Unwanted detuning defect in Nebula libraries: Please Read!

Tips & tricks, working results, technical support
scene
User Level II
User Level II
Posts: 28
Joined: Wed Aug 18, 2010 5:20 pm

Unwanted detuning defect in Nebula libraries: Please Read!

Post by scene » Fri Mar 07, 2014 5:29 pm

From the day I started using Nebula, I noticed that some libraries have a very strange defect that is easily noticeable on acoustic instruments, especially vocals, violins, cello, piano etc.
It usually sounds like some kind of detuning/dithering like grain that is not pleasant at all. Originally, I thought perhaps this is some kind of analog artifact that comes from layering very rich harmonically saturated libraries, but then it just didn't make any sense.

The reason it didn't make any sense because I can stuck up huge analog-hardware chains with up to 9 tube/transformer based devices and drive them really hard without getting this type of artifacts.
First, I noticed this defect when I got some STN tape libraries and I sent him a message about it but unfortunately never heard back. I wanted to investigate it further, but in the end I stopped using Nebula up until recently when I bought Tim's latest delay library, and decided to re-authorize Nebula once again.

Tim sent me his 660 pass through C program and since I know this piece of equipment very well, I decided to try it on the materials where it would normally shine -vocals. I immediately noticed the defect I described above. I pointed this strange behavior to some of my friends at a local high end studio and we all agreed that this is just cannot be a desirable effect. The interesting thing is that despite having en extremely high end monitoring chain they did not notice it until I pointed it out to them.

Anyhow, I sent a before&after file to Tim (cupwise) and I think this should be investigated because it affects many Nebula libraries.

Thank you

Will The Weirdo
Vip Member
Vip Member
Posts: 405
Joined: Tue Apr 06, 2010 10:15 am
Location: On the lake
Contact:

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by Will The Weirdo » Fri Mar 07, 2014 6:12 pm

Please post your examples and your exact Nebula settings, process, OS, DAW... etc. I want to hear what you are describing, then see if I can duplicate the issue on my system, others may want to try also.
System 1 - i7 5820K Win 7 x64, Studio One 3.5, Reaper 5.4, Live 9.72

System 2 - AMD 1100T Win 7 x64, PT HD2 10.3.7

"In cognitively demanding fields, there are no naturals." - Simon and Chase

User avatar
david1103
Expert
Expert
Posts: 517
Joined: Wed Mar 31, 2010 12:26 am

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by david1103 » Fri Mar 07, 2014 7:02 pm

I want to hear this also!

Please post the dry clip as well as the version with the effect.

How about with a Nebula EQ rather than preamp? That will not have dynamic layers, so if you don't hear it on that we can narrow down the cause very quickly I hope.

RJHollins
Expert
Expert
Posts: 4011
Joined: Sun Mar 28, 2010 5:53 pm

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by RJHollins » Fri Mar 07, 2014 8:40 pm

my guess [without seeing anything] ... incorrect samplerate conversion.

? do you see a flashing '!' on the main Nebula GUI next to the sample rate ??
i7-5820k, MSI X99A Plus, 16 GIG Ram, Noctua NH-D14, Win-7 Pro [64-bit], Reaper-64

NVC [Nebula Virtual Controllers]

SpliffSplaff
User Level III
User Level III
Posts: 37
Joined: Fri Dec 28, 2012 5:46 am

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by SpliffSplaff » Fri Mar 07, 2014 9:13 pm

Could you please post the libraries that have this Problem? Thanks

jorismak
Vip Member
Vip Member
Posts: 356
Joined: Fri Nov 16, 2012 4:49 am

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by jorismak » Fri Mar 07, 2014 9:40 pm

the thing is, some devs use null-tests to make sure the library is giving the best result possible. If a preset (almost) nulls out, it's doing its just nicely, and any detuning effects would be apparent during null-testing.

