The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Virtual Drums: beating 7-bit MIDI velocity Virtual Instrument Plugins
Old 26th November 2017
  #1
Here for the gear
 
Zathras's Avatar
 

Thread Starter
Angry Virtual Drums: beating 7-bit MIDI velocity

So, I'm comparing virtual drum software, and remembering the bad old days of mixing in Cakewalk, when there were 128 volume levels to be had, and no more (and volume changes in mixdowns were/still aren't sample accurate, more like 1.5 secs. to respond, and they weren't about to do anything to fix it). I'm also remembering a Korg 168RC mixer that also was limited by 7 bits of volume for its automated mixing changes. At the lower end of a logarithmic scale, these jumps will often be 1 db or more. (BTW: I know you see 127 levels of loudness whenever its mentioned, but that's just wrong. Zero is a loudness level too. Seven bits allows 128 possible values.)

So, sez myself to I, let's get a virtual drum instrument that can give you dynamics without huge jumps in volume. Unfortunately, the pickins are slim. Here's an article on it:

Behind The Groove - Virtual Drum Instrument Developers' Round Table

Here, I quote the relevant part:

Quote:
Do you think that the resolution of MIDI is good enough for reproducing real drum performances, in terms of timing and velocity?

Digidesign: "No, and that is one reason why we do not use MIDI patterns. We use a higher resolution. Internally, Strike's timing is sample-accurate, and the level isn't just represented by 127 velocity steps but is entirely continuous."

Fxpansion (Skot): "The MIDI file spec has more than enough resolution in timing, as you can have time quanta down to microseconds if needed; but the MIDI serial transport layer is woefully inadequate — 31.25kbaud just doesn't do it. All major electronic drum manufacturers need to get into USB MIDI — hint, hint, Roland. Yamaha and Alesis are already producing drum brains with USB MIDI connections. Velocity-wise, 127 levels is probably adequate for our purposes; ..." Wrong answer!

Mixosaurus: "Yes, definitely. Don't blame the machines: if a MIDI instrument sounds robotic and machine-like, it is due to a lack of articulations and sound variations programmed into the instrument (by humans!), ..." blah, blah, blah.

Steinberg: "In GA3, we use a much higher resolution than MIDI for playing back internal patterns and beats, so internally we can do away with many of those limitations. Generally speaking, VST Instruments use sample-accurate resolution, so there is no real resolution issue."

Submersible: "Yes, but we are continually seeking ways to improve it. We've found that the more care and detail we put into taking the samples themselves, ..." blah, blah, blah

Toontrack: "Yes, but we do think that there could be ways to make it even better, such as higher resolution."
Am I hallucinating (without assistance)? Is it really only down to two virtual drum softs that have a possibility of more than 128 discrete levels of loudness: Strike 2 and Groove Agent 4? This is 2017. I know that technology in the civilian, non-corporate sector is regressing (soft Chinese steel, el-cheapo Wal-Marx goods, no graphic EQs on boom boxes anymore - just 5 presets), when we always thought it would go forward; with the exception of technology govts. use to survail or control populations (cell phones, computers, Onstar, and ... oh yeah, bigger TV's). Still, software seems to continue to progress (with the notable exceptions of Windows and M$ Office). There's nothing really to stop virtual drum instruments from having noticeable gaps in volume as volume gradually increases. Enough ranting. My specific questions are these:
  1. Are there any other virtual drum plugins that observe more than 128 levels of velocity?
  2. Are there any DAWs that can support them, or are they needed?

I notice that e.g. Reaper has a MIDI editor that allows you to edit the note volume, but only as 128 levels. Reaper does support some 14-bit continuous controllers, to its credit, but that doesn't help if there are two notes at the same time (at least, not without great effort). Looking at videos of Groove Agent, its "Beat Designer" sees the world in 127 discrete levels of loudness. Strike 2 looks like it might possibly support more, but you are given these short little bars to edit, and it doesn't give you any numbers to tell you how loud it is. I know there are some others (e.g. the Battery series, and Komplete Drums), but I don't see any signs of hope in the ones I've checked by perusing their manuals either.

