2Box at NAMM 2012 this week!

[Krillo]

Who me? 
Plenty of layers to get good sounds in the module from SD2.0 .... and when you build your dsnds you have the choice in Dsoundtool to use up to 99 hits per zone ... like 99 hits on a snare head for example. If you record 99 hits in SD2.0 .... there is plenty of alternating, randomizing of hits to avoid an unnatural sounding/responding result ... and the module has a variable setting beyond that .... I don't even use the 2Box module's variable setting, I just choose fixed since I have already recorded a variable sound in SD2. And as far as soft hits .... the other thing happening is Louis' Dsoundtool works really good now ... it spreads the velocity of the hits over a wide range....

No advantage to me whether people go that route or not .... definitely nothing to gain on my end ... but, I can tell you that I didn't hear anything that tops these sounds from a stand alone module at NAMM 2012. Each time I came home ... no matter how tired I was I had to sit at my kit just to smile and gloat a bit!

Not sure if you have checked out the video demos I've done ... and of course sounds are very subjective.... but I am pleased with results from SD2:
http://www.2box-forum.com/index.php/topic,912.msg8818.html#msg8818

I guess this was a response to me?
My point was exactly that - the 2box module supports 99 layers and SD2 has much less. By that I do not mean that your kit doesn''t sound great, it does! My ddrums sound great too with only up to 16 layers. My findings in regard to SD2 is merely that it didn't have as many unique hits / samples as the hype kind of suggested. Actually only 12 on some sounds and 17 on others.
I'm also sure that your dsnds kick all Namm's e-kits butt, including the stock 2box. I have downloaded the sounds from the 2box site, so do I have an idea of how those sound, also I have checked out tons of Youtube vids of 2box kits which have the stock sounds, and they do sound good.

[Jman]

I guess this was a response to me?
My point was exactly that - the 2box module supports 99 layers and SD2 has much less. By that I do not mean that your kit doesn''t sound great, it does! My ddrums sound great too with only up to 16 layers. My findings in regard to SD2 is merely that it didn't have as many unique hits / samples as the hype kind of suggested. Actually only 12 on some sounds and 17 on others.
I'm also sure that your dsnds kick all Namm's e-kits butt, including the stock 2box. I have downloaded the sounds from the 2box site, so do I have an idea of how those sound, also I have checked out tons of Youtube vids of 2box kits which have the stock sounds, and they do sound good.

How did you determine how many unique hits ... not really arguing this, but not sure? There is a voice and layer limit setting, which I set to unlimited on the layers, and just set it to a high number on the voices ... I have never actually counted the voices of individual drums.... etc...
And of course the no. of Velocities is dependent on your choice of MIDI file, no of hits..
but interested in how you determined the number of actual unique hits.....

[the3D]

honestly i don't understand why everyone is so excited about PS. Are you really using the snare side sound so often? The only song in wich i have ever used it was gavin harrison's 19 days. I am much more concerned about MIDI expandibility, i have a trigger IO and a bunch of pads i can't use simply because it's impossible to assign a MIDI note to a sound that is not bound to a drum channel, therefore midi in is almost useless for me. Please improve this thing i don't want to buy another module!

[nonoduweb]

...i don't understand why everyone is so excited about PS...

According to Jerry, PS won't be implemented soon, so you're right, no need to be excited.
AFAIK, the module has the integrated circuits needed to make PS, so IMO it's more reasonable than asking for more effects or USB-midi connection, wich would add a lot of latency.

[Krillo]

How did you determine how many unique hits ... not really arguing this, but not sure?

Create midi sequence containing the note number of interest. For example wholenotes of C1. One note per velocity, 1 thru 127. Mute all mics and channels exept for the one we want. Bounce to file. Normalize every individual hit. Cut every hit out and save as separate files. Start position must be the same down to the sample. Put each of these files on separate tracks. Phase invert channel one. Check all channels against channel one to see which one phase cancel each other out. Those that do are redundant and can be removed. Then check channel 2 against all the rest, then channel 3 etc. You will at the end of the process be left with those that occur only once.

