Originally Posted by ClawAndTalon
No, even 16 bit head room is 100dB. That can cover the spectrum of human hearing above the noise floor of a typical room or outside (60dB), and the height of human hearing to the threshold of pain and immediate damage with room to spare. 24 bit gives more but that’s not needed, but it’s useful in production and mastering for “elbow room.”
Any modern day DAC is perfectly capable. To reproduce a shuttle launch with accuracy is limited to power and speaker capability, not digital quantization.
I disagree -- a shuttle, at launch, is (say) 160db. If we wanted to accurately record and play back a shuttle launch and immediately there-after record a pin-drop in a quiet room on the same track, you couldn't play both back today in sequence, accurately, without touching your volume knob. I'm not saying this is a "problem" but it is likely true. If you didn't touch your volume, you'd hear hiss or some artifact of your gain stage during the pin-drop (i.e., it wouldn't be accurate) on even the best DACs or you'd hear the pin-drop way too loud (because you had to gain-up the shuttle launch to its "accurate" level).
Is this untrue?
Originally Posted by ClawAndTalon
“That said, I can't assume this is a format limit. I must assume any kind of music we would ever want to create would "fit" into a 16-bit container, right? Katz, I think, said something about the crest factor of music being up to around 20dB, so that would say that 16-bit should be ok, right? At about 96dB dynamic range that gives some (needed) room. So, why, oh, why do so many engineers release "textured" music without putting in a little more..."effort"...for dynamic range when it can't be a format limit...can't it?”
Huh? Lost me.
I'm basically trying to determine why, if this isn't a "format limit", an engineer that is capable of producing music with "texture" can't similarly add dynamics with a little more "effort" in the recording/mixing process. I suppose, though, what I'm failing to grasp is the psychological component. The engineer (probably) "could" if they wanted to, but the artists "don't want it" because they hear their tracks are "softer" and want to "compete" and this is the "problem" just as much as (possible) technical ineptitude on the parts of the engineers.
Personally, I'm still thinking the "acoustic advantage" kind of explains it all...the more you "stuff in" the less dynamic range you're "allowed" to have. Modern music "stuffs a lot in"...at least that's how it seems to me with all the "texture" I hear.
Originally Posted by m. zillch
^That some musical works such as Ravel's Bolero have such a tremendous dynamic range in their natural state that listening to it without any DR compression is extremely difficult, because in order to clearly hear the subtle details in the faint opening part above our typical ambient room noise [HVAC, computer, refrigerator, fans, street noise, etc.] we need to turn up our volume controls to a high point which may (possibly) even clip our amps when the loud ending comes along (or at least generate complaints from the neighbors). The reason I mentioned "typical consumer living rooms" earlier is because in a dead quiet room of say 10dbSPL, or less, this problem is somewhat alleviated, but only a handful of us are afforded such a dead quiet listening environment.
I tried to make this point, several times, in the thread. Thanks for bringing it to a stark reality!
In fact, a recording like this may even be more argument for "HDR-like music" -- your "black-level" (the noise floor of the room) determines the lowest sound you can hear and if there is information in the track that is lower than that then you need to crank it up to hear it. But, at that level, the track may be capable of peaking above either your personal (don't want sound to get louder than XdB or my wife wakes up and puts me to bed, lol) or system limits. Meta-data in the track could help fine-tune the track on-the-fly for your personal "black-level" and "limits"....not too far off, IMO, of what HDR tries to accomplish for video (where the black-level is set by the playback device and room and the limits are set by your eyes and the playback chain's peak brightness).
Originally Posted by Rex Anderson
Where we all get screwed is when we bounce from song to song by different artists jukebox style when streaming. Here is where the circle of confusion comes into play, spectral balance is all over the place. And, the level of each track many be different based on the mastering.
I would suppose this is true just as much bouncing between genre. However, I find that even when I bounce between many different genres, I can pretty much set my volume knob at a spot and, for the most part, listen all night to music from Slipknot to Halsey, to Of Mice & Men, to Korn, to DJ Snake and whatever else and not move it. I guess that's one "benefit" of "compressed" music with a pretty constant "average level". Whenever I feel the need to move the volume up, it's usually when I hit a song with "lots of dynamics" and then it's only because I still want it "loud", but now I have to "worry" when the "crescendo" comes and wakes my bloody wife up, lol.
So, could all these tracks sound "better" if they had more dynamics? Maybe, though, I'm still not sure how true it is, and they sound fairly dynamic to me, still, but, it's becoming ever more obvious that I must not be a "critical listener". Regardless, if they had more dynamics, I'd still pine for compression so I could set the volume and forget it.
That said, I'm still not quite sure if the "spectrally complex" argument I keep trying to make (and maybe it's nonsensical, and feel free to say so, lol) isn't part of the reason modern music "lacks dynamics". The more "spectra" I "stuff into the box" (the less "acoustic advantage" I have, assuming I'm interpreting that term correctly) the less dynamic things will measure...I'd say this is by necessity. I certainly know the EDM tracks I listen to have more going on than a singer-song-writer track!