What ultimately is needed for this to happen, is for individual notes to have more than 7-bit velocities. These will have to be supported by the DAW, or the GUI of the plugin. The plugin itself must be able to respond to these velocities. My last hope is that Pro Tools or another DAW can assign note velocities that are more than 7 bits, and Strike 2 can respond to them; or that Strike 2 has an alternate method of specifying more exact note volumes. Are there MIDI drum pads that have a sensible number of volume levels, and do they implement it in MIDI? If so, how? Presumably, there would be a sound module that uses them too (perhaps by Roland). I expect if you go high end enough, you can do better than 7-bit velocities. Is Glen really right when he says virtual drums suck?

Next Generation Guitar Amp Sim Review - RECABINET

This is surely at least one reason why. Does anyone know of any that don't have this limitation?
Old 26th November 2017
  #2
Drum software is realistic enough.
By far the biggest compromise with virtual drum software is the sampling process?
How many accurate velocities can a human produce in the recording. How much time are you going to spend recording multiple hits at various velocities?
How many different timbres are you going to sample?
Probably the most microscopic, in depth, detailed drum product so far offered to the public was Mixosaurus - and it failed.
You had to pay a lot more and you got two or three drum kits with limited snares and toms. I think most people decided they would rather have more choice in kit pieces than uber detailed but fewer drums.

You can never replace the multiple timbre expressiveness of a real acoustic kit.
But then with drum software you get to use incredible kits recorded in incredible studios by world class producer/engineers. So you lose in one respect, but gain in another.
Old 26th November 2017
  #3
Lives for gear
 
jwh1192's Avatar
whats funny is we go for all these Velocity Levels and then Smash the Drums into 2Db of dynamic range or go through one note at a time and make them all the same or just different by 3-4 velocity levels - if we are feeling Groovy that day ...

now, if the sampling is a big jump from one velocity level to the next that is just pain wrong and sounds bad, resulting in the work needed to make it sound more HUman Friendly ..

and then there is the consumer !!! if you program a good part - they will never know or care about the difference ... we do most of this work for US !!! nothing worng with that just a small percentage of the market .. and if noone ever hears your song then did it ever matter ???

sorry, a little rant before american football starts !!
Old 26th November 2017
  #4
Tui
Gear Guru
 
Tui's Avatar
Anything to do with MIDI is truly prehistoric: The protocol, the hardware, the connectors. Sysex could have been THE standard for allowing studio gear to talk to each other... Hardly anybody is still using it.

I note an ever declining interest in musical subtleties. MIDI timing, velocity, and general functionality are examples. Keyboard players rarely use controllers anymore, i.e. breath, foot, pitch-bend.

As a drummer and keyboard player, I feel like I'm stuck in mud... Manufacturers couldn't care less either.
Old 26th November 2017
  #5
But there has been an explosion in next level midi - like Expert Sleepers and NI Reaktor Blocks. Lifting dodgy data driven midi to sample accurate timing.
Old 27th November 2017
  #6
Lives for gear
 

Quote:
Originally Posted by Zathras View Post
So, I'm comparing virtual drum software, and remembering the bad old days of mixing in Cakewalk, when there were 128 volume levels to be had, and no more (and volume changes in mixdowns were/still aren't sample accurate, more like 1.5 secs. to respond, and they weren't about to do anything to fix it). I'm also remembering a Korg 168RC mixer that also was limited by 7 bits of volume for its automated mixing changes. At the lower end of a logarithmic scale, these jumps will often be 1 db or more. (BTW: I know you see 127 levels of loudness whenever its mentioned, but that's just wrong. Zero is a loudness level too. Seven bits allows 128 possible values.)

