm
2. Added signal 'pre-burst'. If signal on an input is higher then Threshold/2, MegaDrum starts a short burst. If the level doesn't reach Threshold, the burst is cancelled without producing a result. You can see an effect of the pre-burst if Latency set to an odd value - when a pad is hit lightly but not enough to exceed Threshold level the VU meter for the pad set to a maximum but no MIDI signal is sent and the LED stays off.

[Firelord]

2. Added signal 'pre-burst'. If signal on an input is higher then Threshold/2, MegaDrum starts a short burst. If the level doesn't reach Threshold, the burst is cancelled without producing a result. You can see an effect of the pre-burst if Latency set to an odd value - when a pad is hit lightly but not enough to exceed Threshold level the VU meter for the pad set to a maximum but no MIDI signal is sent and the LED stays off.

Can you please explain this feature in more detail? What's the main idea behind it?

[dmitri]

The main idea is to improve triggering for very light hits. MegaDrum does sampling burtsts when it detects a signal above Threshold. It may not be good enough for light hits so this new features starts bursts even before Threshold reached. Be aware, that setting Threshold too low can cause constant pre-bursts, without you realising this, thus making triggering worse since MegaDrum will spend too much time on inputs which in fact don't go above Threshold. This is why I also added a possibility to see 'pre-bursts' if you set Latency to an odd number.

[anttipi]

2. Added signal 'pre-burst'.

I'm just about to update the FW and if this feature does what I think it does, I'm very grateful for the inclusion! I've been struggling to get the threshold right for the hi-hat head - when I try to raise it, I start getting annoying 127 CC value hits occasionally even though hitting very lightly. So the pre-burst will eliminate this.

Please, correct me if I'm mistaken!

[Sabre]

I'm just about to update the FW and if this feature does what I think it does, I'm very grateful for the inclusion! I've been struggling to get the threshold right for the hi-hat head - when I try to raise it, I start getting annoying 127 CC value hits occasionally even though hitting very lightly. So the pre-burst will eliminate this.

Please, correct me if I'm mistaken!

Please advise if it does what you think it does.
I've updated the firmware, but haven't tweaked any settings.
I'd be interested to know your findings on this.

[anttipi]

Please advise if it does what you think it does.
I've updated the firmware, but haven't tweaked any settings.
I'd be interested to know your findings on this.

I will report back on this as soon as I'm done re-tweaking the settings for the new firmware (which will hopefully be this weekend)! I'm not sure yet.

In the meantime, I want to comment on the new algorithm for false triggering suppression: I don't know what exactly Dmitri has modified but it seems like a dramatic improvement with my DrumTec Diabolos! Previously, the minimum values I could use for the pads without getting any false triggering was something like dynamic level 3, retrigger 4, dynamic time 32. Now it seems I can easily use dynamic level 0, retrigger 1, dynamic time 8 with no false triggering!

[anttipi]

Please advise if it does what you think it does.
I've updated the firmware, but haven't tweaked any settings.
I'd be interested to know your findings on this.

I just fiddled with the hi-hat settings and it seems like I understood this right. No sudden 127 jumps anymore when hitting lightly and raising the head threshold.

And as I wrote, the new false triggering suppression certainly seems to have reduced false triggering but I found out it's unusable to me as I can no longer do drum rolls with Alt False Triggring enabled. So now I leave it disabled with retrigger at 4, dynamic level 0 and dynamic time 32 which works fine.

[Firelord]

Now for some bug reporting. I am using the latest version of firmware for MegaDrum 324P, 20_32_20100402.

1. When in menu and a signal is received from a pad, the signal VU meter will sometimes appear inside of the menu;
2. I don't know whether this is a firmware bug or not, but I sometimes get signal activity on random channels when I hit a single pad. It might have something to do with the 'signal pre-burst' feature. Anyway, my pad configuration hasn't changed in a while but I noticed this only recently.

I would also like to know what is the current situation with MegaDrum and MTC. The thing is, I've got problems with recording long drum sessions with MegaDrum. The strange part is that when I play, all notes are in time, however when I'm playing and recording in Nuendo, the resulting recording offset is huge (on the order of 250-1000 ms) and what's worse even this offset is not consistent and it changes over time.

Now it may be that I've got problems with MegaDrum/Nuendo settings. But I also own a simple MIDI-keyboard. It appears not to have this issue with recording.

[dmitri]

Now for some bug reporting. I am using the latest version of firmware for MegaDrum 324P, 20_32_20100402.

1. When in menu and a signal is received from a pad, the signal VU meter will sometimes appear inside of the menu;
2. I don't know whether this is a firmware bug or not, but I sometimes get signal activity on random channels when I hit a single pad. It might have something to do with the 'signal pre-burst' feature. Anyway, my pad configuration hasn't changed in a while but I noticed this only recently.

These two are both a result of a feature introduced some time ago:

Changes in version 20100126:
...
2. Added signal 'pre-burst'. If signal on an input is higher then Threshold/2, MegaDrum starts a short burst. If the level doesn't reach Threshold, the burst is cancelled without producing a result. You can see an effect of the pre-burst if Latency set to an odd value - when a pad is hit lightly but not enough to exceed Threshold level the VU meter for the pad set to a maximum but no MIDI signal is sent and the LED stays off.

[Firelord]

Now it may be that I've got problems with MegaDrum/Nuendo settings. But I also own a simple MIDI-keyboard. It appears not to have this issue with recording.

Never mind this. I did have a similar problem with the MIDI keyboard.