Which is to say, I think I'm trying hard (and maybe failing) to separate the use of DRC (which is "bad") from the measurement of "poor" DR. Is it possible that DR "measures bad" even without ever using DRC on the track? Could it occur simply because there is so much "stuff in the box"? Can we make this "stuff" more dynamic without masking all the sounds?
says "no" more dynamic range wouldn't cause any "masking" of sounds. But, I'm suspect -- if I hear a gun-shot I'm less likely to hear a pin-drop after it. If I hear a loud snare-crack, the same can be said. Of course, certainly, I could be conflating two things, but if the artist wants you to hear all the "elements" of their track, having "loud bits" interspersed may make it harder. Not only that, but the "loud bits" would force their average levels down, requiring their typical audience to turn it up, and then the "loud bits" would be "too loud" in many cases (either causing the system to clip or the wife to awaken, lol). Worse, if those "loud bits" take you out of the tune and cause it to lose its "spectral balance".
Originally Posted by garygreyh
There's a lot going on in this thread. I've read through all of the posts a couple of times, trying to figure out how to post something that will contribute constructively to the discussion. The problem I'm having is that there are more aspects to the whole "loudness wars" thing than are practical to try and address without writing a novel here. That said, I'll try to "compress" (ha ha) my thoughts down to something reasonable to digest. Before I dive in, I'll preface by saying that I'm speaking primarily about pop music. The classical and traditional jazz recording world is still (thankfully) pretty responsible about how they treat the mixing / mastering process. Rex would be much better suited to address what's going on in the classical genre than I am.
I'm just happy you and others in the industry are contributing! Whatever you can post (hopefully) helps us all learn, and thanks for doing so! I'm going to clip out a few pieces of your post that I want to dive in more to.
Originally Posted by garygreyh
The audio example from the Matt Mayfield video is good example of how increasing "loudness" can wreak havoc on a mix. it's important to listen past the obvious increase in average level and dig a little deeper. Not only has the contrast between soft and loud been diminished, the timbre of the track is changing also. Listen closely to snare drum. In the uncompressed version, the snare has a nice clean "pop" to it, in the compressed version, the sound of the snares on the bottom of the drum have been pushed up in volume, causing a gritty, crunchy sensation. The short reverb tail on the snare drum has also been pushed up in level, changing the ambience associated with the drum sound. The attack of the kick drum has also come up in level versus the body of the kick drum sound. It now has a more artificial sound to it, which I semi jokingly refer to as the "click drum". This sound is popular in speed metal, because it allows the kick drum articulation to cut through the wall of chain saw like guitars inherent in that genre. The tone of the guitar has gone from clean and glassy, to edgy and harsh (particularly on the leading edges of the notes) and any sonic space in the original that allowed separation between the mix elements has been filled with a fog of ambient noise. The point of these observations is that excessive compression doesn't discriminate. You tell it to smash the dynamics of the recording and that's what it does.
Totally get this, in that track it seems like they just pulled up everything. I suppose that'd be the equivalent of what a mastering engineer would do to "ruin" the final mix. However, does all that go away when you're dealing with "stems" rather than a "final mix"? I mean, if you have a lot of "stems" and you can control every little bit of the (e.g., snare) because you've recorded it to four different tracks with close-mics all around it, and can completely "engineer" the way the snare sounds in the mix...well...then you may have to "compress" artistically to "fit" the snare in, but...you're not dealing with the whole track anymore.
Originally Posted by garygreyh
That said, some genres of music are far more forgiving to excessive compression than other and many engineers have figured out ways to make "loud" recordings that still sound good. Clearly hard rock, metal, some hip hop and EDM are all genres that are rooted in what I call "sonic physicality". There's a certain weight and energy to the music that begs you to crank it up. These genres by definition usually have much less dynamic range than classical or acoustic jazz. While orchestral performances can have extreme dynamics, most modern rock music does not. That's not necessarily a bad thing, as it can give a recording more power and more muscle. The down side is that when taken to extremes, the effect wears off after a while because (as was mentioned in the Matt Mayfield video), without soft, there can be no loud. While there is no true right or wrong when it comes to artistic endeavors, my fear is that the practice has become so ubiquitous, that lack of dynamic range is now considered a characteristic of some genres of popular music. In my estimation, that's not a good thing.
I kind of disagree with the characterization. I think "lack of dynamic range" is subjective. But, I did ask at the outset of this thread, "how much dynamic range do we need." Still haven't received a good answer.... But, just as an example, if we have:
Two guitars, all the drums, the vocals (and possibly any other layered background vocals or just overdubs), bass guitar, and every single drum in the kit, then we had a handful of electronically generated sounds...well, seems like we have a "spectrally complex" mix and how do we still make the snares pop when it is competing with all the other sounds. How do you make sure you hear everything, while still allowing the, e.g., snare, to "cut through" and be "dynamic"? If you start making the snare "dynamic" as you want, I'd guess you have to start lowering everything else and then we're back to "when the user turns it up, they may clip, or hear the snare too loud".
I just don't think one can "have their cake and eat it"...without, anyway, some form of on-the-fly, end-user-controlled (based maybe on a calibration like "HDR") "compressor" which is "programmed" with "meta-data" put in the track by the mixer. This would allow the artist to hear their mix at several different "levels" on-the-fly. Further, my guess is that some tracks, like the piece Zilch posted, will require compression so every bit can be heard at a "comfortable" level that won't overtax certain playback chains or ears of their listeners.
Quite possibly, lol.