So, sez myself to I, let's get a virtual drum instrument that can give you dynamics without huge jumps in volume. Unfortunately, the pickins are slim. Here's an article on it:

Behind The Groove - Virtual Drum Instrument Developers' Round Table

Here, I quote the relevant part:



Am I hallucinating (without assistance)? Is it really only down to two virtual drum softs that have a possibility of more than 128 discrete levels of loudness: Strike 2 and Groove Agent 4? This is 2017. I know that technology in the civilian, non-corporate sector is regressing (soft Chinese steel, el-cheapo Wal-Marx goods, no graphic EQs on boom boxes anymore - just 5 presets), when we always thought it would go forward; with the exception of technology govts. use to survail or control populations (cell phones, computers, Onstar, and ... oh yeah, bigger TV's). Still, software seems to continue to progress (with the notable exceptions of Windows and M$ Office). There's nothing really to stop virtual drum instruments from having noticeable gaps in volume as volume gradually increases. Enough ranting. My specific questions are these:
  1. Are there any other virtual drum plugins that observe more than 128 levels of velocity?
  2. Are there any DAWs that can support them, or are they needed?

I notice that e.g. Reaper has a MIDI editor that allows you to edit the note volume, but only as 128 levels. Reaper does support some 14-bit continuous controllers, to its credit, but that doesn't help if there are two notes at the same time (at least, not without great effort). Looking at videos of Groove Agent, its "Beat Designer" sees the world in 127 discrete levels of loudness. Strike 2 looks like it might possibly support more, but you are given these short little bars to edit, and it doesn't give you any numbers to tell you how loud it is. I know there are some others (e.g. the Battery series, and Komplete Drums), but I don't see any signs of hope in the ones I've checked by perusing their manuals either.

What ultimately is needed for this to happen, is for individual notes to have more than 7-bit velocities. These will have to be supported by the DAW, or the GUI of the plugin. The plugin itself must be able to respond to these velocities. My last hope is that Pro Tools or another DAW can assign note velocities that are more than 7 bits, and Strike 2 can respond to them; or that Strike 2 has an alternate method of specifying more exact note volumes. Are there MIDI drum pads that have a sensible number of volume levels, and do they implement it in MIDI? If so, how? Presumably, there would be a sound module that uses them too (perhaps by Roland). I expect if you go high end enough, you can do better than 7-bit velocities. Is Glen really right when he says virtual drums suck?

Next Generation Guitar Amp Sim Review - RECABINET

This is surely at least one reason why. Does anyone know of any that don't have this limitation?
If anyone thinks that MIDI drums is a bad thing, then you're really looking for a session drummer.

First, I'm not obviously knocking session drummers, but when I have a song that I'm working on and it's 2am in the morning and it's right now that I want to explore with drums, I can't get a session drummer and rent a studio at 2am, that's why using MIDI drums is great.

After I record with MIDI drums, if I still (since I'm the songwriter) think the drums needs to improve, then I have to decide if I'm going to pay for a session drummer.

That's why I have no issue with MIDI anything.
Old 28th November 2017
  #7
Here for the gear
 
Zathras's Avatar
 

Thread Starter
So far, I've seen five reasons not to want what I'm asking for. I'm not saying you're wrong in most or all of your basic assertions. MIDI IS pre-historic in terms of modern computing; much like CD's for music (and iPods and iTunes are worse in sound quality). I'm just asking how much $$$ does one have to pay for such an animal, and sayin' we all should be calling for better.

I'm not sure if ValliSoftware is making the argument 'Let's keep MIDI bad, since it will help session drummers.' I don't find this reason not to want what I'm asking about is compelling, however.

Sequenced drums have a place IMHO. I'm not sayin' it's good music, but look how many hits use 'em these days. Yeah, some people squash their drums into narrow ranges, but, e.g. suppose you have drums going on below the threshold of your compressors. Agreed, for most stuff jumps of maybe 4.5 dB at quiet levels won't make much difference, since dynamics are all over the place. Yet, below the threshold, responses are linear, and medium-fine dynamics are utterly lost.

These arguments 'We do this stuff for us. Most people can never hear it.' don't convince me. Experts can make tweaks rank amateur casual listeners can pick up right away. After maybe dozens of listens, it will be a bigger deal. It wasn't until dozens of mixes with early Cakewalk that I became frustrated that I couldn't nail the volume of some tracks (although the lag between the drawn curve and the volume actually changing did frustrate me earlier). My argument isn't people will never hear it. My argument is 'Why did Glen say virtual drums suck!'? I don't think using e.g. a dozen samples for 127 velocity layers is near as big an issue as having only 127 velocity layers; since drum hits of similar loudness often sound similar - especially when a humanize option chooses from different hits for a layer. Now, having hits of varying distance from the center of a drum out toward the rim would make a difference. I'm just sayin' it wouldn't be so much harder to have a virtual drum machine that records hits in high precision velocities, and e.g. lets you define them as such in some GUI. I would be astounded if MIDI drum pads didn't have it somewhere. I'm just askin' "Where are such virtual instruments?" and sayin' "We should be asking for better."

