Messages

Buy two 2box modules BadgerOfTruth? 

MIDI is a different beast compared to sound channels (or instrument slots)

It's little work for a CPU to process MIDI signals, you can have 56 MIDI slots running off a single ARM chip on Megadrum like it's nothing.

Loading stereo sound channels is different. It requires more CPU, loads of memory, A-to-D convertors, a low latency platform / bus to push through huge amounts of data.

Think about how large an uncompressed WAV file is compared to a MIDI note line on a txt file, it will make sense. Those wavs need to keep playing while others are triggered on top as well!

Even Roland dropped pad patterns for the TD-50 because the modules are at their limits even with the few inputs they allow. Roland's samples are tiny compared to what 2box allows. It's rather impressive tech in 2box all things considered. 

Hope that helps explain things!

I've come from Megadrum to 2box so this may help:

What good triggering means... really needs to be expanded when considering VST vs 2box.

I find the triggering in 2box overall to be better than Megadrum and more consistent, kicks especially, but worse in respect to the losing advanced features. Both systems can be made to be as sensitive, that isn't an issue.

If you want to make some compromises, then 2box allows you to build a perfect drum module that competently covers all the basics with the VST experience 80% intact.

You will have to sacrifice the advanced features you would otherwise take for granted; better hihat playability, positional sensing, 3 zone snares. Do these modern features make the eDrumin / VST experience for you? Then you should expand in the VST direction and put modules aside for another few years.

I think for 2box to be worth abandoning VST triggering, you really need to go all in. Convert the kits you want from all VSTs (especially the ones that suck to trigger) using SDSE so it's worth investing the time and effort. Otherwise stick to playing SD3 through eDrumin.

You predominantly need to feel like leaving VST triggering for a simpler experience on 2box. That will help make deciding easier. 

[several times]

All cards shown here are formatted working for 2box (32k cluster, 4k align)

Please notice the change in scale for the Unbranded and Samsung EVO+ firstly. Highlighted in red are the issues that cause glitches / lag.

This isn't just the average random access time that A1 and A2 do better at here - I ran this test several times and noticed I could get those spikes far more occasionally with the uncertified cards than I could with both the Kingston, only rarely would those spike up to 0.8ms, whereas the others could spike anywhere from 3 to 10 sec in worst case scenario.

Highlighted in white on the CANVAS Go Plus 64 GB A2 - this grouping of slightly higher access time appeared only with 4k align formatting (just interesting to note, it was more consistent without alignment but that doesn't boot 2box OS)

Both the 32 GB cards also seem to have a common split in access time due to something, which nether of the 64GB cards have.

I read that too halftime, which is exactly why I punted for an A2 so I had a comparison.

I will post some pictures here later if I can pass the moderation check. There is also uncertainly around when a manufacturer sells MLC or TLC in the same product line as well, as they often switch technology between sizes and don't specify this, annoyingly.

I do have a significant enough sequential write speed boost on the A2 card vs the A1 card, but read access latency is more uniform.

[as long as they are A1 or A2 certified]

Reporting here that Kingston Canvas Go! Plus microSDXC 64GB works very well in a 2box Drumit5 MKII.

I had a clue from a replacement card I was given by 2box support.

Basically since 2018, manufacturers have only recently started to realise a number of people aren't using memory cards just to store data, but to use them to install and run apps. A helpful standard was invented called "Application Performance Class Specification" to certify which cards are suitable for this; rated A1 and A2.

Having high input-output access per second isn't the complete picture, however, there is very reasonable link between high read IOPS and low random access latency - which 2box needs to be below 1ms on absolute terms. Quite a few older 64GB SD Cards actually meet the criteria to be certified A1 already from my research, but they were made before the specification came about. A1 requirement is at least 1500 IOPS read. My Kingston Canvas Go! Plus is A2 (at least 4000 IOPS read) The official 2box card is rated A1.

I have 4 microSD cards now; the ones that are rated A1 or above, work without issue. The cards that aren't certified at all, have glitches and lag. My testing tools aren't sophisticated, but they work well enough to reveal the secret is in read latency (at a particular chunk size) and also read consistency - that I'm just going to say it; buying any 64GB card (or even 128GB provided you restrict the partition size) as long as they are A1 or A2 certified, should work in 2box without any issues.

