Pink noise is an audio test signal that contains all the frequencies in the audio spectrum with equal energy in each octave. Unlike white noise, which contains equal energy at all frequencies, pink noise contains less energy in the higher audio frequencies than in the lower ones. This is because the higher octaves have wider frequency content than the lower ones.
if your were to slice the audio spectrum into one hertz "slices", the octave between 100Hz and 200Hz would contain 100 such slices. The octave, between 1KHz and 2KHz would contain 1000 slices, while the octave between 10KHz and 20KHz would contain 10,000 slices. It can be seen that if each slice contained the same energy, there would be much more overall energy in the higher octaves than in the lower ones.
White noise can be converted to pink noise by passing it through a low pass filter with a -3db per octave roll off.
Let suppose that an orchestral sample library is balanced for all instruments on all octaves with the Pink noise test signal, does it imply that Eq would be obsolete ?
You cannot render audio engineering obsolete with technology. Microphones are silly dumb sticks with vibrating foil converting energy from the air into electronic signals.
No matter how pristine, how balanced, how good, how anything the capturing process is, the engineering processes will always be necessary, coupled with a musical ear, to turn this crude capturing of signal into something which resembles music.
EQ obsolete. God, I hope never. EQ is the king daddy of all engineering processes.
Not in the capturing process. Once you got all your samples on disk a pink noise (kind of) treatment would be applied to each articulation on each instrument. For example, at Velocity 35 the "A" 440 Hz would be attenuated by -3db and made the perceived loudness sound equal to the "A" 220 Hz. In fact it's all about the perceived loudness in ears.
So at Velocity 35 for any instrument at any frequency what you set is what you get.
Like a pink noise wav file having all frequencies stock in the file. When played the Eq is uniform all over the graph.
Well i'm not clear about how to explain the point.
I think you answered it just fine Bruce.... and ditto that!
The fundamental point, Serge, is that music can never be an exact science. Too many variables. The best way to get a great sounding mix is the old fashioned way ---> "ears".
Same as a chef wouldn't create fabulous meals measuring ingredients scientifically (and such). They "wing it", in a sense. Grab some spice and toss it in, taste, repeat. But only "after" they've gotten extensive training. They use the old fashioned "taste" method. Does it taste good? Cool! So I technically added too much butter.... so what, it tastes "right".
same with music....
As Well ----- same as you wouldn't brick wall limit an orchestral piece. What you're describing is essentially "brick walling" the frequency content. It's just not gonna sound very good. I don't think music is meant to be "perfect"... after all, we are all trying to recreate "imperfection" with our samples. Why? IMO because humans aren't perfect, and neither is music. It sounds best when not perfect... (see: quantizing )
Besides, I like knowing my mixes aren't gonna sound the same as the next guy.
Ok we're getting closer. The goal is not to set all instruments at Velocity 35 and have a flat shape. The question is if a chef tell to his student "Grab some spice ", how much is "some" for the student ? How much is ff for a cello or a trumpet ?
Cooking is an art but it is also rules like chess.
Here the rule side is having an audio reference valid for all instruments in all frequencies. Then we can say how much is ff for a cello and a trumpet no matter if we cover low or high range. With this rule scoring software developpers would find interesting path to explore.
The art side is once the chef grab the good spices for a song (pp flute and oboe, ppp violins) he can experiment other combinations in a fast way at any range while keeping relative proportions among instruments. And how about mixing and mastering steps ? Well you get what you set.
White noise is great for checking to see if your system components are flat. Put white noise into the start of a chain, look at the result at the end of the chain, and you can see where things dip and peak.
Pink noise is good for checking physical things near max power. Note that woofers and bass amps are big and powerful, while tweeters and treble amps are smaller and weaker. Pink noise has more power in the bass than on the high end too.
You can put pink noise into a speaker system and push it closer to the limit. Record the result. Apply the anti pink noise conversion. Now you can see where the peaks and dips are when your system is closer to the max. It also approximates the energy that we expect from real music better than white noise does.
Pink and white noise are technical test signals. Nothing more. They don't really help when it comes to art.
I just don't see it, because there is absolutely no way to account for context.
The "rule" already exists, simply in good ensemble skills...most often called the "pyramid of balance." Same basic theory, that a low frequency sound needs more source energy to project in a given situation than a higher frequency sound. This is one of the first things a person learns in pedagogy and conducting courses, how an ensemble is balanced...the basis not being so much pink noise as the natural level of partials in the overtone series. Both chord balance and tuning are related to this phenomenon. I'm inclined to think that any mechanism for automating this process would be musically wrong 99% of the time.
The problem with the pyramid of balance--actually not a problem, but simply the application--is that there are several independent pyramids existing in any given musical moment. And there are exceptions to the pyramid, as in when an instrument of a given tessitura is featured against tutti or concentrations within its tessitura. At that point, it becomes necessary to detach that voice from the pyramid, and to further deconstruct the ensemble into many smaller sub-ensembles, all maintaining their independent pyramid of balance in addition to balancing their combined singlular choir against the other choirs.
If you just apply a strict bottom-up balancing scheme to a library, you create many problems. Not the least of which is that you want maximum dynamic resolution in each instrument. Applying the pyramid of balance to a tuba and a piccolo, say, you would end up with almost invisible waveforms in the picc against virtually full-bandwidth waveforms in the tuba. But what happens when you want to simply put a piccolo part over some Sargeant Pepper type production? Suddenly, you have no power, and you're cranking down everything to match.
The only time a strict, ensemble-wide pyramid of balance would ever be workable would be on a tutti chorale, like the Bach chorales which are so often set for full ensembles and used as warm-up and balancing exercises. But the moment you start moving forward on the historical timeline, and working with more polyphonic and horizontal structures, there is almost no point in any given piece of music where instruments in proper musical balance would be anywhere near the proper "acoustical" balance of the pyramid/overtone series type.
This is the reason people need to actually learn to engineer, and not just take the "I'm a composer, and why should I do that" kind of approach. A person who is creating end product with sample libraries is not an orchestral composer. He is an electronic musician. To deny it, or to somehow wish that technology will make it not so, is to ask the technology to purposefully compromise itself from maximum flexibility and effectiveness.