Login

Transient loss

Tips & tricks, working results, technical support

Re: Transient loss

Postby Cupwise » Wed Mar 14, 2012 11:48 pm

the render mode with timed mode would probably be the best solution. an instance that only uses timed would be bad because it forces you to use high cpu mode for every single instance, and it would bring your cpu to it's knees. with render mode, you can tweak nebula parameters within a program to how you want, then it would keep that setup but use better mode on render. it's the obvious choice, IF:
1)everyone agreed that it was actually better sounding.
2)if it could be done without changing the sound in any way besides quality. so the kernel length and program rate and whatever else would have to be exactly the same. most pre programs have 50ms lengths so that would have to be kept with timed.

back to the first point though, i've personally never really checked into it much myself because my cpu can't handle timed, and i've been satisfied with freqd enough that i've never bothered to render stuff with both to a/b. as long as this thread is, i haven't seen anyone organize a blind test like they do over at gearslutz, nor have i seen that most people have taken it upon themselves to really compare the two modes thoroughly and come to a decision. again, you'd have to make sure that kernel length and prog rate were the same with both timed and freqd or it wouldn't be a straight up a/b. maybe if enough people actually mentioned doing this comparison and hearing a difference, eventually something would be done.

basically i'm saying that i think this thread needs a little focus (understatement). you still have people not knowing whats going on after 9 pages. thread won't get any where that way. topic interests me enough that i'd participate more myself but i'm busy with sampling :)

people HAVE said they hear improvement (specifically with transients), but they usually haven't specifically said they kept length/prog rate the same (which may alter freq response and envelope behavior). other people just make comments without mentioning if they actually tried it, and still other people somehow don't know what to do yet. even other other people now don't want to use nebula at all anymore because they think it isn't highest possible quality, which is a little ridiculous because just before this thread they thought it was better than other plugs. if this thread were a petition of some kind, signatures would be written in crayon, paw prints, X's, hieroglyphics, and the stated request would be a '?'.

timed was an extra when it was first included, it was there for tweakers. G didn't have to include it at all. he explained it a little in the manual, said it can sound a little better but uses more cpu. it's ALWAYS been there for people who wanted to experiment with it. it's not some cryptic mystery. it's right in the manual. there's no reason to get all outraged about some secret quality parameter that was kept hidden from you, because it's not secret- it's in the manual. it's been explained over and over. it takes more cpu. it might sound a little better. it's not any more complicated than that. back when nebula first came out it was pretty much impractical to use at all except for splith, but maybe cpu speed has advanced enough that timed can start being thought about for actual use in some way. this thread really needs some organization though...

just my 2 cents, nobody take offense.
Cupwise
Expert
Expert
 
Posts: 773
Joined: Tue Nov 23, 2010 2:03 am

Re: Transient loss

Postby yr » Thu Mar 15, 2012 12:28 am

http://www.4shared.com/zip/ejn7Azfe/NebulaFun.html

- A small test using MWC (line+buss) followed by GMC (out) and NAG30 on a short drum rhythm. Included are the unprocessed file, a "timed" version and the "Freqd" one. I kept the kernel lengths identical. I'll make some additional loops when I have more time..
Reuven | post-production & sound-design | scenography |
website | nebula presets
User avatar
yr
Beta Tester
Beta Tester
 
Posts: 438
Joined: Mon Mar 29, 2010 9:04 am
Location: Amsterdam

Re: Transient loss

Postby dpclarkson » Thu Mar 15, 2012 1:51 am

he explained it a little in the manual,
said it can sound a little better but uses more cpu.


Cupwise, that's EXACTLY what I feel it does.
Using the TIMED parameter thickens the sound
a bit. It's no 'night and day'-experience, but it
fills a hole, sonically.
Not all programs will automatically sound
better, but I found that equalizers do a better job
with the split-hybrid/timed kernels.
Some programs react 'funny' and introduce ringing
sounds when all kernels are selected to TIMED.
And besides, what can be better than tweaking
the engine to your own specific taste?
dpclarkson
Member
Member
 
Posts: 320
Joined: Thu Apr 22, 2010 9:53 am

Re: Transient loss

Postby RJHollins » Thu Mar 15, 2012 2:42 am

yr wrote:http://www.4shared.com/zip/ejn7Azfe/NebulaFun.html

- A small test using MWC (line+buss) followed by GMC (out) and NAG30 on a short drum rhythm. Included are the unprocessed file, a "timed" version and the "Freqd" one. I kept the kernel lengths identical. I'll make some additional loops when I have more time..


Thanks for posting a test 'yr'.

I'll reserve comment for the moment [till others check it out ... minimize bias].

One question, for clarity ... which filename represents the default NEBULA setting ??? I thought it was 'splith' ?!?.

