Solo violin 2
Posted: Mon Aug 22, 2016 8:50 pm
The second solo violin is done. I'm still working on the first. When I get the other violin done, I'll issue the update torrent.
The two violins are based upon on the existing (old) two NBO violins. I made a list of everything I thought needed improvement. Then I took those violin samples, and used a barrage of FFT filtering and noise reduction to try to eliminate the aspects I didn't like. For example, I dislike the bow sound of violin 2. It's too low pitched and scratchy. Also the sound of the violin itself is a little too nasal. (Could be due to the miking technique). So I had to isolate and remove as much of the original bow sound as possible without losing too much of the "good parts" of the samples. Then I used some bowed string physically-modeling algorithms to create a new bow sound I liked. I also created a modeled violin sound with a strong fundamental, and only the highest harmonics. Finally I remixed my modeled sounds with the processed original to recreate a new composite real/modeled violin.
The result turned out really nice. This sounds nicer than a number of commercial products I've heard. Here's a piece played on the new violin:
https://soundcloud.com/j_e_f_f_g/no-bud ... straviolin
And here's the new violin for those who want to test it now:
http://wikisend.com/download/885286/soloviolin2.zip
In fact, nearly all of the improvements I've made to the other instruments for the past year have been done by using several physically-modeling algorithms (plucked string, bowed string, wind tube, mallet) to create hybrid real/modeled instruments. I'd say about 1/3 of NBO's musicians aren't even human. But it sounds real.
The two violins are based upon on the existing (old) two NBO violins. I made a list of everything I thought needed improvement. Then I took those violin samples, and used a barrage of FFT filtering and noise reduction to try to eliminate the aspects I didn't like. For example, I dislike the bow sound of violin 2. It's too low pitched and scratchy. Also the sound of the violin itself is a little too nasal. (Could be due to the miking technique). So I had to isolate and remove as much of the original bow sound as possible without losing too much of the "good parts" of the samples. Then I used some bowed string physically-modeling algorithms to create a new bow sound I liked. I also created a modeled violin sound with a strong fundamental, and only the highest harmonics. Finally I remixed my modeled sounds with the processed original to recreate a new composite real/modeled violin.
The result turned out really nice. This sounds nicer than a number of commercial products I've heard. Here's a piece played on the new violin:
https://soundcloud.com/j_e_f_f_g/no-bud ... straviolin
And here's the new violin for those who want to test it now:
http://wikisend.com/download/885286/soloviolin2.zip
In fact, nearly all of the improvements I've made to the other instruments for the past year have been done by using several physically-modeling algorithms (plucked string, bowed string, wind tube, mallet) to create hybrid real/modeled instruments. I'd say about 1/3 of NBO's musicians aren't even human. But it sounds real.