These jumps in volume will show up in works that have slow builds or drops in loudness, e.g. in a drum roll - especially when it is at low level. Not so frequent in rock, and especially not electronica or rap, but it does happen. I look forward to recording some of my stuff with a real drummer again, and expect to. I'm just sayin', if I put the effort into sequencing beats for something I may be releasing, if even only to YouTube, I'd like to do it with a virtual instrument that has a smooth volume curve rather than big steps!!! Remember, we're only talkin' multiplying by an int, from a programmer's point of view, rather than cramming velocity into seven of the eight bits in a byte, where it doesn't fit! I have long despised MIDI keyboards for the same reason. Analog ones, mechanical ones, and some samplers undoubtedly do better; but I'll bet you gotta pay the big bucks for such sampling keyboards.

I suppose some software might let you change a patch, and thus, the velocity curve, so you could sequence a special curve for quiet stuff. Still, it limits our expressiveness, ultimately; and it needen't be too difficult. I think maybe it's just a matter of enough people not being fobbed off with easy answers, vs. explaining that it really would improve things.

Part of where I'm coming from is this: I used to do analog mixes, but was a computer programmer, too. I was very intrigued when SAW, the four track version came out - but not yet impressed. Fast forward to the present day, and you can get a pretty good virtual console with Slate VCC or probably Harrison Mixbus. With inputs for yer computer, and/or Kazrog Thermionic, Axe FX II, or a Van Kempen, you can get pretty good virtual guitar and bass amps, not to mention various bass (but not guitar) samplers. You can even get a pretty good Neuman U87ai-inspired mic from Michael Joly for under $500. The dynamics of drums and synths are now the weakest link: by keeping little guys and pros using sampled drums and MIDI keyboards from getting good dynamics in their works. Yes, a real drummer can do things a virtual drum machine can't (and vice versa); and for someone sequencing beats, he or she probably won't bother to. Still, IMHO, big steps in volume is kinda basic.
Old 28th November 2017
  #8
Lives for gear
 

Quote:
Originally Posted by Zathras View Post
I'm not sure if ValliSoftware is making the argument 'Let's keep MIDI bad, since it will help session drummers.' I don't find this reason not to want what I'm asking about is compelling, however.
That wasn't my point at all.
Read what I wrote.
If you guys are so preoccupied with some sort of limitations with MIDI drums, then don't use MIDI drums or MIDI anything.
Use session players.

Are you producing any music at all with any MIDI drums?
I'd like to hear those songs with the MIDI drums.

Post your Soundcloud page or Bandcamp link here.
Old 28th November 2017
  #9
Quote:
Originally Posted by ValliSoftware View Post
That wasn't my point at all.
Read what I wrote.
If you guys are so preoccupied with some sort of limitations with MIDI drums, then don't use MIDI drums or MIDI anything.
Use session players.
Agreed.
Virtual drums have usually got very good dynamics in my experience.
I'm not a fan of hardware midi. I try not to use it if I can.
The so called limitation of 0-127 I don't find a problem at all. The real limitation is the inability or unwillingness to sample every single nuance a real drummer can get out of a drum or cymbal. Dynamics? Not really a problem in my experience.
Old 28th November 2017
  #10
Here for the gear
 
Zathras's Avatar
 

Thread Starter
Quote:
Originally Posted by ValliSoftware View Post
That wasn't my point at all.
Read what I wrote.
If you guys are so preoccupied with some sort of limitations with MIDI drums, then don't use MIDI drums or MIDI anything.
Use session players.

Are you producing any music at all with any MIDI drums?
I'd like to hear those songs with the MIDI drums.

