Hi, all! I'm new to this sampling thing so please bear with me. I'd like to know how to increase dynamic range (difference between soft and loud) in relation to velocity (1 through 127) in Maize Sampler.
For instance, if I map a sine wave (just one sample stretched from velocity 1 to 127) to A4 and I'd like to increase the difference in volume between soft (low velocities) to loud (higher velocities), how do I do it?
How to adjust velocity dynamics?
-
- KVRer
- Topic Starter
- 5 posts since 16 Oct, 2007
-
- KVRian
- 858 posts since 14 Sep, 2004
Try putting a lp filter on the sound and modulating it by velocity. After setting the filter cutoff freq at the right side of the envelope (experiment with various cutoff points), you'll want to visit the knobs to that control's right toadjust both the ratio of velocity to filter and the key to filter setting. It may take you some time to hear the difference in these settings. The first can be said to control the extent to which a high velocity overcomes the filter. (There are other ways of putting this.) The second controls the degree to which the filter is active from right to left across the keyboard. (If the filter cuts equally across the keyboard, mellowing a brassy sounding bass register may cut the high frequency fundamentals in the soprano range.)
More generally, if you want a wide dynamic range, you'll need to include more than one velocity layer. A few much older piano samples had only two; most people prefer at least 3-5 layers. (And of course there are piano multisamples with over 50 layers, these days...) The problem is partly that a filter cannot distinguish between high frequences that you want to cut (the transient sounds and upper harmonics) and those that you want to keep--the ones that make up part of the basic sound. When you set the filter low to kill the brassy sounds of a hard strike, it therefore often dulls the overall sound. If you set it higher, the low strikes sound too metallic. The many methods of trying to find ways around this situation are more or less the history of creating piano mulitsamples. Recording low level strikes gets around the problem, in part, by just not including as many transients in the first place. Welcome to the journey.
More generally, if you want a wide dynamic range, you'll need to include more than one velocity layer. A few much older piano samples had only two; most people prefer at least 3-5 layers. (And of course there are piano multisamples with over 50 layers, these days...) The problem is partly that a filter cannot distinguish between high frequences that you want to cut (the transient sounds and upper harmonics) and those that you want to keep--the ones that make up part of the basic sound. When you set the filter low to kill the brassy sounds of a hard strike, it therefore often dulls the overall sound. If you set it higher, the low strikes sound too metallic. The many methods of trying to find ways around this situation are more or less the history of creating piano mulitsamples. Recording low level strikes gets around the problem, in part, by just not including as many transients in the first place. Welcome to the journey.
-
- KVRist
- 456 posts since 15 Oct, 2004 from SF, CA
The default one in maize sampler is a linear fader. That means the amplitude of the sample decrease 6db every 20 velocity. I think we'll add more options for users in the next version.
-
- KVRer
- Topic Starter
- 5 posts since 16 Oct, 2007
Jake,
Your tips are really appreciated.
I'd like to do it without use of a low pass filter, so I tried your suggestion of adding more velocity layers. I tried 3 layers and added 3dB gain to the topmost layer and subtracted 3dB to the bottom layer, leaving the middle layer untouched. It did increase the dynamic range but there was a noticeable jump in volume at the velocity transition points. Perhaps this would improve if I increase velocity layers to perhaps 10 or so.
Your tips are really appreciated.
I'd like to do it without use of a low pass filter, so I tried your suggestion of adding more velocity layers. I tried 3 layers and added 3dB gain to the topmost layer and subtracted 3dB to the bottom layer, leaving the middle layer untouched. It did increase the dynamic range but there was a noticeable jump in volume at the velocity transition points. Perhaps this would improve if I increase velocity layers to perhaps 10 or so.
-
- KVRer
- Topic Starter
- 5 posts since 16 Oct, 2007
So that's a maximum of about 36dB dynamic range if I understand correctly.caoxiang wrote:The default one in maize sampler is a linear fader. That means the amplitude of the sample decrease 6db every 20 velocity. I think we'll add more options for users in the next version.
It would be better to have a way to manipulate dynamic range, and I eagerly await the next version.