That makes me think there is something somewhere not right on your setup. I hope we find it, cause it's annoying to have problems like that.

mathias
Expert
Expert
Posts: 2736
Joined: Wed Mar 31, 2010 12:25 am
Location: South-West Germany

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by mathias » Sat Mar 08, 2014 12:57 am

i am interested in testing this also.
please more specifics and a testscenario we all can work with.

as the soundissues you are describing are not very obvious, it could be many things, from samplerateconversion to kernelswitching (possibly related to dynamic layers), or some issues that come from sampling.

i am not a fan of listing programs with "potential" defects, as this may mislead users, to just avoid them. and this can be very unfair to developers!

so please let's just try to test this more thorough, before we come up with any conclusion.
and then the first adress would be the developers, to give them a chance to acknowledge and maybe update the programs.

best,
mathias
system 1: windows 8 32 bit - samplitude pro x3, waveform 9 (tracktion), reaper
system 2: mac osx el capitan - reaper(64bit), waveform 9 (64bit), mixbus 6

both systems on: macbook pro (late 2009), core 2 duo 3,06 ghz, 4 gb ram, graphic: nvidia geforce 9600M GT 512 MB

scene
User Level II
User Level II
Posts: 28
Joined: Wed Aug 18, 2010 5:20 pm

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by scene » Sat Mar 08, 2014 1:59 am

I had this problem with about 5 distinctly different computer setups plus the studio that I mentioned has yet another totally different system. However, I would not exclude the possibility that something could be wrong with my system or there something wrong with the internal settings of Nebula.

This is not a sampling rate error, I checked it and rechecked it again.

All my recordings are done at 32/96 and currently, I use only programs that are sampled natively at 96khz so there is no need to sample convert.

I haven't noticed it with any EQ libraries but perhaps because nowadays I mostly use VSTs algo and hardware EQ.
This happened with Tape libraries from STN, some tape/tube programs from CDSM(not entire libraries).
Cupwise 660...I'm sure there is more but in order to streamline and simplify my Nebula Menu, I normally uninstall libraries that I don't use. Over the years, I bought just about every library out there but I mostly use 5 perhaps 6 programs at the most.

Here is my current computer set up:

Windows 7 64, all latest updates
Asus SABERTOTH Z87
Intel Core i7
ASUS GTX770-DC2OC-2GD5 GeForce GTX 2gb

Samplitude Pro 11
Reaper 4.x
RME UCX latest drivers
RME HDSPe AIO latest drivers
Latest Nebula (installed few days ago)
this particular vocal snippet was recorded through a very clean path:

N 49-----> Forssell pre----> Purple MC77----->Cranesong HEDD.

One is before and the other one with Tim's 660 pass through C, everything at default value.
Pleas listen to the phrase "Do what you were told..." and pay attention to what is happening. It is especially notable when she sings "you".
Attachments
Nebula test without 660.wav
(1.78 MiB) Downloaded 168 times
Nebula test with 660.wav
(1.78 MiB) Downloaded 165 times

jorismak
Vip Member
Vip Member
Posts: 356
Joined: Fri Nov 16, 2012 4:49 am

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by jorismak » Sat Mar 08, 2014 2:29 am

it seems _very_ noticeable when you flip the phase on one of the files :). Little robotic something that gets added. Still sounds like a rounding error or changes from vector to vector. Did you play around with the liquidity (liqdt) control?

Do you have it in both Reaper and Samplitude? Wouldn't be surprised if it had something to do with certain buffer sizes in your DAW or inside nebula. But other than those (useless) remarks, I'm at a loss here. Something is clearly wrong with those files though.

edit: Tried it with CDSM 160 D*x program (only nebula compressor I have), STN Studl, STN NagTape, AlexB CLC and AlexB VBC + SideCar console. I tried it by applying the nebula preset to your 'without 660.wav', and bouncing to a mono 32bit FP 96000khz wav file. Comparing them (and doing null tests) does NOT reveal the effect you describe, while it's clearly there on your 'with 660.wav' file. So I don't seem to have the problem... keep searching.

mathias
Expert
Expert
Posts: 2736
Joined: Wed Mar 31, 2010 12:25 am
Location: South-West Germany

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by mathias » Sat Mar 08, 2014 4:15 am

thanks for the files!

it sounds like a resonating type of sound, that comes in in the words you mentioned. it is clearly audible.
like someone is humming (through a plastictube :D ) in a different pitch together with the voice.

pm me a few other libraries where you get this, so i can try to reproduce it, when i have them too. i don't have the 660 you used here.

i remember a few times hearing such resonating in nebula, but it was with real rooms i sampled and had to do with roomresonances and the sampling exaggerating it.

