sash poask
Member
- Mar 24, 2008
- 440
- 0
- 16
nothing useful to say, but this is becoming a really interesting read, and i hope, people carry on.....this seems really promising
Re-opening the Nebula for cabs discussion
- Thread starter Ermz
- Start date
as you wrote, differently from other companies we are using volterra series and not just impulse responses. Our models are ACCURATE till 9th harmonic. We have a sort of "dynamic volterra series" which is a replacement of dynamic conv theory. You have a sort of "impulse response" for each harmonic. The first kernel is the impulse reponse. Than you have a kernel for the first harmonic, second harmonic and so on...
Here a tube emulation:
AeternusEternus
Member
- Oct 8, 2008
- 601
- 0
- 16
You guys are right. I guess I just had a bit of a blown top yesterday. I just look at the photos and get all ansy about doing this. Giancarlo, let me beta test and do what I can to get closer to the holy grail here. I'm getting PMs about doing it here and on the Nebula Programs forum, and I just can't let anyone down.
I want the best cab emulation possible. I sent you a PM, Giancarlo, about getting more involved and going further.
I want the best cab emulation possible. I sent you a PM, Giancarlo, about getting more involved and going further.
Thanks bro.
In the mean time, Marcus, would you be willing to give this thing one more shot if giancarlo were to give you a more exact starting point? I kinda feel like we're sitting on our hands here, not able to push this any further without AE.
In the mean time, Marcus, would you be willing to give this thing one more shot if giancarlo were to give you a more exact starting point? I kinda feel like we're sitting on our hands here, not able to push this any further without AE.
GarethSE
New Metal Member
- Jul 5, 2008
- 7,595
- 1
- 0
- 33
I just got Aeternus' reply.
I'm saying no more except for I apolagise for my crankiness in my previous post in the topic. Sleep deprivation does weird things to you, and I realise that's no excuse for being an asshole, but hey.
My point still stands though, but with less harsh wording if that's any consolation. Believe me, I'm behind nebula 100% of the way, it just came out a bit wrong is all.
Apolagies once again. My pm to you still stands however Aeternus.
I'm saying no more except for I apolagise for my crankiness in my previous post in the topic. Sleep deprivation does weird things to you, and I realise that's no excuse for being an asshole, but hey.
My point still stands though, but with less harsh wording if that's any consolation. Believe me, I'm behind nebula 100% of the way, it just came out a bit wrong is all.
Apolagies once again. My pm to you still stands however Aeternus.
Fuck yeah, I'm totally down - someone just tell me what to do
giancarlo, is there any chance that you could give us a good NAT template to use for sampling cabinets?
Metaltastic is quite thorough and will likely post a direct comparison of the real cab alongside the nebula program, so any inconsistencies will be highlighted clearly.
Cool stuff giancarlo, thanks a lot for sharing! I'm curious, what do you guys do for time-invariant stuff like chorus/flange?
we sample a "train" of kernels, like a camera. A single collection of kernels is a snapshot = a photo. We are making a movie, and it's looped automatically by nat (nat finds the loop point automatically). Than nebula synchronizes it with the host tempo. You are able to change the rate, but a program could "quantize" the rate time, so that a chorus or a flanger or a moving panner will be perfectly synchronized with your song even if the host tempo changes. We are proud for a CSL polyphaser and an electro harmonics emulation, they are quite good replicas. This (like many other things, for example a dynamic reverb) can't be emulated using dynamic convolution (or simple convolution)
I add:everything is sampled by your users and betatesters. We are not changing anything. They sample, they test the emulation, and if they find it enough close and good they release it. Programs (libraries) are shared using bittorrents. Everything is free, we don't ask royalties. Only you should own a commercial player, the minimum price is 20 eur. But we released nat3 free with limited features, you could share free programs (till 2nd harmonic) with free users (nebula3 free).
Ok, if I`m getting this right:
- Someone should set up a cab, make a recording of a cab and an impulse which we already know how to do
- Then we should probably start with a simple 1 kernel Nat template and try to get a Nebula program that sounds, spectrum-wise, like an impulse, without the high end harshness that we previously were getting
- From then on, start adding gain dynamics into equation and try to keep the harmonic content equal to the previous program, thus getting the dynamic behaviour while retaining the original sound.
- possibly try to emulate the speaker distortion as well (although, I`m pretty sure that if the previous points go as planned, we`d all be pretty satisfied already)
- Someone should set up a cab, make a recording of a cab and an impulse which we already know how to do
- Then we should probably start with a simple 1 kernel Nat template and try to get a Nebula program that sounds, spectrum-wise, like an impulse, without the high end harshness that we previously were getting
- From then on, start adding gain dynamics into equation and try to keep the harmonic content equal to the previous program, thus getting the dynamic behaviour while retaining the original sound.
- possibly try to emulate the speaker distortion as well (although, I`m pretty sure that if the previous points go as planned, we`d all be pretty satisfied already)
How though? I made samples last time with 1 kernel Nat templates and it had the high-end harshness; what else can I do?
Well, that`s where we expect Giancarlo to kick in, and give us a starting template.
So, how about it, Giancarlo?
A template that should be just a simple impulse without all the mumbo-jumbo to get us started.
Then we can do a round one and get to compare a real cab recording, an impulse, and a simple nebula program. Nebula program should sound about the same as an impulse.
So, how about it, Giancarlo?
A template that should be just a simple impulse without all the mumbo-jumbo to get us started.
Then we can do a round one and get to compare a real cab recording, an impulse, and a simple nebula program. Nebula program should sound about the same as an impulse.
The Unavoidable
jättebög
Ok, here my suggestions:
FIRST STEP
before sampling test your system. Follow this topic:
http://www.nebula-programs.com/viewtopic.php?t=382&start=0
only testing your system we'll be sure that everything is working as expected.
Post there your results (a bypass program).
SECOND STEP
When everything is working as expected (flat frequency response) you should sample a completely wet hardware reverb. You should be sure that no direct signal is present in your path. In general harshness is caused by a direct signal which is going from the outputs to the inputs of the audio card. It could a trouble of asio or even nat (it could be caused by a bug for example). For this reason we ask to sample a wet reverb. We should be sure you are sampling reverbs correctly.
tell me when you are ready with those 2 steps. Than we'll analyze your cabinet and your results, comparing with standard IRs. The result should be perfectly the same (not similar! they are numbers, and we expect the same numbers)
AeternusEternus
Member
- Oct 8, 2008
- 601
- 0
- 16
Ahhh, the good ol' system test. I remember reading this thread, but I never did it. I'm doing it now, and I'll be posting my results in that thread.
The hardware reverb, however, is another story. I don't actually have a hardware reverb unit. It may be a little while before I can acquire one.
The hardware reverb, however, is another story. I don't actually have a hardware reverb unit. It may be a little while before I can acquire one.
do you have something hardware completely "wet" to sample? A lot of times the dry signal is outputted to inputs again, creating a sort of light feedback. Sometimes it's for an internal setting of the asio card, sometimes it's a bug, sometimes it's simply a bad connection. It happens very often. You'll have "harshness", exactly the thing you are describing.
