I get what you’re saying and it makes sense. I’m just wondering why some things seem to play so well and others don’t. As an experiment I just went played around with the snares from the BFD3 library and I’ll try to explain.
So when I load up the Tama Tepesta snare on kit piece or the Bell Brass one, both of those basically play as well as the Platinum Samples snares. Softest hits are clear and defined and there is a perfect progression as I ramp up the dynamics that sounds and feels pretty much exactly like a real drum would. Great. No problems there.
But when I load up the Mapleworks snares, the soft hits sound almost too soft and when I ramp up the dynamics there’s a weird gap where it sounds and feels like there are samples missing between really quiet and medium, and then another gap between medium and harder hits. It’s just not a smooth progression. And the Mapex black panther almost has no mid range at all. Just quiet to loud. And yeah I can turn on the loudness and mess around with the range and curve settings and improve both of them, but they never seem quite right like the Tama snares which just work beautifully simply by loading them.
Out of curiosity I opened the library and literally counted the number of “hit” samples for the Tempesta and the 14 mapleworks and they were both 80, so I’m not sure what’s going on.
This same thing applies to certain hi hats, toms etc…It just seems that overall the vast majority of the platinum samples kit pieces, like almost all, have that perfect sounding and feeling dynamic range without any messing around. There are two snares that I remember don’t play that great and maybe a HH or two. With the BFD stuff it seems to be the opposite. There are some pieces that play perfectly but most sound and feel a bit lacking.
oooooooo… you’re opening one of my favourite can of worms! Sampling philosophy.
First of all… all of the kits in BFD3 Core Library - aside from the Mapleworks one - were recorded by Platinum Samples. So it stands to reason that they’d playback similar to the other Platinum Samples libraries. John Emrich recorded the Mapleworks drumkit.
Number one to me is how the drums are tuned and how the tuning is monitored throughout the session. Makes a big difference! If you are tracking rimshots and you’re not fixing the tuning in between passes, your rim shot tunings are going to be all over the shop. Not in a good way.
Secondly, how the drums are hit. This is super important. You can’t just sample the drums at various intensities - quiet, medium, loud, extra loud, METAAALLLLL … it just doesn’t work. You end up with those “stepped” dynamics like you’ve noticed. You have to record with the knowledge that these sample groups need to blend nicely together. I have a few playing techniques I do when I’m sampling drums, to ensure I’m not creating stepped dynamics.
I actually have it as part of a longstanding wish-list, to revisit the dynamics of all of our content and fix things. Bear in mind we have 17 years worth of content across 35+ expansion packs, with at least 4 different production teams making content throughout that period. There is quite a lot of variance.
Super interesting.
I always wondered why different developers have their own particular sound to their libraries - I assumed the sampling process would be the same for everyone, but apparently not.
Whatever your secret sauce technique is, it’s great, especially on the cymbals.
For me, the trick is to not hit them too hard until the very last 10 layers. Most of the time when a drummer is playing a beat, they’re not aiming to absolutely destroy their cymbals… coz cymbals are expensive!! So why make loads of samples where you’re properly gorilla-ing the cymbals with the sticks? Makes no sense.
Rides are a tricky one too. So many ways to play it, that software doesn’t adequately cover (yet) - you can hit the bow of a ride with the tip of a stick or the shank of a stick, and get very different sounds. Same with the bell and edge.
haha…cool! Didn’t know this was a “thing” but yeah that obviously makes a lot of sense that just in the way the whole process is done will have a great effect down the line. I would imagine the whole “stepped” dynamics thing is much more apparent for drummers on ekits in real time than would be for people who produce with software as you can just “smooth it out” with parameters.
That would be so cool to have the older samples “remastered” to smooth them out because there are so many great sounds. I’d love to have those mapleworks playing smoothly as I really do love the sounds. How has the development of AI tech affected you? At some point it seems like jobs like this might be perfect for AI. (based on my amateur use of chatgpt…lol) Throw the samples at AI and tell it to smooth everything out? haha.
It’s interesting too because most of the time when you over hit cymbals (and drums) it starts to choke them anyway. The hardest hits on a cymbal should still be “reasonable”. With things like the Ride samples, what’s the future? Is it some kind of merging of real samples with modeling to achieve as many possible “colours” in terms of shank/tip/bow/edge etc…? Just based on youtube demos I wasn’t really all that impressed by a certain big company’s modeled drum software just in terms of sound.
It’s very fascinating and your feedback about the issue is very much appreciated! Helps me understand more about why I like certain kit pieces more than others. I’ll have to change my view with regards to thinking velocity layers were mostly the main difference.