i'm loving the sound of the S*L console and eq programs at the moment, they truly do provide a more dynamic, less flat sound in the box. however I'm having the age old Nebula dilemna of running out of RAM early in the mix and having to make render decisions. working in Logic, I have to bounce 24 bit and what I end up with is a flattened 2d sound with the same eq curve which really goes against the reason for using Nebulas in the first place. I could probably best describe the sound as using the Waves S*L channel strip instead of your programs once it's been rendered. Cubase 32bit bounces are a bit better but in the same realm of loss.
I'm hoping you might have some advice on how to work around this somehow until the day we can work in 64bit and it's not an issue. thanks for your time and keep up the great work!
Seriously you think S*L waves identical to nebula at 24bits? I have Waves plugin S*L and V3/V4, and frankly, I record in protools and I make my own mixes it also. Since I put nebula3 bundle pro on my machine,waves do not use for anything. Much digital, PSPaudioware and S****x Oxford is better than waves. I worked with S*L and Neve1073 for a long time, recording samba and mixing. I will not say that nebula is 100% identically, But I assure you that is 95% hardware. When I finish the mix, in protools does not really know the difference. when the songs play on the radio and in wheels samba, in P.A, size is the weight and heat. Frankly, I hate waves plugins by many factors, mainly by the sound. Waves never be equal to the nebula, in this reality. And if you still feel that sound nebula in 24 bits, is skinny or not 3d,I do not know what is happening. several papers posted for download, and can determine the outcome. That agree with me, For what 96khz, is overhead for nothing. cd is 16bits 44.100, Mp3 worse still. and the way the world will walks, already saw eh? what counts is the music, and today what arrives closer the analog is nebula with sure. for a 100% analog sound, better the hardware... if you want the sound closer next the analog, use the nebula. Take a look in my posts, I put music download there,all mixed with nebula. And look who is a core2duo3.0mhz,with 4 gigs of ram. road about 20 nebulas. intend to increase the power of my machine, a amd phenom x6. then I will turn 50 nebulas. I thank every family nebula.
If you have the time, you could re-record the track that you are using with nebula onto a new track. Assign the output of the nebula processed track to a bus, and route that into a new audio track. Record in real time to that new audio track. It takes time, but it saves the audio quality within Logic. you will need some processing power to spare, though, so if you are already maxing out the cpu, you'd have to do a track a time or something.
Our ears can hear many things when we believe something is different. Placebo effect is very common. Take the time to invert the polarity of the 24bit render against the "active" original. See if they cancel out or not. If they don't there is something flawed in the rendering process. Unless your clipping your track over 0dbfs, rendering to 24bit isn't the problem if it exists. 24bit audio has 144db of dynamic range. Much greater then the tools & the gear originally sampled.
yes, while it sounds great, it doesnt work for me because I get processor overloads too often in addition to the RAM maxing out.
"Seriously you think S*L waves identical to nebula at 24bits?"
no, not identical, comparable in relation to the change in depth. S*L waves could've been interchangable with an array of non-Nebula plugins but I used that as an example.
"..intend to increase the power of my machine, a amd phenom x6. then I will turn 50 nebulas."
you won't be able to run 50 96khz sampled Nebulas unless you're 64bit.
"I render files with Nebula over here all the time and it sounds exactly as processed in real time"
well, I'm jealous. what DAW are you using? sample rate/bounce bit rate? are you monitoring carefully at somewhere around K System levels in a treated room on good monitors? and finally, have you bothered to do a null test? I personally favor using my ears over null tests but at any rate, the files dont null.
"If you have the time, you could re-record the track that you are using with nebula onto a new track. Assign the output of the nebula processed track to a bus, and route that into a new audio track. Record in real time to that new audio track."
wow very interesting! I thought i'd have to go out digital and back in through my converters to do this. definitely will give this a go and report back.
so if there's render preview software in development, then I assume the purpose is for my reasons stated above...and to know what you're going to get pre-render instead of being surprised/disappointed.
if anyone wants to test this on their own, take a good sounding drum track, apply generous boosting with your Nebula eqs of choice, render and a/b the render vs live. after you've done proper listening tests, flip the phase on one and see if it nulls. I haven't experienced carbon copy renders in 32bit Cubase or 24bit Logic ever.
rhythminmind wrote:Our ears can hear many things when we believe something is different. Placebo effect is very common. Take the time to invert the polarity of the 24bit render against the "active" original. See if they cancel out or not. If they don't there is something flawed in the rendering process. Unless your clipping your track over 0dbfs, rendering to 24bit isn't the problem if it exists. 24bit audio has 144db of dynamic range. Much greater then the tools & the gear originally sampled.
I wish it was placebo but it's not. my colleague and I have done much testing on this and it's obvious to us, both our ears and null tests are telling us. I'm not accusing 24bit to be the culprit because it happens in 32bit Cubase bounces as well.
I'm not clipping the files or anywhere near it, I'm aware of optimum analog levels. so you say that if the files dont null, something is flawed in the bouncing process..well, yes, I am saying that something is inherently flawed in the bounce process
Give some audio examples will be more specific. How do you null the real-time play with the rendered file when rendered files sometimes introduce unexpected time-shift when nubula(due to large latency I guess) is used??
I tried this experiment in Sonar8.5: Playback in 64bit and record in 24bit File A: render drumloop under 64bit engine and truncate to 24bit File B: Loopback via ADAT out and record the in real-time 24bit (I usually work under 24bit to save more CPU power and render in 64bit engine)
Nudge B forward to align with A. Invert phase of B or A, the result of null is around -40dB. Listen by ear, I can't tell which is which and which sounds better.