hey, interessting!Kingston wrote:The reason I tend to get itchy about this particular subject is because,
1. I design and build (tube) preamps and understand summing busses and buffers very well.
2. I program and design various DSP processes which obviously include summing and clipping.
which are the products?
i'd really like to take a listen!
no need to be sorry.Kingston wrote:Hence when I see obvious erroneous information on these particular issues I jump on it. Sorry about that. I simply can not read all of your posting seriously because of the analog vs. digital assumptions you have.
however, what i've been talking about is no myth.
f.e. when feeding a signal into an analog console, the signal is running through a certain ammout of units, such as elko's, transistors, ect, i don't have to tell you.
do you really want to tell me, that all these unit's dont apply a certain ammout of nonlinear processes, when overdriven?
i.e. we have a harrison console in our studio, where i work often (_the_ console where queen were mixing their stuff btw).
when i feed a vocal signal into a channel, and i level it exactly to 0db peak (no eq or compression applied) and then compare it to the same signal, which goes straight to the speakers, it sounds nearly the same, also same loudness.
but when i push up the gain of that channel in the harrison console (which i btw could push up to the close to max without any _bad_ influences to the sound), then readjust the level on the masterfader, so that the signal again is on exact 0db peak, the vocal track sounds very different, and it is subjectively louder, and way calmer in peaking.
i analysed this signal in wavelab then, again comparing it to the source signal.
the analog-gained vocaltrack was greatly shaped in frequency and peaks.
do that in the digital domain.
just raise the fader in the channel, and readjust it on the masterfader.
what you have is the _exact_ same signal again (besides maybe some calculation errors of the mixerengine, but that's another story), no nonlinear processes are applied.
as _noone_ is able to _exactly_ adjust a signal so that the unit tolerances are not overdriven by a tiny ammount in the analog domain (as we allways try to feed the loudest signal to the analog channel, to the max noisefloor reduction), these nonlinear processes are applied to mostly every channel on a mixing console.
that _is_ definately audible.
and, more, most people like that sound, so they just do it because of even that reason only ...
i don't say that this cannot be done in digital, but one must concioulsly and wiseley apply that, otherwise it is just not happening.
what kind of music do you do?Kingston wrote:I just can't see a situation where a mastering compressor needs to be protected and "work harder" because of a few sharper transients here and there. And I can't see its effect on final mastered loudness and sound as bad.
please, again, reread my posts. i explained it in depth why this is so important.
there are signals that can influence the masterprocess greatly.
thing is, that you even might not be able to hear it.
f.e. take a signal where a great ammout of dc offset is applied (you must know, if you develop tube gainstages).
you don't hear it.
this signal can destroy your mix, as the mastering compressor reacts on the mathematical signal. you are just not able to get the mix as loud as you want.
same with transients. there are transient peaks sometimes that you are not aware of the ammount, as they might not be in the frequency range one can hear.
however, too many transients make your mix less loud, as the _body_ of every rich-transiented sound (the part after the attack) _must_ be quieter. otherwise you wouldn't have transients at all.
so, if the body of all signals is quieter, but the transients hit the peak nevertheless, the mastering compressor _has_ more audible, more fast work to do.
now, in this case, you have to crank up the mastering compressor to get the desired level.
but what you achieve whith doing so is, that the transients (which allready had the desired level)
are raised up, too, so the mastering compressor has to bring them down again in a greater ammout, as if you'd wisely would have shaped the transients to a healthy ammout in the _individual_ cannels.
this is automatically the case if you raise the gains on an analog mixing console, and that in a so pleasant way (it's not one unit that overdrives, but many in one channel, which sounds diffrent, too), that people just do it because of that fact.
i mean, you must know, you develop tube preamps, which allways apply nonlinear processes, thats what actually makes the sound ...
i cannot believe _you_ are denying this ...
Sounds like a 'bomb' spoken in an airport. 
We are definitely not serious! Stick to killer audio plugs, Aleksy!
).
) ...
