I am growing up, sonically speaking, in this digital age and daw's have made it possible ($$$) for me to quit my day job (so to speak) and focus on my first love.
This paper verifies my thoughts about hard limits versus aranging and mixing to deliver dynamics. But a few questions I have are: How does the hard digital limit "damage" the audio?
1.) Is there audible evidence of damage?
2) Is this simply the digital version of analog hard limits? i.e., are those apparrent "clips" in the diagrams really clips or are they fast compression as you might find in digital limiters?
Are those mastering engineers just tricking our brains into thinking that these hit songs sound better louder; an auditory illusion?
Thanks for the reference, I'll be using it often when I'm trying to explain why my mixes aren't as loud as others'.
I didn't read it as I already read a very similar article, but I looked at the waves - it's hard to believe someone cripples a song that much! That was way more than a lot of compression and a lot of limiting (at least what I'd rate "a lot")
As there's no date, I guess this article was initiated by Prorec's article "Over the limit"... you'll find it on the main page at http://www.prorec.com it's a bit more detailed, I think!
If you follow the electrical signal as the wave pushes through your system, those "flat tops" are clipping of the signal.
As you view a non clipped wave, the electrical signal to your loudpeakers follow each positive 1/2 wave as a PNP [polarity through the amp to the speakers) and the bottom half is NPN from the amp to the speaker. The sharp corners of clipping, invite DC into the signal chain (not DC offset...that is something entirely of it's own) and this DC of the squared off wave causes excessive heat in the amplifier, excessive heat in the voice coils of your speakers (due to DC osscillation in the signal path) and "asks" a loudspeaker to "freeze" in mid motion. This does burn out tweeters if their is approachable current and causes for a raspy, nasty sound, even under lower volume conditions. The signal chain only can do what is fed into it and if clipped waves are fed into it, the amp and loudspeakers and the sound quality will suffer.
It is unprofessional, unmusical at best.
(It is also being a ^#$%ing idiot and not an engineer)
On some systems, the wave of overdriving that trash inverses itself to either a normal look but with hash or to a triangle (urrrhhhggg) wave!
In any case, it is not representative of the source recording...at all. Even when you put a pure square wave (intended) into the mix, it needs to not be altered to clip level beyond the source sound or original wave shape.
It needs to be proper to begin with.
Trash in, trash out. (don't you just loathe the simplistic term of that?)
I'd love to see how some of the cd's for which I love the mix, compares (like chris boti's latest). Is there a specific kind of software used to view this kind of detail or can I simply use an editing software like wavelab, peak, or soundforge to view waveforms.
I think I'm gonna take another look at the recommended reading section.
Looks like the amplitude zoom was used. It would be more meaningful if the dB scale was used instead of samples.
Also, when you drop to your .wav editor, you need to open the sample from a .wav file that is converted from the .cda first to the hard drive.
If you use winamp and your record mixer is above 50%, you are boosting gain in the record range and that is inaccuate of the true value. One computer I used had true unity gain with the record mixer set to 33%.
Take the CD and open it with http:// and save as .wav Make sure the normalize funtion is set to "off". Then reopen the .wav file in Cool edit pro and see what you have.
Anytime I am dropping a real time source into the wave editor on the fly, my record level is at less than 50% (or unity gain for win mixers)
The givaway is how short the fade was......
Not a good represntation of proper gain management on the posters part.
The poster knows what he's talking about. He uses EAC, *THE* tool for getting .wavs from CDs literally bit-for-bit, even by rereading erroneous parts over and over again.
There wasn't any fade - the music just ends with one last beat - the fading off is actually just the reverb of that last beat.
I've got the MP3 now, and it sounds like ass. Granted, MP3s aren't the best way to judge quality, but this one was encoded without all that auto-normalize/compress crap. The pumping is very very noticable, and the bass has gone down the drain.
The following was found as an explanation within RO:
"As you view a non clipped wave, the electrical signal to your loudpeakers follow each positive 1/2 wave as a PNP [polarity through the amp to the speakers) and the bottom half is NPN from the amp to the speaker. The sharp corners of clipping, invite DC into the signal chain (not DC offset...that is something entirely of it's own) and this DC of the squared off wave causes excessive heat in the amplifier, excessive heat in the voice coils of your speakers (due to DC osscillation in the signal path) and "asks" a loudspeaker to "freeze" in mid motion. This does burn out tweeters if their is approachable current and causes for a raspy, nasty sound, even under lower volume conditions."
Now, I don't mean to be insulting, but if anybody cares, this is fraught with misunderstandings. I especially got a kick out of the oxy-moron 'DC oscillations' Peace
Understanding basic Electronics? That might be a toughie since it HAS been a while since graduating with my EE DEGREE, and designing electronic circuits all my cognizant life. Let me think about that one for a while.
I guess it just seems a little counter-intuitive that a direct current (which is fixed and unchanging, by definition) would also be considered as oscillating (which is an alternating current, AC, by definition).
By the reasoning that "A strai[gh]t line at amplitude is DC", a square wave could be considered alternating DC. Oh, wait, that's AC; point taken?
Well, you certainly see things differently than an EE would. In our world, half a square wave is just that, not DC. There is a tool known as Fourier analysis that reveals components in signals. (time domain <-> frequency domain) If one were to transform a (symmetrical, inherently) square wave, it would be evident that there is no DC component. The same would be true for a symmetrically clipped sine wave, or any wave that had no DC component to start with. No additional DC component is generated by clipping, unless it's asymmetrical. I can see why one might come up with the reasoning you did, but sorry, I'm seeing the engineering side of it.
It can always be argued that taking any arbitrarily short period of time, a changing entity can appear unchanging. The whole story is that it is, however.
Clipping contains many DC components..especially if it is amplifier induced. Power supply ripple for one. inductance and capacitance of a cable, reluctance, reactance, resistance. Nothing about clipping is symmetrical.
In a textbook VS real world, their are differences. We have not even entertained the thermodynamic properties.
From the Source signal through all of the components to the transducer, clipping behaves differently in each.