| Author
|
Less The Better / More the Better ???
|
WAVELOGIX
Wavelogix
Started Topics :
136
Posts :
1214
Posted : May 17, 2005 21:46
|
hello guys ,
well ... i have this question lingering in my mind since many days .. just thought id ask and get stuff sorted out 
i have a synth line playing .. that i get from a typical synth like vanguard / vsti ... now ... these vsti have enough controls to tweak around , so that one can design an appropriate sounding patch and then filter it just right , so that the patch sits properly in the mix ...... then offcourse a little 1-2 band eq to make things a little more better .....
now i use the WAVES plugin V5 all the time , for all the processing .... my question over here is ... once i have a decent sounding patch comming off albino / vanguard ..... is more processing thw way to go or one can actually settle with less .....
well .... for example ... i could use a compressor - some tube amplification plugin - 4 ban REQ - CHorus - PHaser .. etc etc ...
now when i follow such an effects chain .... dosent my sound signal get degraded with every component that the sound passes through ? becasuse every plugin introduces some noise / artifacts into the signal .. due the processing rates difference ...
if the above is the case .... isint getting a proper sounding patch and then little eq the way to go ? to keep the sound as clean as possible ....
and if multiple effects do not degrade the sound , then i think one could actually use multiple plugin to get the synth to sound way different and better [ depending on how you use them !! ] .........
so just wanted to know ... how you guys go around using plugins .....
btw .. more plugins also mean more audio bounces ! [ another reason for me to use less plugins all the time ]
Thank You ,
respect .. chandan !
  http://www.youtube.com/watch?v=CTPJuMSwzUQ |
|
|
|
Colin OOOD
Moderator
Started Topics :
95
Posts :
5380
Posted : May 17, 2005 22:55
|
Noise will (generally) only be added to a signal when it passes through the analog domain. If your source and processing are all inside VST, no noise will be added, in the traditional sense. However there are other kinds of noise that can be introduced if you're not careful with your bitdepth and the levels to/from each plugin in the chain, so for best results make sure that the input level to each plugin in the chain is as high as it can be without clipping.
Actually, now I think of it, this isn't so important since most (if not all) of the sequencers we use have a 32-bit float data path which pretty much eliminates quantisation noise when processing, but if you're using 16-bit audio you still need to make sure the first plugin in the chain is getting a good healthy level.
There is a valid discussion as to whether it's best to use more or fewer plugins and you will hear good arguments on either side. Me? I use what I need to get the sound I want. What there can be no argument about however is the point you made about the source - getting the initial patch or sample right in the first place is the best possible start for your sounds.
Mastering - http://mastering.OOOD.net :: www.is.gd/mastering
OOOD 5th album 'You Think You Are' - www.is.gd/tobuyoood :: www.OOOD.net
www.facebook.com/OOOD.music :: www.soundcloud.com/oood
Contact for bookings/mastering - colin@oood.net |
|
|
|
Freeflow
IsraTrance Full Member
Started Topics :
60
Posts :
3709
Posted : May 18, 2005 02:01
|
isnt it simple?`
just A/B the sounds?
does it sound better or worse after a certain effect chain.
i my self stick to, "what sounds good to me, is good"
if i need a phaser effect then ill use it, if i need to cut a certain frequency ill do it.. no theory will stop that or fact for that matter...
if i like errors i use them aswell...
i say dont think about theories, break them or reinforce them.. dont let a theory stop you! Take the theory to test first..
but ofcourse ill be true, if i have 20 channels of errors its bound to add up to a real mess, but you wont notice it if its just a few...
i bet there are lots of matematical errors in digital music... but it can still sound good and clean..
|
|
|
|
WAVELOGIX
Wavelogix
Started Topics :
136
Posts :
1214
Posted : May 18, 2005 09:57
|
well , thank you for those insights.... but as i mentioned earlier ... i always use the Waves Plugins Pack and i read the manual which says .... all internal processing of the Waves Plugins is done @ 56khz / 64 khz depending on the plugin and then again is dithered down to 44.1/48k ... depending on the project resolution ...........