[edtc]

... yes , clever procedure !  cant lie this way...

[Jman]

Create midi sequence containing the note number of interest. For example wholenotes of C1. One note per velocity, 1 thru 127. Mute all mics and channels exept for the one we want. Bounce to file. Normalize every individual hit. Cut every hit out and save as separate files. Start position must be the same down to the sample. Put each of these files on separate tracks. Phase invert channel one. Check all channels against channel one to see which one phase cancel each other out. Those that do are redundant and can be removed. Then check channel 2 against all the rest, then channel 3 etc. You will at the end of the process be left with those that occur only once.

Sounds pretty thorough to me

[digitalDrummer]

In case anyone wondered about the 2box external triggers, here you go...

[puttenvr]

Nice dessign

[Pascaldc]

i wonder when those external triggers will be available , and how much they wiil be..  Interesting that it doesnt seems to hold on a lug bolt

[Slap the drummer]

Create midi sequence containing the note number of interest. For example wholenotes of C1. One note per velocity, 1 thru 127. Mute all mics and channels exept for the one we want. Bounce to file. Normalize every individual hit. Cut every hit out and save as separate files. Start position must be the same down to the sample. Put each of these files on separate tracks. Phase invert channel one. Check all channels against channel one to see which one phase cancel each other out. Those that do are redundant and can be removed. Then check channel 2 against all the rest, then channel 3 etc. You will at the end of the process be left with those that occur only once.

I keep wondering about this........

A) Did the process involve using SD's "humanization" features when generating the files ?

B) What do the "humanization" features actually do?  If all they do is randomize the velocity then they
will indeed be cancelled out by Krillo's technique - but I imagined the features doing more than this
(so in that theory they shouldn't cancel even if based on the same physical recorded strike).

C) What if there are, say, 17 discrete velocity layers, but each layer comprises multiple hits which have
been normalized as per their velocity?

Just thinking aloud....... 

[Krillo]

I keep wondering about this........

A) Did the process involve using SD's "humanization" features when generating the files ?

B) What do the "humanization" features actually do?  If all they do is randomize the velocity then they will indeed be cancelled out by Krillo's technique - but I imagined the features doing more than this (so in that theory they shouldn't cancel even if based on the same physical recorded strike).

C) What if there are, say, 17 discrete velocity layers, but each layer comprises multiple hits which have been normalized as per their velocity?

Just thinking aloud....... 

Sorry for the late answer, I haven't been logged in here in a while.
A) I don't know. I didn't spend much time with the software (this whole process was time consuming enough  ) I left everything as it were, so if humanization is on by default, then it was on...
What I can say is that there was a quite clear range in which a number of samples would occur and re-occur. Beyond a certain velocity level, the earlier ones became less frequent, and some new ones was introduced.

B) The actual level of playback of the samples would vary. Let's say that sample A would be triggered at velocity 80 and then again at 83, the one at 80 could actually be louder than the one at 83. This is probably the humanization feature.

C) I don't think so, but the thought did strike me, that there would be different samples played back at different times, so that some of them actually were missing in my test. Since the ear is the most sensitive to rapidly re-ocurring hits of the same sample, it would make no sense in having some samples play back more rarely than once every 127 hits.

[Slap the drummer]

Thanks Krillo - interesting experiment for sure.

Shame SD can't be a bit more open about their procedure.  BFD for instance state very clearly
that their basic library has (from memory) 48 separate hits per instrument (where appropriate)
up to the Deluxe extension which has 127 hits.

For all that .dsnd files from SD still sound good to my ears.

I guess the key question for some of us is:- is it better to turn the humanization features off
when we're exporting for .dsnd ?  Sounds like the answer is yes from what you're saying.

[raptor]

I also noticed some velocity variation when trying to make a kick from SD 2.2. After making the dsnd file, I noticed that velocities varied when they should have increased from low to high. I think I will turn of humanizing and maybe the randomizing feature and try again, especially since I'm only using 11 to 16 individual hits.