We're good.

So you are saying that people who "overcompress" do it

Could you link me to some examples?
I find overcompressed material tends to peak closer to 0, not -30dB, don't you?

 

You are a smart kid, but I don't know if you are in a hurry but are having a logical disconnect.... but you need to check your math here.

Inside of a fixed integration window (say 5ms) with a time of 60seconds where the average of both signals must equal 1.5 volts, which signal will have higher peaks a signal based on square waves or one based on triangles.

No real world just pure math. As the integration window approaches the limit of zero your math starts to works, but RMS as a value needs to live in the real world because in pure math or physics it's a shit unit that never has a place. It would be about as logical as making an arbitrary unit called dicks and trying to explain electron fields and states in relation to the dicks unit.

 

At equal RMS, triangles will have higher p-to-p V, not sure where you're heading with this. Not following your dick analogy, could we try something else?

 

Big non sequitur alert!!!
Many moons ago I went to a party, and when I walked in heard a banjo and saxophone trading licks and jamming out to a bluegrass tune. It was Dave Brubeck playing bluegrass sax like he had been playing it his entire life. And yea, it was very dynamically played. I only wish that I had a guitar with me if only to have joined together in the fray.

 

Almost everything does, either I'm open-minded or my music taste is shit (leaning toward the latter).

 

thanks for the info & sharing the vids my friend,
my 2 favorite tracks from the Reccess album are "stanger" & "f*** that",listening to the cd,mp3's or vids of his music makes me proud of what he brought to the electronic music scene & i would care less about what someone has to say about his versions (sonics/dynamics )as the guy on your second vid.
(this is not a skrillex thread but everyone tries to copy his 'dubstep/brostep' style and be a "mister me too").

 

Match you: Took me awhile to connect "Take 5" with 5/4. Still not sure if that's where the name came from.

 

It is. And once we get into PWM spikes with W -> 0, even jazzers won't beg for more :D

 

I'll be good. Assumed that what I was saying (compressed stuff tends to be louder) is pretty self-evident. I mean, does anyone really think YT's response Loudness War (Good god y'all! What is it good for? Absolutely NOTHING!) -- normalizing audio -- was to make all the overcompressed stuff LOUDER?

 

So when they are referencing RMS post line level output which means its not caped by Voltage Under they are indirectly saying apparent volume normalization. With respect to the integration window a more dynamic track can be louder and more responsive than a more limited track if both are capped to same apparent volume.

So if you measure average SPL over 5 minutes with an integration window of 500ms and a maximum average of 90SPL and use attenuation to match two 5 min tracks they more dynamic one will actually have elements that are louder and supposedly sound better than the one which is more limited. Because the average is what matters not the constant when you match volumes.

BTW I don't care about the philosophical differences between DR and no-DR. When we start talking about shit speakers, smartphone speakers and playback systems with 15watts... large dynamic range becomes a no-go as if you can't hear it at all doesn't matter how good it sounds. Which is why compressors and limiters were first invented, back in the day on AM radio the noise floor was just 12dB bellow the max amplitude of the signal, FM radio which was Hi-Fidelity about 20dB of usable signal and people thought it was the shit. Compression in broadcast radio was first and still foremost used to increase the range of transmitted signal... because that made it clearer for the people who were listening... and regulatory agencies tend to get mad when you breach your broadcast license and over-modulate. This is where the louder makes it better myth first comes up, and for radio it is 100% true. A clear signal that you can listen to in noisy environment such as a car that is much better that a signal that dips bellow the noise floor constantly, etc. . .

Regardless... TLDR... Things are a lot simpler than you are trying to make them, and if you actually understand the math, then I will let you figure it out yourself. You seem smart enough to separate loudness from SPL while still taking the real world consideration of playback systems into account and separate business decisions from best practices. I'm taking the old lady out to a nice Vietnamese restaurant where they don't even play music, because for me (personally) no music is better than the being blasted by over compressed tunes that are mostly distortion, that makes me want to find some pre-teens kids and tell them to get off my lawn... even though we are no where near the house.

Apparent volume has no limits outside of how physics treats pressure waves as long as you stay withing a range that people can actually hear and doesn't make them loose their bowels instantly you compare the two recordings on equal footing, the less compromised dynamic track will generally sound better - because all loudness comes at a price with a loss of fidelity and an introduction of distortion. Regular people might respond to louder is better - but when they are at the same AVERAGE SPL - the track with larger peaks actually will sound louder.

This is meant to be how loudness normalization works, though it still has room to grow, it is somewhat of a cart before the donkey procedure as well.... but it is what we have.

 

Google doesn't normalize everything yet.

