Tools for Analyzing the Quality of Mastering

Sep 21, 2014 at 4:56 AM Post #16 of 209
I think that's just where our ears have to come into play. I can't say what's flat based on looking at the visual data alone, but when I switched from the CD to the Vinyl, back and forth, my ears kept thanking me. The graphic analysis can only correlate to subjective listening impressions, there is no objective flat within the data. 
 
Edit: No fair! I said the same thing as you before you edited it! lol 
 
Sep 21, 2014 at 6:20 AM Post #20 of 209
I wonder if we'll eventually start seeing the "Sausage" award for the song that most completely fills all the available bandwidth.
 
Edit: Perhaps we could give extra points for tapering the ends [Edit some more: "or constricting random places along the body"] so it looks more like a turd? 
 
Sep 22, 2014 at 8:48 PM Post #22 of 209
How can dynamics be negatively affected if nothing ever hits 0db? I thought... the problem is that peaks would go over 0db but of course it's not possible, so they stay at 0db, whereas the quieter parts of the track gets louder. So, the difference between the loud vs quiet parts of the track is reduced, so there is the reduced dynamic range.
 
Off topic, but seeing the wave forms reminds me of me going to Audcity a few hours ago, testing my mic. I wanted to see when I talk, whether the audio picked up ever clips.
 
Sep 22, 2014 at 9:56 PM Post #23 of 209
I'm working to put stuff together in the original post, and it's frustrating because it seems to be taking longer than I want it to, but here's something that I wrote so far, and an accompanying real-life example. Also, if you want more information on this I would definitely check out Bob Katz's discussion on the Loudness War linked in the op. He's a great speaker on this subject, and he goes into detail about dynamics in that presentation. 
 
From text that will be posted in op soon: A waveform, like a peak meter, is a measure of signal amplitude, but it changes over time and is designed to highlight changes and dynamics in a signal’s amplitude/loudness. Large changes in amplitude indicate a dynamic track, while the flat-line plateau we see on many new records indicates a loud, compressed mix. When a mix is always loud, it is really always quiet. Perception in dynamics and energetic shifts in music require us to perceive a loudness shift, and an increase in the signal. In the words of Bob Katz, “things need to get quiet before they can get loud, you need a decrescendo after every crescendo”. But if the source is always loud, always hovering near it’s peak, and there are no “quiet” moments, our ears adapt with their own kind of sound normalization, they literally tighten up their anvils and stirrups and shut off to the constantly high amplitude in the signal. The nuances of the track become hidden, it becomes hard for the brain to pick out details, so what do we do? Turn up the volume! Then our ears adapt again. What do we do next? Turn up the volume again! The dog keeps chasing its own tail until we eventually need to sit in a quiet room with a Tylenol.
 
So, with that in mind, this is an example of some rock'n'roll. The infamously compressed Red Hot Chili Peppers Californication and below it Eagle's Hotel California vinyl, and below that Eagles' Hotel California modern CD release. I didn't really think about it, but just realized Californication and Hotel California are extremely similar in their lyrical themes, both being about moral decrepitude in California. So I guess some things never change... (but loudness and dynamics sure do!)
 
 

 
When people were joking about sausage waveforms, well... it's true, it's not even a joke anymore, that's actually what certain songs look like :sigh: As you can see, Californication has no dynamics through a vast amount of it. There is no headroom left, there is no quietness, there is no change. Does it clip? It clips like a madman, all throughout the track, but that's a separate issue from the dynamics. With "music" like that, it's impossible to really enjoy it or sense changes of energy.
 
Now looking at the Eagle's stuff, you can see a loudness change from vinyl to CD, but the dynamics are still there in both versions! The Eagles CD release does in fact clip at one point in the song (when the bongos come in after the line "they stab it with their steely knives, but they just can't kill the beast"). But other than that, it actually looks pretty dang good. It makes good use of CD's dynamic range, and it sounds excellent to my ears. Probably the best if not close to the best you could probably hope for in a rock'n'roll CD master (it's an old CD though, and I personally admit to liking the quietness in the vinyl more but that is purely subjective at this point... both, for the most part, are quality masters).
 
Hope that helps. I will have more info in the op soon, including tools for measuring dynamic range in a track. This website, the mean time, has a list of albums with their dynamic range measured: http://dr.loudness-war.info/    
 
Sep 22, 2014 at 10:43 PM Post #24 of 209
  So, with that in mind, this is an example of some rock'n'roll. The infamously compressed Red Hot Chili Peppers Californication and below it Eagle's Hotel California vinyl, and below that Eagles' Hotel California modern CD release. I didn't really think about it, but just realized Californication and Hotel California are extremely similar in their lyrical themes, both being about moral decrepitude in California. So I guess some things never change... (but loudness and dynamics sure do!)