Thanks.
i7-5820k, MSI X99A Plus, 16 GIG Ram, Noctua NH-D14, Win-7 Pro [64-bit], Reaper-64

NVC [Nebula Virtual Controllers]
RJHollins
Expert
Expert
 
Posts: 2627
Joined: Sun Mar 28, 2010 5:53 pm

Re: Transient loss

Postby edbilleaud » Thu Mar 15, 2012 6:08 am

Cupwise wrote:basically i'm saying that i think this thread needs a little focus (understatement). you still have people not knowing whats going on after 9 pages. thread won't get any where that way. topic interests me enough that i'd participate more myself but i'm busy with sampling :)

people HAVE said they hear improvement (specifically with transients), but they usually haven't specifically said they kept length/prog rate the same (which may alter freq response and envelope behavior). other people just make comments without mentioning if they actually tried it, and still other people somehow don't know what to do yet. even other other people now don't want to use nebula at all anymore because they think it isn't highest possible quality, which is a little ridiculous because just before this thread they thought it was better than other plugs. if this thread were a petition of some kind, signatures would be written in crayon, paw prints, X's, hieroglyphics, and the stated request would be a '?'.
just my 2 cents, nobody take offense.


OK, no offense taken, but I stand by all my previous statements, both pro and con. I am a paying customer, just like you, and entitled to my own evaluation of the product, just like you.
edbilleaud
User Level VIII
User Level VIII
 
Posts: 89
Joined: Fri Sep 03, 2010 8:38 pm

Re: Transient loss

Postby giancarlo » Thu Mar 15, 2012 7:13 am

nebula is already tuned for the best compromise between quality and cpu load. You should take in account 3 things
- minimum cpu requirements (99% cpu on i7 is 400% on single core)
- it's still a realtime plugin (you want to use in realtime)
- you want the same sound from a single core or from i7, and it should be enough stable on different sources (bottom drums, piano....)


obviously everything is a compromise
User avatar
giancarlo
Founder
Founder
 
Posts: 9171
Joined: Mon Sep 21, 2009 10:40 pm
Location: Italy

Re: Transient loss

Postby musicgreator » Thu Mar 15, 2012 1:18 pm

On the EVF page i can flip between 5 pages on the right upper corner. For the suggested EVF 17 tweak, do i need to choose EVS17 on every of those pages? Page 5 is "Env Type: OFF" on default. When i want to use SPLIT mode with TIMED engine for the first milliseconds, do i need to do anything to keep the original kernel length?
musicgreator
User Level XI
User Level XI
 
Posts: 161
Joined: Sun Mar 28, 2010 5:16 pm

Re: Transient loss

Postby yr » Thu Mar 15, 2012 1:44 pm

http://www.4shared.com/zip/sAUKHW91/AnotherTest.html

-Another test using a longer chain: MWC(line)-MWC(buss)-r2r(studer -10 30) -MLC(line)- MLC(buss)- NAG30

Average sound levels are matched. Classic mode (default) was used for all instances. Kernel lengths are identical (50ms for all other then the NAG30-33ms).
Reuven | post-production & sound-design | scenography |
website | nebula presets
User avatar
yr
Beta Tester
Beta Tester
 
Posts: 438
Joined: Mon Mar 29, 2010 9:04 am
Location: Amsterdam

Re: Transient loss

Postby david1103 » Thu Mar 15, 2012 1:54 pm

thanks for your long post cupwise, this is all good! and thanks yr for the example, no time to check it right now.

i have spent a whole night testing TIMED last week and am hoping to give a load of input soon, i am just really busy right now.

dpclarkson
Using the TIMED parameter thickens the sound a bit. It's no 'night and day'-experience, but it fills a hole, sonically.


THIS sums up what i thought modding the fate comp PREAMP (line in program) with vocals and drums as sources.

i think across a mix it would make a big difference.

i won't be able to be happy without using the timed version now on this pre at least, and it takes 90% of my quad core cpu, so would REALLY like a way to render in highest quality very much!

render mode could have all sorts of cool stuff going on. for example, i have always thought in a program like reaper why not on render have it upsample everything to 96k! (yes, i know about all the arguments, i am talking about advantages of VST/VSTi mostly running at a higher rate)

i doubt nebula would re-load the 96k versions of programs at the moment, but each standard program could have a link to another higher quality program that it loads on render mode? THIS could be the timed version, the 96k version, whatever mod you like :)

is it possible?
User avatar
david1103
Beta Tester
Beta Tester
 
Posts: 516
Joined: Wed Mar 31, 2010 12:26 am

Re: Transient loss

Postby yr » Thu Mar 15, 2012 2:15 pm

http://www.4shared.com/zip/PeS_MDeh/last_test.html

- same chain as the previous test, this time with classical music (live recording).
Reuven | post-production & sound-design | scenography |
website | nebula presets
User avatar
yr
Beta Tester
Beta Tester
 
Posts: 438
Joined: Mon Mar 29, 2010 9:04 am
Location: Amsterdam

PreviousNext

Return to Working with Nebula

Who is online

Users browsing this forum: No registered users and 1 guest