Post your Soundcloud page or Bandcamp link here.
That sounds like what you are in effect saying to me. Don't use MIDI. Don't ask if it can do this. Hire a session drummer. Don't ask if what can be done is being done. Don't call for it to be better. IMHO this is not helpful; just preachy. Never file an enhancement request, just use a session drummer. Quit complaining, and take what you are given. Don't even ask if there's better. Still, people DO use MIDI drummers, and still people DO ask for enhancements.
Old 28th November 2017
  #11
Tui
Gear Guru
 
Tui's Avatar
Quote:
Originally Posted by ValliSoftware View Post
That wasn't my point at all.
Read what I wrote.
If you guys are so preoccupied with some sort of limitations with MIDI drums, then don't use MIDI drums or MIDI anything.
Use session players.

Are you producing any music at all with any MIDI drums?
I'd like to hear those songs with the MIDI drums.

Post your Soundcloud page or Bandcamp link here.
Quote:
Originally Posted by chrisso View Post
Agreed.
Virtual drums have usually got very good dynamics in my experience.
I'm not a fan of hardware midi. I try not to use it if I can.
The so called limitation of 0-127 I don't find a problem at all. The real limitation is the inability or unwillingness to sample every single nuance a real drummer can get out of a drum or cymbal. Dynamics? Not really a problem in my experience.
Self-absorbed much?

As a classically trained percussionist and piano player, I'd very much appreciate better dynamics, timing and sonic expressiveness of electronic gear since I play on it.

I thought the idea of gear was to make music with it, not merely produce background noise.
Old 28th November 2017
  #12
Quote:
Originally Posted by Tui View Post
Self-absorbed much?

As a classically trained percussionist and piano player, I'd very much appreciate better dynamics, timing and sonic expressiveness of electronic gear since I play on it.

I thought the idea of gear was to make music with it, not merely produce background noise.
Like I don't make music?

We obviously have a difference of opinion, no need to make it personal.
I played with the National Youth Orchestra of Wales, studied for four years at music college and have been professional since 1980.

I just find there are much bigger compromises in drum software than the 0-127 velocity range.
I agree, hardware midi timing is appalling, and I stopped using it about 8 to 10 years ago. I actually find that a much more pressing reason to find a new machine protocol than dynamics.
Old 28th November 2017
  #13
Here for the gear
 
Zathras's Avatar
 

Thread Starter
Quote:
Originally Posted by chrisso View Post
Agreed.
Virtual drums have usually got very good dynamics in my experience.
I'm not a fan of hardware midi. I try not to use it if I can.
The so called limitation of 0-127 I don't find a problem at all. The real limitation is the inability or unwillingness to sample every single nuance a real drummer can get out of a drum or cymbal. Dynamics? Not really a problem in my experience.
My suggestion of samples at different distances from the center I think would get a lot of those nuances. I admit that MIDI drum dynamics have been adequate in my experience too. I'm also sure there are cases where I might have humanized drum rolls better than what I was getting by default; or by turning humanizing off and doing it myself. Yes, I know there are samples of rolls, but not necessarily my cadence. I am saying it has really been a problem in my experience, if a minor one. I'm glad it works for you.

I'm saying I, and I'd bet real money, many other musicians, when the sometimes large volume jumps are actually explained to them, would want better, and use it. It may be out there. From a software point of view, it is very doable, even with current plugin architectures; although DAW velocity editing may not yet support it. I know VST3 has added a more fine-grained controller architecture. I'm sure it is at the high end. As a developing artist, I have a decided dislike for artificial and unnecessary limitations on my expressiveness - even moreso from an antiquated MIDI standard that is and can be extended. I believe MIDI notes have an extra byte that could be being used. Isn't that enough, to pose these questions? Could I get more straight answers, and fewer, "Not a problem for me's" or "Hire a session drummer's"? Even negative answers: 'This DAW, or this virtual drum software does not support it' would be more helpful. 'This or that drum module does or doesn't do it.' 'This or that standard could/couldn't be used.'
Old 28th November 2017
  #14
The actual major point being....
The compromise of not being able to tune the drums to your liking. Not being able to find multiple timbres on one cymbal. Not being able to play multiple timbres on snare and toms... ALL make quibbling about 0-127 seem a big sidebar.
If you are a classically trained percussionist, are you so happy about the limited timbre and kit piece choice in drum software that your main gripe is the velocity curve.