Some stuff gets normalized to -14LUKS sometimes, but there is money involved and these kind of things take at least 1/2 a decade to get rolled out and worked out.

 

I love Yello's 'Flag' and used to hear it on vinyl, which I bought back then. Lately I heard the remaster on Spotify. It sounds like shit. Not just overcompressed, rather multiband compressed, mangling the whole tonal balance and then limited to kill people. Had to hunt down the first cd release on ebay to get that awesome album in decent (digital) quality. If I'd had it ordered new on amazon I would have got the crappy remaster.
So... gone forever.

 

Why would they do that? This thread is about [over]compressed program material, not power amps. I'm sure it's possible to build an amp that would play loud stuff quieter than quiet stuff -- human ingenuity knows no bounds, but that's neither here nor there. Overcompressed stuff is made for the majority -- the great unwashed who like it that way. No one's overcompressing chamber quartets, AFAIK.

[Over]compressed material tends to be LOUDER, why does this even need to be said?

No opinion re. what sounds better, but no matter the size of the integration window, compressed stuff (pegging 0dB -- that's how it tends to be done) will never be quieter, and always be louder overall. On this, both math & common sense concur.

The shittier the system, the more compression helps. If I made music, and my target demographic was kids who like loud & listened to laptop speakers, I'd compress & bandpass hard (as in mix *for* laptops speakers and not, say, for Tannoy reds), and forget about dynamic range/that elusive unverbalizable thing one allegedly gets from unity-coupled NOS Gold Lions @ around 15k. Because, as you said, real world

 

This is where you keep making the same mistake - there is no such thing as 0dB in the real world.

Some people are having a conversation about what happens after the power amp, if you choose to not acknowledge what happens after line level, you are technically correct, more compressed will always be louder at line level. Now if you could invent an input jack that I can plug into some orifice into my body that would allow me to bypass my ears to allow me to listen to line level without the aid of any kind of speaker, I can quite easily set you up for life.

Once again your math only works until you get to a power amp, in which case you are 100% correct. Alas, once that signal gets converted and you get to a power amp KFINZG 0dBFS doesn't mean anything anymore and neither does line level.

0dBFS is a modulation maximum and has nothing to do with real sound.

Once again and for the last time loudness has nothing do with SPL and people respond to SPL.

You can get more SPL from a loud track with less energy ( a shitty amp and shitty speakers) however once you move away from shitty amps and shitty speakers and can produce as much as SPL as a human can tolerate .... its already been said.

I am starting to loose my faith in your ability to coalesce semi abstract concepts together to understand what is at best a fundamentally simple concept. In truth your mention of 0dBFS after all the time I wasted on this thread trying to explain the difference between line level loudness and SPL has severely taxed my patience and my desire to give a shit. You might as well be talking about the moon being made of cheese. I don't know if you are taking the piss, or just so in need of attention or a win that you need to construct some sort logical fallacy in which you can construct a logical argument that it its root is solid but only so by ignoring every other condition that must follow course after that argument.

Cheers mate, no point in pissing into the wind.

 

Last edited: Dec 11, 2017

Thank you @mozee for excellent posts. You've saved me some time to try and explain it for those who still don't understand. The main problem is that people these days don't understand how analogue audio works any more. I mean real-world audio, which is analogue. Not to confuse it with analogue-digital synths debate, although there are some similarities.

Digital audio is not supposed to be used all the way up to 0dBFS as they use it today. It should be used at analogue levels to get the most of it. It's a perfect copy machine and that's its biggest advantage that we never get with analogue audio. I love digital audio!

Good thread with lots of great and knowledgeable posts!

 

Sure, of course there is. For CDs, it's 2^16, or 65,536; for 24-bit int, it's 2^24, or 16,777,215.
inb4 "but 32-bit float!" No. No such thing as a floating point DAC.
How is this not obvious?
Would it be worth mentioning that when your DAW's rendering channel meter hits 0dB while rendering 16-bit deep file, the number 65536 will appear in said file?
Would pointing out that while the meter itself may go past zero (into the red), the same 6553 will continue getting spat out of the encoder, 65536 being the highest 16-bit int number possible? Should I explain why 0dB is so darn meaningful & important, that 0dB means you hit the ceiling and can't go any higher -- will continue spitting out 65536 65536 6553665536, i.e. straight line, aka clipping, until you go below 0dB?

You tell me, 'coz I don't wanna gloss over anything, but also don't want to seem patronizing. Don't like being patronized myself & sure would hate to inadvertently do it to you. But if this 0dB thing still confuses & you continue to doubt it exists IRL outside Equestria? Don't be afraid to ask -- I'm here to help

​

 

Last edited: Dec 11, 2017