best,
mathias
system 1: windows 8 32 bit - samplitude pro x3, waveform 9 (tracktion), reaper
system 2: mac osx el capitan - reaper(64bit), waveform 9 (64bit), mixbus 6

both systems on: macbook pro (late 2009), core 2 duo 3,06 ghz, 4 gb ram, graphic: nvidia geforce 9600M GT 512 MB

mathias
Expert
Expert
Posts: 2736
Joined: Wed Mar 31, 2010 12:25 am
Location: South-West Germany

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by mathias » Sat Mar 08, 2014 4:22 am

i would call it a distortion artefact, maybe harmonic kernels driven to resonance on specific frequencies.
system 1: windows 8 32 bit - samplitude pro x3, waveform 9 (tracktion), reaper
system 2: mac osx el capitan - reaper(64bit), waveform 9 (64bit), mixbus 6

both systems on: macbook pro (late 2009), core 2 duo 3,06 ghz, 4 gb ram, graphic: nvidia geforce 9600M GT 512 MB

mathias
Expert
Expert
Posts: 2736
Joined: Wed Mar 31, 2010 12:25 am
Location: South-West Germany

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by mathias » Sat Mar 08, 2014 6:23 am

i was able to recreate the resonating artefact with the demo1 of the 660 on your "without" sample.
a bit quieter, but still audible.
in this case it seems to be related to the program.
it is related to the attack time here. when i set the attack higher, the artefacts go away. setting it to 10 ms was enough to make them dissappear.

i don't know if you have the attack available on the passtrough program, when you do, set it higher and all is good. it doesn't change the sound, when compression is not happening.

best,
mathias
system 1: windows 8 32 bit - samplitude pro x3, waveform 9 (tracktion), reaper
system 2: mac osx el capitan - reaper(64bit), waveform 9 (64bit), mixbus 6

both systems on: macbook pro (late 2009), core 2 duo 3,06 ghz, 4 gb ram, graphic: nvidia geforce 9600M GT 512 MB

Cupwise
Expert
Expert
Posts: 1045
Joined: Tue Nov 23, 2010 2:03 am

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by Cupwise » Sat Mar 08, 2014 10:13 am

so i've been doing some testing

* it's not a problem with this or that specific library having some kind of 'defect', imo. but i guess it would come down to semantics. i'm thinking it has to do with how nebula handles harmonics, and that programs with higher harmonics cause it to happen.

* it doesn't have anything to do with dynamics (so it can happen with static EQs)

* it has to do with the sampled harmonics

* you can hear it best with programs with high level harmonics

* if you disable the harmonics in 'kern' page by reducing number of them to 1 (clean), it completely goes away

edit- ok actually i seem to be wrong on a few of those points. it doesn't seem to go away when you switch to 1k clean on kern page... i'm still pretty sure it's coming from the harmonics though, as weird and contradictory as that might sound.

also, the attack/release time doesn't seem to affect it over here. you can adjust that by going into edit/evf page.

edit again- ok sometimes it does go away when switching to 1k, sometimes it doesn't, which is weird and confuses me, but i still think it's coming from the harmonics.

scene
User Level II
User Level II
Posts: 28
Joined: Wed Aug 18, 2010 5:20 pm

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by scene » Sat Mar 08, 2014 5:41 pm

Tim, you're the only developer who actually responded to my query. I didn't mention it, but about a year ago I also sent a message to Michael Angel (cdsm) and AITB. I didn't hear back from them.

The point is that "real" equipment does not produce this type of d-effect. I have a good selection of tube/transformer based unites and I can stuck them up and drive them anyway I want and they would never sound like that.

As I said it is very easy to hear this "artifact" on any monophonic, acoustic sound.

P.S Sorry for repeating myself

kels
User Level XI
User Level XI
Posts: 165
Joined: Fri Oct 11, 2013 8:43 am

Re: Unwanted detuning defect in Nebula libraries: Please Rea

Post by kels » Sun Mar 09, 2014 9:47 am

Hi,

Here, the problem disapear completely if you use the 660 Pass-Through C_SHQ version of the preset.

If you switch the 660 Pass-Through C (non SHQ version) to pure time mode, kernels are limited to something like 10ms in that case, seems to help but that ringing robot fx occurs at a higher frequency anyway, less obvious at first)

Post Reply