Shopping only for A1 / A2 certified memory cards should be the silver bullet.

[I would like an alternate zone setting please.]

I would like an alternate zone setting please.

It would work by alternating through all zones on a dsnd instrument, when only the head zone is triggered.

Example 1: kick sound: ++_ = would alternate between head and rim in sequence. [A B A B A B]
This would make double kick drums sound more authentic when a left and right kick sample is available. Both cannot be merged into a single zone because 2box round robin doesn't alternate samples in sequence, so alternating zones is the best solution while keeping a single input for kick.

Example 2: loop sound: +++ = would play a music loop / chord sequence with available zones. [A B C A B C]
4 zones / loops would be ideal here for progressive chords

[File saved in 2box editor, Maximum tune set to 0.]

Excellent support once again, thanks!

Pleased if I can continue to provide any new feedback that help makes SDSE even better. 

So I've been running some experiments on minimizing background noise with the 2box module, and I discovered something interesting about SDSE / DSound Manager along the way.

Basically I created a 24bit 44100hz WAV in audacity, 15 seconds of silence.

Then I saved multiple versions of this file to see if I could isolate if the noise was generated by the preamp in the 2box module, or if it was contained in the sample and I found something interesting...

File saved in 2box editor, Maximum tune set to 0.
https://www.mediafire.com/file/3czbtbk2voz4u63/15s_2boxEditor_Tune0.dsnd/file

File saved using DSound Manager / SDSE, Maximum tune set to 0, maximize off.
https://www.mediafire.com/file/48c7rdtgjrit960/15s_DSndMgr_Tune0.dsnd/file

The file saved in DSound Manager contains noise. The file saved in 2box editor does not.

Importing a 16bit 44100hz had a little more noise in both programs, but the difference was increased when saving 24bit.
Both programs convert the 24bit WAV in 16bit for 2box, but something seems to happen during this conversion in DSound Manager that adds noise. As far as I can see the only difference is the reported peak value being -9223372036854775808 from the editor and peak of 1 from DSound manager. Strange. 

I'm thinking it's because DSound Manager doesn't support negative peak values and the correction boosts the signal but I have noticed some samples exported using SDSE do cut off abruptly which lead me to this experiment.

I'm keen on isolating noise contained in the recording from the VST from noise contained in the noise floor of the wav.

Even on the newly released Roland CY-16R-T you have to use HIGH cup sensitivity to get good bell response but it's performance is very acceptable - as long as you aren't playing like a feather on it. I'm going to write a short review I think over at vdrums.

Bug to report: Convert > Sort in folders > instr only (this doesn't save, and defaults back to instr + lib for export)

Can I also request a feature? Zones, Snare articulations: Support for Ruffs?

Basically I've been playing around with custom sample stacking and placing a few ruffs samples on the bottom couple of layers of the snare to add a really authentic "autopilot" buzzroll (akin to pressroll notes on other drum modules or dynamic attack smoothing on Roland COSM) so I'd like to export those easier.

One day maybe we could have the same Snare Vel switch feature applied to the head zone with a percent control for ruffs or roll articulations?

For now I'm doing it manually because it requires an careful ear. Ruffs often exceed in volume over the minimum velocity layers on normal head articulation, so you have to bin all but the lower 2-3 ruffs and about 15-20 lower head hits. An algorithm could do this of course by matching the nearest peak volume, but currently the switch happens at about 20% with selective binning. So although it sounds natural, the crossover is too high. When I get the switch happening near 5% it will play much more naturally I presume, so I'm working on exporting ruffs at a reduced volume next, to see how close I can get to 5% without a stark jump in timbre.

So requesting ruff export support, as more snares seem to have ruffs than rolls.

From owning the 2box for a while now I've come to a similar conclusion. Having had experience in both worlds I think there one main issue I'd like to see addressed with 2box; cymbal muting.

Why would this improve hi-hats?

2box has a simplified hi-hat engine (Roland may do this also) that plays a pedal splash sample as both a chick and splash.

This compromise means there cannot be any natural ambience on chicks (or mutes) which makes it sound bad next to a VST which can play the full chick sample (that chick sample might last up to 4 seconds long) But this system also means very reliable splash events as a positive, when there is no transmuting support - which itself isn't a perfect solution either.