That's why I agreed with ValliSoftware - if you are concerned over relatively surmountable compromises, maybe you should not use drum software.
Old 28th November 2017
  #15
Quote:
Originally Posted by Zathras View Post
My suggestion of samples at different distances from the center I think would get a lot of those nuances. I admit that MIDI drum dynamics have been adequate in my experience too. I'm also sure there are cases where I might have humanized drum rolls better than what I was getting by default; or by turning humanizing off and doing it myself. Yes, I know there are samples of rolls, but not necessarily my cadence. I am saying it has really been a problem in my experience, if a minor one. I'm glad it works for you. I'm saying I, and I'd bet real money, many other musicians, when the sometimes large volume jumps are actually explained to them, would want better, and use it. It may be out there. I'm sure it is at the high end. Could I get more straight answers, and fewer, "Not a problem for me's" or "Hire a session drummer's"?
At the moment, I think the top drum software has three head zones, rimshot, middle and off centre.
I'm not saying you don't have a complaint, I'm just not hearing these 'big volume' jumps, not even on EZdrummer.
When we record the drums (sampling session). We record hundreds of hits at every velocity I can coax out of my body.
My point is that it's actually hard to play more than about 25 velocities from soft to loud. I think it would actually be impossible for a human to play 127 different velocities on a drum. So if you have 127 velocity levels in drum software, I can't see how it's a major compromise.
Old 28th November 2017
  #16
Here for the gear
 
Zathras's Avatar
 

Thread Starter
Quote:
Originally Posted by chrisso View Post
At the moment, I think the top drum software has three head zones, rimshot, middle and off centre.
I'm not saying you don't have a complaint, I'm just not hearing these 'big volume' jumps, not even on EZdrummer.
When we record the drums (sampling session). We record hundreds of hits at every velocity I can coax out of my body.
My point is that it's actually hard to play more than about 25 velocities from soft to loud. I think it would actually be impossible for a human to play 127 different velocities on a drum. So if you have 127 velocity levels in drum software, I can't see how it's a major compromise.
You sound as if you have great experience at this sort of thing; but I'm not believing your argument. If drum pads had sufficient resolution, every performance would have tens of thousands of different velocities. Without better equipment, it would be hard to test if the human ear can hear them, but my examples of humanizing quiet drum rolls or slowly building crescendos should be sufficient. Maybe it's like DVD audio: since it's top frequency is above human hearing, the argument was made that we can't hear the difference. After conducting actual tests, it was found that we can hear the difference. Sure, I'd like the other things too, but my point is this: the other things you say I should be more concerned about would take thousands more samples, much more time to record, and many more gigabytes to sample. What I am simply asking about, in its most basic implementation, is something that boils down to multiplying by an int in C++, changing the recorded pattern data, and changing the GUI. Existing sample libraries could still be used. Much easier.

I don't see why I'm still arguing that it would help, when Steinberg and the Air Strike guys have already admitted that they already seen fit to do better in their recorded patterns, in the quote I provided in my OP.
Old 28th November 2017
  #17
I'm not against extending velocities beyond 0-127.
I only argue how much difference it would make to already quite compromised products (drum samples vs real drums).

You could be right about the ability to hear a difference, even if it's subliminal.
As a player who has recorded samples for drum software - if I play 100 hits in a row, I'd be surprised if there were 100 different velocities. When I've had to play my own velocity curves on a real drum, after about thirty hits from whisker quiet, I'm already at full volume. So it's actually harder than you think.
Old 28th November 2017
  #18
Here for the gear
 
Zathras's Avatar
 

Thread Starter
Quote:
Originally Posted by chrisso View Post
I'm not against extending velocities beyond 0-127.
I only argue how much difference it would make to already quite compromised products (drum samples vs real drums).

