After exprimenting for a while with nebula compressors,
I came up with this tweak for lowering artifacts a while back and is used in TNT Dynamics.It helps to mask distortions artifacts. You can try this with any compressor or preamp.
for example take a compressor 1.Use timed kernals
2. Go to the global page , select the program rate and fine tune it to 0.725ms (this program rate for 44.1 and 88.2) use the shift key for fine tuning.For 48 and 96 use 0.750ms
These values can be halved for faster program rates, so 0.725ms would be 0.362ms. Then even faster 0.181 and half again 0.090ms then fastest you can go at 0.045ms. At 0.045 it's better to turn off any smoothening.
Please note that you need to do a neb compressors mod to go below 0.250ms
3. You can use the ahead with the same values to capture transients.
Thats it .
Make sure your ahead is set to 1.0ms in master page.
Also another tweak for the EVF page, this slightly tightens up the sound but its very subtle
If you have any questions about these tweaks let me know. If you want to expriment with a preset you can use my focus comp. I did a update sometime ago taking advantage of the first tweak mentioned here. So you may already have that mod. The np2 for that is here https://dl.dropbox.com/u/21072780/FOCUSCOMP.N2P
Last edited by Tim Petherick on Tue Nov 12, 2013 1:07 pm, edited 9 times in total.
I thought G had said that alaising would be a problem with too faster program rates and I assumed that nebula stopped at 0.250 for that reason? I know when you go very fast things don't sound quite right. So on the release I tried to work the ahead times better so it would still capture the transient. On analysis the attack samples still let a little bit through, around 50 micro seconds. Which is what the hardware is doing.
Anyway I found going down to 0.181 actually sounded close too but then the ahead was acting strange, also CPU goes up by a lot,so I wanted users to not have to worry about making a seperate neb compressors. As it could be a bit confusing for new users to nebula.
About the ahead by all means you can go faster if you want to experiment with longer ahead times!
The other thing I wonder if nebula reverb(not very much at all) is alaising anyway. I saw some analysis on gearlutz showing neb reverb did a small amount but nebula regular did not?! It was in a thread comparing alaising of different plugins. Although nebula reverb appeared to be very low on alaising compared to a lot of the plugins.
Last edited by Tim Petherick on Fri May 03, 2013 6:05 pm, edited 2 times in total.
i always thought that nebula reverb and nebula regular were the same, except for the xmls. the dlls are identical i think. something about the settings in the xmls could cause more aliasing though.
i've noticed that if you switch prog rates it alters attack and release times. so if you take a comp program that's at 2.5ms prog rate and take that down much lower, you will be getting different attack/release times.
Last edited by Cupwise on Wed May 08, 2013 1:42 pm, edited 1 time in total.
I thought that if you are going to faster program rates the speed is going closer to realtime reaction. So, it would be affecting the attack and release in that way? Or is it that it actually changes the characteristic, I did not know that was the case. So I thought the ahead time was made to compensate the slower program rates I maybe completely wrong here?
don't know if anyone has noticed the diffrence in selecting different dsp buffers too.....
Yes it should improve. But not totally eliminate the ringing found in many library's. The more reduction the library is capable of the more of this ringing you get, I'm guessing it comes from fast sample changes in the engine.....
The ringing is generally not noticed on drums, because of drums complex frequency response at any one time (drums are note one musical note) ,on a clean low end sine tone is where you'll here it the most. I've noticed this artifact for a couple years when I was making nebula run at faster program rates. On bass guitar you will hear it the most in general.
Hi Tim, this looks interesting, do you still recommend this tweak? If so I have some questions.
1) If using a 96k preset in a 44.1k project, do I set the program rate to 0.725ms or 0.750ms?
2) How does this relate to the program rate in the xml file? I've read discussions in a number of threads on setting it to a low value like 100, but I'm unclear about how it relates to the program rate in a preset's global page.
3) Does setting the ahead to 0.725/0.750ms as well have any benefit for tapes and preamps, or is this just for compressors.
The program rate i'm speaking of is in nebula's global page not a xml/master page setting.... This tip essentially masks ringing, so really is not always going to be perfect, it was a work around. Since this tip can only achieved in timed mode it's very high cpu. I recommend that you only try this tip if something had a relatively low kernal length like a compressor. With preamps you'll get too high cpu in most cases.
The program rate recommendations for xml were something else...but are related.
each preset's program rate can be tuned in the global page. When I wrote this tip nebula was only able to get to 0.250ms without modification. We then needed a mod to the nebula xml to get even faster rates. So this is where the 100 you speak of came from. Now we don't have to mod nebula to get faster rates, the limitation was unlocked last year.
Now I tune things differently from the tip above but some of my old releases still use this tip..