im not just too sure ... as to how this conversion from 56k--> 44.1k does to the signal with "each" plugin applied to the sound .....
respect .. chandan ! |
|
|
|
billy ambulance
IsraTrance Junior Member
Started Topics :
43
Posts :
560
Posted : May 18, 2005 11:19
|
some plugins really improve your signal and some only make it worse.
choose them wisely and the force will be with you!!
|
|
|
|
Boobytrip
IsraTrance Junior Member
Started Topics :
39
Posts :
988
Posted : May 18, 2005 17:59
|
|
It's the internal dithering that worries me, ideally you only want to dither once, at the very end of the signal processing chain: as the last step before leaving the digital domain. If you place several plugs in a series and the sound is processed 'inside' a plugin (mabye at a higher bitrate) and dithered, processed and dithered and processed and dithered again, doesn't this repetitive dithering ruin your sound ? In my ears it does, but it could also be a psychoaccoustic placebo effect. |
|
|
|
Colin OOOD
Moderator
Started Topics :
95
Posts :
5380
Posted : May 18, 2005 18:11
|
|
Hayez
Started Topics :
8
Posts :
393
Posted : May 18, 2005 21:32
|
Quote:
|
.... all internal processing of the Waves Plugins is done @ 56khz / 64 khz depending on the plugin and then again is dithered down to 44.1/48k ... depending on the project resolution ...........
|
|
I don't think Waves plug-ins change the sample-rate anytime. Some plug-ins (like the REQ) have double presicion processing. it means that the internal process is done in 64bit float rather then 32bit float. in that case there is bit reduction on the output and dither is applied. i'm pretty sure that Waves has done it in way that won't introduce any noticable noise.
"a new art came into my mind which only you can create, the Art of Noises, the logical consequence of your marvelous innovations." Russolo, 1913 |
|
|
|
UnderTow
Started Topics :
9
Posts :
1448
Posted : May 19, 2005 01:05
|
WAVELOGIX, as others have mentioned, trust your ears to make decisions on this.  And indeed it is usualy best to get the source sounding as good as possible. That having been said, I wouldn't add phasing and chorus and flanging on a sound if it is in a busy mix. that would probably just make a mess of things. It would be best to add one of these effects so that you can actually hear what is happening. But if carefully tweaked it might work on individual prominent sounds like a lead line that is playing on its own.
Colin, Boobytrip, in a fixed point maths system, dither should be applied every time the the bit depth is reduced. The point of this is to _increase_ the resolution beyond what is normaly possible at any given bit depth. This way you end up with a signal that has more resolution and less quantisation errors.
As Boobytrip mentions, you do add a slight amount of noise to the signal when you do this (4.77 dB to be exact) but if you think of 24 bit audio with processing working at 48 bit (Waves L2 for instance), that noise is at -139 dBFS. Not only that, the noise is shaped so as to be minimised in the frequency bands our ears are most sensitive to. So in practise it isn't really an issue unless you use ALOT of plugins that use increased word length.
Hayez, from what I understand, avoiding distortion by adding dither to floating point math systems is very hard and maybe even impossible. With integer maths, you add shaped random noise at a fixed level so as to toggle the Least Significant Bit (LSB) of the words of your signal. The problem with floating point maths is that the LSB doesn't have to be at the same level for each sample.
If you let the level of the dithering noise follow the level of the LSB, the noise is correlated to the signal and thus isn't random any more. You effectively introduce harmonic distortion.
On the other hand if you add dithering noise at a fixed level, you can't know if you are toggeling the LSB or not and worse, you don't know where the noise will end up in your digital word. It can easily end up somewhere in the (very) audible range.
UnderTow |
|
|
|
Colin OOOD
Moderator
Started Topics :
95
Posts :
5380
Posted : May 19, 2005 01:31
|
|
Hayez
Started Topics :
8
Posts :
393
Posted : May 19, 2005 01:47
|
@UnderTow, I don't know how Waves are doing dithering on doubles, but it does sounds like they're doing it the right way.
I also don't think WAVELOGIX was talking about fixed point.
This is a quote from the REQ manual:
"Renaissance Equalizer is a dual-precision processor. This means all calculations are carried out to a
48bit value within the REQ (on fixed-point DSP chips); for native processing it is performed with
64bit floating point values. This preserves greater resolution of fine details, and definitely affects the
sound of the equalizer. By working at this level of precision, the most exacting details are maintained
throughout the processor. Only at the output does this dual-precision data path become reduced. This
reduction is performed by dithering the 48bit internal data to the desired output. (Native 64bit internal
processing is dithered to 32bit floating point output)."
"a new art came into my mind which only you can create, the Art of Noises, the logical consequence of your marvelous innovations." Russolo, 1913 |
|
|
|
UnderTow
Started Topics :
9
Posts :
1448
Posted : May 19, 2005 02:09
|
Hayez, Waves might first convert to fixed point, do the processing and dithering, then convert the output back to FP. Not sure really. (And I'm not sure what the trade-offs/gains are).
Errr actually, I hadn't finished reading your post when I wrote the above paragraph. That is what they are describing. Oh wait, no. Not exactly. Well I don't know exactly what they do. I'll run some tests to see what it does to the signal.
But by the way, you know the old addage, you get for what you pay for. In some cases what you pay for might be added harmonic distortion.  Check these graphs out:
http://home.casema.nl/ajohnston/limiting/
UnderTow
|
|
|
|
WAVELOGIX
Wavelogix
Started Topics :
136
Posts :
1214
Posted : May 19, 2005 11:39
|
so ,waht exactly is the bottomline here ? are we actually "fucking" up the signal just to make it sound bettter ? ........
btw , now i see the cons of digital audio compared to analog domain .....
respect .. chandan ! |
|
|
|
EYB
Noized
Started Topics :
111
Posts :
2849
Posted : May 19, 2005 11:51
|
I don't see any cons
If it sounds good just chain fx and see if it sounds better. If it sounds perfect without fx let it without, or try some stuff.
And most fx have A|B comparison so u can always test different settings and compare them. Just make it sound perfect.
Signature |
|
|
|
vox
Started Topics :
2
Posts :
114
Posted : May 19, 2005 13:30
|
Quote:
|
On 2005-05-17 21:46, WAVELOGIX wrote:
hello guys ,
well ... i have this question lingering in my mind since many days .. just thought id ask and get stuff sorted out
i have a synth line playing .. that i get from a typical synth like vanguard / vsti ... now ... these vsti have enough controls to tweak around , so that one can design an appropriate sounding patch and then filter it just right , so that the patch sits properly in the mix ...... then offcourse a little 1-2 band eq to make things a little more better .....
now i use the WAVES plugin V5 all the time , for all the processing .... my question over here is ... once i have a decent sounding patch comming off albino / vanguard ..... is more processing thw way to go or one can actually settle with less .....
well .... for example ... i could use a compressor - some tube amplification plugin - 4 ban REQ - CHorus - PHaser .. etc etc ...
now when i follow such an effects chain .... dosent my sound signal get degraded with every component that the sound passes through ? becasuse every plugin introduces some noise / artifacts into the signal .. due the processing rates difference ...
if the above is the case .... isint getting a proper sounding patch and then little eq the way to go ? to keep the sound as clean as possible ....
and if multiple effects do not degrade the sound , then i think one could actually use multiple plugin to get the synth to sound way different and better [ depending on how you use them !! ] .........
so just wanted to know ... how you guys go around using plugins .....
btw .. more plugins also mean more audio bounces ! [ another reason for me to use less plugins all the time ]
Thank You ,
respect .. chandan !
|
|
there is no 'noise' introduces in signal chain. if most todays application internal processing is 32bit or even 64bit, so quantisation noise is definitely not an issue here. if your plugins make the sound better, you should definitely use them as you use them already. and using waves delay, reverbs, compressors and eqs is most certainly better than using built in effects in your synths. and why does more plugins mean more audio bounces? what do you mean? i use cubase with tons of plugs, and it all works realtime.
http://myspace.com/voxproject |
|
|
|