There's an unofficial "unmastered remastered" version of Californication floating around on the internet, or there once was, made from a leak of the unmastered audio. You might want to look for it (I won't try to find a link for legal reasons).
 
Much better dynamics. DR11 according to the engineer (I don't keep the plugin around anymore).
 

 
Sep 22, 2014 at 11:28 PM Post #26 of 209
All of this makes me sad.  I some how managed to try out an Oasis album without realizing they're credited for some of the worst brick wall (I like sausage waveform more) mastering of ALL TIME.
 
Picked some random song from their discography:

It really isn't listenable, I'm pretty sure I'd like their music if it wasn't mastered to such an extreme.  I read the whole "walking down busy streets with lots of ambient noise and earbuds jammed in your ears" but why does anyone master their stuff like this?  
 
I can't imagine going to a university for audio engineering with the intent of mastering music.  
 
Sep 22, 2014 at 11:38 PM Post #27 of 209
  All of this makes me sad.  I some how managed to try out an Oasis album without realizing they're credited for some of the worst brick wall (I like sausage waveform more) mastering of ALL TIME.
 
Picked some random song from their discography:

It really isn't listenable, I'm pretty sure I'd like their music if it wasn't mastered to such an extreme.  I read the whole "walking down busy streets with lots of ambient noise and earbuds jammed in your ears" but why does anyone master their stuff like this?  
 
I can't imagine going to a university for audio engineering with the intent of mastering music.  

Yeah, it's a shame about Oasis. I really like a few of their albums.
 
Another shame is Muse. Not for brickwalling but for generally poor mastering (yes the two can be exclusive).
 
From Supermassive Black Hole, one of my favorite Muse songs. There seems to be plenty of dynamics, but check out the red. Mm, clipping!
 

 
Sep 22, 2014 at 11:44 PM Post #28 of 209
  All of this makes me sad.  I some how managed to try out an Oasis album without realizing they're credited for some of the worst brick wall (I like sausage waveform more) mastering of ALL TIME.
 
Picked some random song from their discography:

It really isn't listenable, I'm pretty sure I'd like their music if it wasn't mastered to such an extreme.  I read the whole "walking down busy streets with lots of ambient noise and earbuds jammed in your ears" but why does anyone master their stuff like this?  
 
I can't imagine going to a university for audio engineering with the intent of mastering music.  

 
They are absolutely breaking the rules of their own craft, violating the most basic standards of audio broadcast and distribution. I don't think most sound engineers think it sounds good, but there might be a producer sitting behind him who only has the sales numbers and his own promotion in mind, and unfortunately that commercial marketing entity sitting there has a stronger voice than the artist, the engineer, or good taste itself. As Bob Katz said, "Loudness is a drug." The industry realized that, and sold it like one. The only thing that can change the tide at this point is the critical ear of the audience. 
 
Edit: BTW, it makes me sad too. 
 
Sep 23, 2014 at 5:00 AM Post #29 of 209
  Yeah, it's a shame about Oasis. I really like a few of their albums.
 
Another shame is Muse. Not for brickwalling but for generally poor mastering (yes the two can be exclusive).
 
From Supermassive Black Hole, one of my favorite Muse songs. There seems to be plenty of dynamics, but check out the red. Mm, clipping!
 

 
The red stuff in Audacity just means that the waveform reaches 0 dBFS. It doesn't mean that the waveform is actually clipped.
 
Sep 23, 2014 at 6:35 AM Post #30 of 209
  The red stuff in Audacity just means that the waveform reaches 0 dBFS. It doesn't mean that the waveform is actually clipped.

 
Although if it reaches 0 dBFS a lot, chances are that it is indeed clipped. It is statistically not very likely to have so many non-clipped peaks at exactly 0 dBFS (+32767 or -32768), especially when runs of multiple samples at that level can be found. Even with dynamic compression, that distribution of sample values is not natural.
 
Some tracks can have clipping without reaching 0 dBFS, as they may be compressed and clipped first, and then attenuated to match the loudness of other tracks on the same CD.
 
Also, dynamic compression with aggressive settings (such as very fast attack and decay times) that are used to squeeze out the last dB of loudness can cause audibly distorted sound without obviously visible clipping on the waveform.
 

Users who are viewing this thread

Back
Top