You could be right about the ability to hear a difference, even if it's subliminal.
As a player who has recorded samples for drum software - if I play 100 hits in a row, I'd be surprised if there were 100 different velocities. When I've had to play my own velocity curves on a real drum, after about thirty hits from whisker quiet, I'm already at full volume. So it's actually harder than you think.
Impressive. Now, here's an easy test for you. Record 100 hits in a row as a .wav file, preferably at 24 bits. Now, slice them up into different samples, and get some software that can express maximum amplitude in terms of 24-bit numbers. I'll bet you real money right here and now, you'll have at least 99 different maximum amplitudes for these hits, but almost certainly 100. I can take that farther. You may be able to find some software that quantizes the amplitudes of those hits. If I ever get more free time, I may do it for you. Now, record a series of separate drum hits, slice them up, gradually going from quiet to loud. You might try it with a real drum line, micing different drums too. Quantize them, and set up blind listening tests (ideally), and see if a difference can be heard, and which is preferred.

You could do a simple test yourself with a series of hits from soft to loud yourself, with 128 levels over the entire range. More definitive tests are possible.

This issue has not progressed since the MIDI standard was drafted way back in the last century, and it really should.
Old 28th November 2017
  #19
Here for the gear
 
Zathras's Avatar
 

Thread Starter
Wait. This is even easier. Record a nice drum line in a 24-bit format. Now downsample it to 7-bits; although it might be a little tricky to find out how to do it. A simple hand-written script in ReaJS or Audacity should do the trick: just zero or randomize the least significant 17 bits. Now compare the two on good speakers. You'll hear the difference. 'Nuff said! There's a reason DAW's "waste" those extra bits of accuracy on dynamics for drum tracks. That should be a much easier test, and should prove to you that dynamics in real drum tracks past 7 bits does matter!!!

You might argue that once it's mixed, you are chopping off more bits of quieter drums, so just lop off all but seven bits of the stems. You'll hear the difference.

Edit: I suppose that is comparing apples and oranges, due to the jagged nature of the resulting waves. Still, all I'm getting is empty assertions that it doesn't matter, instead of help.

When I sequence rock drum lines, most hits are at 100 or more, since the softer hits are not called for most of the time in most passages. That gives me a whole 28 levels of volume. Whoop-de-doo!
Old 28th November 2017
  #20
Lives for gear
 

Quote:
Originally Posted by Zathras View Post
Wait. This is even easier. Record a nice drum line in a 24-bit format. Now downsample it to 7-bits, and compare the two on good speakers. You'll hear the difference. 'Nuff said! That should be a lot easier, and should prove to you that dynamics in real drum tracks past 7 bits does matter!!!
That's complete nonsense.. PCM audio quantization bitdepth and how it affect audio is absolutely different thing, which isn't remotely related to whether you can perceive differences among various triggered samples according to incoming velocity.

Michal
Old 28th November 2017
  #21
Quote:
Originally Posted by Zathras View Post
Impressive. Now, here's an easy test for you. Record 100 hits in a row as a .wav file, preferably at 24 bits. Now, slice them up into different samples, and get some software that can express maximum amplitude in terms of 24-bit numbers. I'll bet you real money right here and now, you'll have at least 99 different maximum amplitudes for these hits, but almost certainly 100. I can take that farther. You may be able to find some software that quantizes the amplitudes of those hits. If I ever get more free time, I may do it for you. Now, record a series of separate drum hits, slice them up, gradually going from quiet to loud. You might try it with a real drum line, micing different drums too. Quantize them, and set up blind listening tests (ideally), and see if a difference can be heard, and which is preferred.

You could do a simple test yourself with a series of hits from soft to loud yourself, with 128 levels over the entire range. More definitive tests are possible.

This issue has not progressed since the MIDI standard was drafted way back in the last century, and it really should.
Err, this is exactly how most drum software is made.
So yes, I've done it many times.
You record as many hits as you can at as many different velocities.
Then you go home and arrange them in velocity order in your DAW. You can use advanced fader software to output a numerical read out of exactly the db of each hit, and you can then lay them out from soft to max volume.