There is nothing you can do with eDRUMin that will resolve this as it's not a problem with triggering. 2box either needs a function to adjust the fade out time on all mutes. So I can choose to use chick samples with long ambience and sacrifice splash functionality. Crash muting might also be affected though if it's universal (programming dependant).

An alternative least resource effecting solution: Make a mute event jump the currently playing sample forward a few seconds (or scroll 75% through) instead of killing the sample completely - this should sound more natural all round. The same should occur to currently playing semi-open hihat samples when closing the pedal.

You could also fake this with a digital reverb to mask the sudden cut off but I doubt 2box will program reverb as it would have been implemented already if it was possible.

So I discovered what the audio anomaly is; my speakers peak protection kicking in! 

Funny story, I somehow managed to play just below the threshold that it would only cut out when I thought the module was overloaded - of course, it was just loud enough to happen playing 5 stroke tom rolls.

Secondly, due to an incompatibly with my subwoofer (which I am resolving) I hooked up these balanced speakers directly to 2box. Since they were crossed over at 80hz previously - the problem never occurred before!

So I'm putting that new 64GB SLC card on hold since my Samsung EVO 64GB seems to work flawlessly, at least till it comes down in price.

Well I tried to get my MicroSD 64GB Samsung EVO Plus working in 2box. Even with my IT expertise I wasn't able to crack the formatting problem after a couple of hours, stopping short of making a Linux bootdisk. So I cloned the original using MiniTool as recommended. Easy, extended to 64GB, works, but inefficient.

Surprisingly while I was testing, one of the default snares on the original MicroSD cracked up really badly. Alarming that my MKII had an audio anomaly that early - I hope it isn't a problem.

The original 2box MicroSD beats the Samsung by about 0.4ms less on average read access - so I am still looking to procure a better 64GB card.

Best bet I found is an WD IX QD334 Industrial microSD 64GB SLC however it has proactive error checking so it could cost some latency. Difficult to discern details at this technical level without sending a few emails.

Worth a punt at £92?  It's not that bad considering the higher seq write speed SLCs go for £500+

Most other manufacturers I've researched cap their SLC line ups at 32GB unfortunately.

First I've heard of Lemon cymbals, they look like the new Triggera and very good value! I'm pleased they exist but zero suppliers in Europe is annoying. I don't want a downgraded CY-15R to be honest, I'll pay for the sequel and hope to make enough back on my used gear.

I'm not expecting 2box to fix anything here, I just wanted to know if a good triggering 3 zone performance is possible on 2box? I.e. bell triggering at the same strength as bow with stick tip only?

The rest of my triggers are pretty good except the rim performance, which is acceptable, but not DTX-Pro great. But then I've ruled that module out for 2box's sample prowess which I hope to exploit as soon as I can clear this hurdle. 

Yes, I have it on H.

L and M effectively kill all bell response.

Before I read the manual, I intuitively thought L, M, H referred to the different gain levels for edge, bow, bell. But it's looking increasely likely I may have to bite the bullet and get a 16R-T since ATV's 18 inch ride is absolutely atrocious to buy on it's own. Just really hard to justify when I can't see anything wrong with the CY-15R I have and no way of knowing if the same wouldn't occur on something like a Mimic Pro.

Hi folks, no prizes for guessing why I'm writing here. 

Is there any hidden function to further boost bell response?

I was lead to believe my CY-15R would be great on the MKII - but that just isn't the case.

I'm using zourman's recommended settings; CyPSB 13H -36. These do indeed seem to be the best settings after experimenting. 2box's calibration is incredibly simple compared to what I'm used to lol.

At the moment, I'm looking at best roughly 20dB signal reduction from the same strike strength from bow to bell. Oddly, bow/edge goes into 2A, bow/bell into 2B to have the zones correct on 2box. I have disassembled my CY-15R partially, to check the 3 ribbon connectors are in great condition.

I need the ability to make the bell response parity with bow response. Am I dealing with a broken CY-15R I that was just able to calibrate better on Megadrum?

I feel I have exhausted all the settings available (which is why I'm posting here) I need to fix this issue soon because I'm not sure what's the expected normal on 2box. It's not acceptable performance to whack the bell. Alternatively I'll be peeved if I have to buy a better cymbal since I know my CY-15R is capable of great performance on Megadrum.

Does anybody have any suggestions?