| Author
|
master volume question
|
charles
Triptych
Started Topics :
9
Posts :
155
Posted : Feb 10, 2006 12:46
|
Quote:
|
On 2006-02-10 12:19, assaf_zo wrote:
thats the thing i use limiter on -0.2 always
offcourse i dont start with him but at the end of my track ill add him
i think a good limiter wont squash ur highs that over 0db but soften them
|
|
the thing is that even by limiting at -0,2 , converters, some processing on the signal chain (change rate convertion and mastering for exemple) and even digital processor themselves can overload internally without give you an over on the output meter.
Also when a D/A converter will read your signal, for exemple a cd player to play your music, this one may create peaks between the samples for the same reason detailed above.
So your -0,2 db signal will be read by another D/A converter with peak and then will be distorted.
So to avoid this, don't stay too close of the 0db because L3 meter won't be a proof to ensure you won't have intersample peak after on the signal chain.
  "And i saw the angel come down unto me"
www.soleadmusic.com |
|
|
|
assaf_zo
IsraTrance Junior Member
Started Topics :
46
Posts :
141
Posted : Feb 10, 2006 12:49
|
yup, charles we have much to learn from u
keep posting !!
 |
|
|
|
black_cat
Started Topics :
2
Posts :
101
Posted : Feb 10, 2006 12:50
|
Quote:
|
the thing is that even by limiting at -0,2 , converters, some processing on the signal chain (change rate convertion and mastering for exemple) and even digital processor themselves can overload internally without give you an over on the output meter.
Also when a D/A converter will read your signal, for exemple a cd player to play your music, this one may create peaks between the samples for the same reason detailed above.
So your -0,2 db signal will be read by another D/A converter with peak and then will be distorted.
|
|
is some way to be certain using software that signal will be not distorted/clipping ?
for instance global analysis in wavelab with peak detection ?? |
|
|
|
charles
Triptych
Started Topics :
9
Posts :
155
Posted : Feb 10, 2006 12:59
|
special meter used for professional mastering application can solve this.
or best is just to don't care about this, to not use L3 or any limiter, to trust the mastering guy, and to leave him a lower signal (-3db) for him to deal with those possible peaks.
and don't worry, the mastering guy will send you the volume back even if your mix is -3db 
"And i saw the angel come down unto me"
www.soleadmusic.com |
|
|
|
index
IsraTrance Junior Member
Started Topics :
36
Posts :
548
Posted : Feb 10, 2006 13:23
|
Quote:
|
On 2006-02-10 12:32, assaf_zo wrote:
btw do u compress each channel?
|
|
i never compress,i only find it usefull for fx.
BTW if my master clips with the Compressor it will suck more
HTTP://www.decadancerecords.it/audioplug |
|
|
|
UnderTow
Started Topics :
9
Posts :
1448
Posted : Feb 10, 2006 13:50
|
index, black_cat, assaf_zo, go and read the whole thread before asking questions that have allready been answered. You are just adding junk to a topic that was nice and informative. Lazy gits!
Oh and a limiter always squashes. That is what it does.
orange-atropa, -10 dB peaks in 24 or 32 bit is fine.
-10 dB on the master fader in 32 or 64 bit floating point is also fine if there are no plugins in the master bus (and is even ok with some plugins).
As for being sure a plugin or an application doesn't clip, the easiest way is to send a > 0 dB signal to its input, bounce the audio and zoom in to the waveform to see if it is clipped.
UnderTow |
|
|
|
index
IsraTrance Junior Member
Started Topics :
36
Posts :
548
Posted : Feb 10, 2006 13:55
|
Quote:
|
On 2006-02-10 13:50, UnderTow wrote:
index, black_cat, assaf_zo, go and read the whole thread before asking questions that have allready been answered.
|
|
If its already answered so what do u talk anymore?
No bad feelings
HTTP://www.decadancerecords.it/audioplug |
|
|
|
assaf_zo
IsraTrance Junior Member
Started Topics :
46
Posts :
141
Posted : Feb 10, 2006 13:57
|
but surly he can complain huh!
so yeah we lame so what hehehe
no bad feelings |
|
|
|
vox
Started Topics :
2
Posts :
114
Posted : Feb 10, 2006 14:36
|
Quote:
|
On 2006-02-10 12:46, charles wrote:
the thing is that even by limiting at -0,2 , converters, some processing on the signal chain (change rate convertion and mastering for exemple) and even digital processor themselves can overload internally without give you an over on the output meter.
|
|
this is true for outboard mixing. but you can overload channels if you are doing it all within a single application like cubase since most of them support 32bit internal mixing which prevents clipping from happening.
Quote:
|
Also when a D/A converter will read your signal, for exemple a cd player to play your music, this one may create peaks between the samples for the same reason detailed above.
So your -0,2 db signal will be read by another D/A converter with peak and then will be distorted.
|
|
how can this be possible if a converter is working properly? the converter only converts digital stream to analog signal, there is no resampling or any digital processing involved which could result in change of digital data, which is happening if the signal clips.
if this is true, then the limit for your signal clipping is different for any d/a converter.
http://myspace.com/voxproject |
|
|
|
UnderTow
Started Topics :
9
Posts :
1448
Posted : Feb 10, 2006 17:12
|
Quote:
|
On 2006-02-10 14:36, vox wrote:
... you can overload channels if you are doing it all within a single application like cubase since most of them support 32bit internal mixing which prevents clipping from happening.
|
|
You are right in theory but many plugins don't have enough headroom to handle the high levels. There are also still many plugins that don't support 32 (or 64) bit floating point. The host application speaks to them in 24 bit. Any thing above 0 dB will clip in these plugins.
Quote:
|
Quote:
|
Also when a D/A converter will read your signal, for exemple a cd player to play your music, this one may create peaks between the samples for the same reason detailed above.
So your -0,2 db signal will be read by another D/A converter with peak and then will be distorted.
|
|
how can this be possible if a converter is working properly? the converter only converts digital stream to analog signal,
|
|
You can have digital values that are within the limits but that cause peaks beyond 0 dB FS during the reconstruction phase of the D/A converter.
I think the worst case scenario is a signal with the following consecutive sample values: 001100110011 where 0 is the minimum word value (24 zeros in a 24 bit system) and 1 the maximum value (24 ones).
Imagine drawing the curves between these points. You get a sinewave with huge overshoots between two consecutive 0s and between two consecutive 1s.
That having been said, not all DA converters are made equal. The good ones have more headroom in the analogue stage of the DAC. The really bad ones have zero headroom and will clip with any modern music you feed them.
Quote:
|
there is no resampling or any digital processing involved which could result in change of digital data, which is happening if the signal clips.
|
|
Yes there is. Typicaly in a DA there is upsampling then a digital brickwall filter then an analogue brickwall filter. This isn't processing as we know it when we mix and make music. It is essential to how digital audio sampling and reconstruction works.
Quote:
|
if this is true, then the limit for your signal clipping is different for any d/a converter.
|
|
And indeed it is. ADCs are all different and will react differently to the (hyper)compressed and (hyper)limited signals of modern music.
UnderTow
|
|
|
|
charles
Triptych
Started Topics :
9
Posts :
155
Posted : Feb 10, 2006 17:44
|
Under Tow is 100% correct and it's very well explained
"And i saw the angel come down unto me"
www.soleadmusic.com |
|
|
|