Like I say, it's surprising how many hits are the same velocity and how few velocities there are from soft to loud.
When listening back, the difference of 1db is almost impossible to hear. In addition, timbre variation is much easier to hear. So you can often have a snare hit that appears louder than another one, when it is actually (numerically) lower in volume - but it may have been a fuller tone hit, or have more rim in the rimshot mix.
Old 28th November 2017
  #22
Here for the gear
 
Zathras's Avatar
 

Thread Starter
And yet still, in actual numbers, if expressed as 24 bit numbers (actually 23 are needed), the velocities between, let's say -3 and -4 db are still not usually duplicated, but are still all in between. You haven't proven that we can't hear the difference between 128 level quantized beats and natural beats in a completed work; and I'd wager you can. Again, the difference in quieter material is per force greater than 1 db in quieter material, sometimes much greater.

Again, I am still trying to humanize with just 28 levels of velocity. You still aren't attempting to answer my questions, and you are not being as helpful as you think. I wouldn't mind, but you keep whipping a dead horse. I don't care if you feel that way about it. Without real listening tests of completed works quantized to 1dB and more increments, and without, you are just pontificating; and I don't believe you. I am asking for software and or samplers that does this. I am not asking somebody to troll me and imply I should be asking it.
Old 28th November 2017
  #23
I'm talking about using ACTUAL drum software, played from v-drums, using all the 127 levels of velocity I've been given.
How that is somehow negated by your convoluted extra processing, quantising and 7bit-ising I don't know.

I', talking about an actual acoustic drummers experience of using actual drum software.

The lack of alternative timbres and the fact most packs are very rock oriented, are far bigger compromises than the velocity response IMO.
Old 28th November 2017
  #24
Here for the gear
 
Zathras's Avatar
 

Thread Starter
You won't let up, and you're not even close changing my mind. Let it go.

I am talking about a user's experience of using actual drum software, and cramming my attempts at humanization into 28 loudness levels.

For the second time, the lack of alternative timbres is important. It would be hard to account for this, without needing lots of extra sampling. Accounting for more smooth velocity curves can boil down to just multiplying by an int.

Even yet again, my "convoluted" thought experiment of quantizing an actual drum performance to 1 dB and often more, increments, might be the only way to compare apples to apples, and say that these large volume steps don't matter. It would help if it incorporated a few series of beats that got slightly louder and softer. If it was a natural track to a virtual drum track comparison, too many other variables would interfere. I'm not sure you have the capability of recording a drum performance on drum pads, and getting high resolution velocities, to compare 7 bit velocities to high precision drum velocities on a drum sampler; although if you did, you could speak from facts. You're still not even trying to answer my questions!!!
Old 28th November 2017
  #25
Tui
Gear Guru
 
Tui's Avatar
God, this is depressing. People who claim to have enjoyed a musical education think it's "hard to play more than about 25 velocities from soft to loud". Good grief.

Try programming this with 127 velocity steps:

Old 28th November 2017
  #26
Gear Maniac
 
clump's Avatar
 

Great drumming.....but wasted on a very mediocre song, therefore Jeff might as well been hitting a plastic bucket with a dead fish. This song epitomises everything I disliked about 80s 'AOR' music. And it proves, in my opinion, why individual musical dexterity means nothing and being part of a groove means everything.....I would cite Charlie Watts and 'Honky Tonk Women' as a perfect example of what I mean.
Old 28th November 2017
  #27
Tui
Gear Guru
 
Tui's Avatar
Old 28th November 2017
  #28
Gear Maniac
 
clump's Avatar
 

Quote:
Originally Posted by Tui View Post
Well of course as a "Classically trained percussionist and piano player" (Is that similar to a pianist?) your opinion is far more valid than anybody elses.

Incidentally, the piano is a percussion instrument, so no need to add it to your statement....it's like me saying "I am a classically trained percussionist and triangle payer".
Old 28th November 2017
  #29
Tui
Gear Guru
 
Tui's Avatar
clump, you are now on my ignore list. Do me a favour, stop quoting me.
Old 28th November 2017
  #30
Quote:
Originally Posted by Tui View Post
God, this is depressing. People who claim to have enjoyed a musical education think it's "hard to play more than about 25 velocities from soft to loud".
You'd have had a point if you hadn't had to twist my point and completely misquote me.
I said no more than about 100 velocities - that is still less than the 127 you get in most drum vst's.
But of course you are on shaky ground, which is why you had to make up the 25 velocities figure.
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump