Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: A new discussion on over-compressed CD titles (Read 19501 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

A new discussion on over-compressed CD titles

This looks like one for the list of vinyl that beats the CD issue:
http://forums.stevehoffman.tv/threads/cali...-4#post-8180551
(note - it's only the new vinyl, not the old vinyl).
Is there a reason for not putting this new version out on CD or iTunes? Other than finally giving the world the evidence that Vlado Meller has devastating influence on rock music?
In my opinion there's never a reason why a good-sounding master, as found on any LP, shouldn't be put to CD or iTunes. (And I say that as one who likes and purchases vinyl as much or more than digital.)

EDIT: I don't mean any LP has a better sounding master than the CD. Poorly-worded perhaps.
Yes, fans with competent hearing feeling obliged to buy a new version would be insult enough, but that’s so many times worse when it costs $35 (to quote one figure) and is not properly (=100%) archivable.

And now for something completely different: Is Coldplay’s Mylo Xyloto covered in clipping, or was I imagining it? I have yet to examine any waveforms, but I seem to have noticed a distinct fizziness a couple of times while listening to it, which really doesn’t fit the style (as if clipping ever does, but you know what I mean).

A new discussion on over-compressed CD titles

Reply #1
And now for something completely different: Is Coldplay’s Mylo Xyloto covered in clipping, or was I imagining it? I have yet to examine any waveforms, but I seem to have noticed a distinct fizziness a couple of times while listening to it, which really doesn’t fit the style (as if clipping ever does, but you know what I mean).
If I remember correctly, "Parachutes" was pretty badly mastered, and I'm not so sure about "A Rush of Blood" either.
It's only audiophile if it's inconvenient.

A new discussion on over-compressed CD titles

Reply #2
And now for something completely different: Is Coldplay’s Mylo Xyloto covered in clipping, or was I imagining it? I have yet to examine any waveforms, but I seem to have noticed a distinct fizziness a couple of times while listening to it, which really doesn’t fit the style (as if clipping ever does, but you know what I mean).
If I remember correctly, "Parachutes" was pretty badly mastered [...]

Pleade define 'bad'?

I did a quick scan on my mp3 version of the album (Lame -V4, I have the CD at home if you're curious):

Replaygain = -8.64 dB (foobar v1.1.7 so AFAIK that is with the 'new' replaygain scanner)
Dynamic Range = DR8

Both values are not that bad for a modern master...

I personally enjoy the dynamics and sound colour on this master.

A new discussion on over-compressed CD titles

Reply #3
Pleade define 'bad'?
Excessive use of filters (reverb/chorus/etc.), DRC and clipping.

I did a quick scan on my mp3 version of the album (Lame -V4, I have the CD at home if you're curious):

Replaygain = -8.64 dB (foobar v1.1.7 so AFAIK that is with the 'new' replaygain scanner)
Dynamic Range = DR8

Both values are not that bad for a modern master...
I might have remembered incorrectly, do you have "A Rush of Blood" at hand?
It's only audiophile if it's inconvenient.

A new discussion on over-compressed CD titles

Reply #4
Pleade define 'bad'?
Excessive use of filters (reverb/chorus/etc.), DRC and clipping.

I did a quick scan on my mp3 version of the album (Lame -V4, I have the CD at home if you're curious):

Replaygain = -8.64 dB (foobar v1.1.7 so AFAIK that is with the 'new' replaygain scanner)
Dynamic Range = DR8

Both values are not that bad for a modern master...
I might have remembered incorrectly, do you have "A Rush of Blood" at hand?

Not at hand, but I could scan that album later.

I don't think reverb was applied in the mastering stage on Parachutes. To me, it sounds that reverb was only applied during mixing, since the lentghs of the reverberation tails seem to vary between instruments/vocals.

A new discussion on over-compressed CD titles

Reply #5
I don't listen to Coldplay, but my wife has a lot of their albums. I had in the past casually noticed that when she played X&Y on our main system is sounded absolutely dreadful.

So this thread prompted me to do a quick clipping analysis of all the albums she has. And on first inspection, Mylo Xyloto has the least amount of clipping. See the various attached images. Clipping here is defined as 4 or more consecutive samples at or above -0.5dB which have the same sample value within a variance of 0.2% (ie. comprise a "flat-top"). As you can see, there is a fairly shocking amount of clipping on all the Coldplay albums, and indeed Parachutes is the worst of the lot:
[attachment=7323:myloxyloto.jpg]
[attachment=7326:rushofblood.jpg]
[attachment=7325:parachutes.jpg]
[attachment=7327:vivalavida.jpg]
[attachment=7328:xandy.jpg]

However, the apparent less-damaged nature of Mylo Xyloto is partially hidden by the fact that it has a lot of clipping well below full scale. All of the above scans were looking for clipping at or above -0.5dB. If we rescan Mylo Xyloto looking for clipping at or above -1.5dB, we get a worse result:
[attachment=7324:myloxyloto_1.5.jpg]

My personal opinion is that except in extreme circumstances (eg. perhaps the cannon shots in the 1812 overture) there can be no excuse for any clipping whatsoever.

A new discussion on over-compressed CD titles

Reply #6
I don't listen to Coldplay, but my wife has a lot of their albums.
We each have our cross to bear.

How did you generate those graphs?


A sample distribution plot is also often very revealing - but seeing that there are continuous runs of 21 clipped samples is very valuable (and staggering) information.

Cheers,
David.

A new discussion on over-compressed CD titles

Reply #7
...except in extreme circumstances (eg. perhaps the cannon shots in the 1812 overture) there can be no excuse for any clipping whatsoever.

This should be elevated to Quote of the Week-status.

I've been using some of my spare time on the 'net to harass kindly ask some of the bands I follow on Facebook (who are posting updates about new albums) to not let the engineers/producers brickwall the recordings. They're listening, but I'm not too hopeful that they will care enough. I think sometimes it's the artists themselves who choose the clipped mixes.
The Loudness War is over. Now it's a hopeless occupation.

A new discussion on over-compressed CD titles

Reply #8
I think sometimes it's the artists themselves who choose the clipped mixes.
I think they could be less likely to if they listened with loudness matching enabled. (SoundCheck in iTunes, ReplayGain in foobar2k and various other places).

Without it, people usually prefer the louder one in an A/B test, even if its worse.

Whereas, of course, just given one of A or B, a consumer sets their volume control to whatever level they want to listen at, and any loudness advantage of the clipped mix is lost - they just get the clipping.

Cheers,
David.

A new discussion on over-compressed CD titles

Reply #9
I think sometimes it's the artists themselves who choose the clipped mixes.
I think they could be less likely to if they listened with loudness matching enabled. (SoundCheck in iTunes, ReplayGain in foobar2k and various other places).

Without it, people usually prefer the louder one in an A/B test, even if its worse.

Whereas, of course, just given one of A or B, a consumer sets their volume control to whatever level they want to listen at, and any loudness advantage of the clipped mix is lost - they just get the clipping.

Cheers,
David.

Well, that's the nice thing about Spotify; it also has some kind of replaygain built-in. This is why, for example, old U2 albums sound so much better and more dynamic on Spotify than the new ones. Hopefully, this will create awareness as well.

BTW, I would be curious to see these 'clipping analyzer' results for What's the Story Morning Glory... that is definitely the worst sounding album in my CD collection. No dynamics and clipping all over the place.

A new discussion on over-compressed CD titles

Reply #10
Well, that's the nice thing about Spotify; it also has some kind of replaygain built-in.
I read somewhere that it is just ReplayGain. I know I should know, but I don't. You get no automatic say or royalties in the use of free software and public domain ideas!

Quote
BTW, I would be curious to see these 'clipping analyzer' results for What's the Story Morning Glory... that is definitely the worst sounding album in my CD collection. No dynamics and clipping all over the place.
If you haven't seen it already, read and weep: http://www.oasis-recordinginfo.co.uk/?page_id=6

I only just saw that there's an SACD surround version, and mixing it down into stereo can supposedly give a nice result...
http://live4ever.proboards.com/thread/70515

Cheers,
David.

A new discussion on over-compressed CD titles

Reply #11
How did you generate those graphs?

It's an option in Wave Repair's clipping detection function. Some time ago I added some fairly sophisticated clip detection code, mainly with encouragement from Krabapple. You can get it to look for any number of consecutive samples at or above a defined level where the samples have the same value within a defined variance. All this is to cope with the fact that there are plenty of CDs out there with clipping below 0dB, and where the clipping doesn't manifest as ruler-flat chopped-off waveforms. My guess is that the mix gets clipped and then passed through subsequent stages (EQ, perhaps?) which turns perfect flat-tops into slightly ragged flat-tops.

On reflection, adding this to what is basically a vinyl restoration program seems a bit weird. My only defense for doing it is that a valued user asked for it, and it was interesting.

A new discussion on over-compressed CD titles

Reply #12
BTW, I would be curious to see these 'clipping analyzer' results for What's the Story Morning Glory... that is definitely the worst sounding album in my CD collection. No dynamics and clipping all over the place.

Sorry, I don't have this album so I can't do it. (I am as big a fan of Oasis as I am of Coldplay).

Maybe someone else who has the album and Wave Repair installed would care to run the analysis?

A new discussion on over-compressed CD titles

Reply #13
Well, that's the nice thing about Spotify; it also has some kind of replaygain built-in.
I read somewhere that it is just ReplayGain. I know I should know, but I don't. You get no automatic say or royalties in the use of free software and public domain ideas!

I've read that too somewhere.

Quote
If you haven't seen it already, read and weep: http://www.oasis-recordinginfo.co.uk/?page_id=6

Yeah, I've read an interview with the engineer before... doesn't seem like he really knew what he was doing, sadly.

So, here are some numbers:

Coldplay - A rush of blood to the head (EAC rip to WAV)
Album gain: -8.75 dB
Dynamic Range: DR7

Again, not very shocking for a modern pop record. However, my evaluation copy of Wave Repair finds 359 'clippings' in the first track. That is more than 1 per second, on average!

Oasis - What's the stoy, morning glory? (EAC rip to WAV)
Album gain: -12.45 dB (!)
Dynamic range: DR5 (!)

Now we're talking. These values are really bad and explain the squashed sound. Suprisingly, Wave Repair does not find clipping in the first track!! So, we probably are listening to analog clipping in hardware like EQs and limiters. cliveb, what do you think?

** Just to add: The least dynamic track on the Oasis' album is the closer, Champagne Supernova. Replaygain: -12.63 dB, Dynamic Range: DR3 (wow!). But also in this song Wave Repair finds no clippings **

A new discussion on over-compressed CD titles

Reply #14
Talking of clipping and distortion: Emeli Sandé - My Kind of Love - WHY?!

You can have very low DR numbers without clipping, and even without compression, if the music isn't very dynamic.

Cheers,
David.

A new discussion on over-compressed CD titles

Reply #15
You can have very low DR numbers without clipping, and even without compression, if the music isn't very dynamic.

Of course, but on the Oasis' album clipping is clearly audible. But apparently those are harmonics introduced by clipping in analog hardware, rather than brickwall/soft clipping in the digital domain. At least, that is my explanation why Wave Repair does not detect any clipping.

A new discussion on over-compressed CD titles

Reply #16
Talking of clipping and distortion: Emeli Sandé - My Kind of Love - WHY?!

What the... This sounds like the mastering engineer had everything set correctly and then accidentally hit the +6 dB gain knob on his desk... This must be an error?!

A new discussion on over-compressed CD titles

Reply #17
Coldplay - A rush of blood to the head (EAC rip to WAV)
Album gain: -8.75 dB
Dynamic Range: DR7

Again, not very shocking for a modern pop record. However, my evaluation copy of Wave Repair finds 359 'clippings' in the first track. That is more than 1 per second, on average!

Oasis - What's the stoy, morning glory? (EAC rip to WAV)
Album gain: -12.45 dB (!)
Dynamic range: DR5 (!)

Now we're talking. These values are really bad and explain the squashed sound. Suprisingly, Wave Repair does not find clipping in the first track!! So, we probably are listening to analog clipping in hardware like EQs and limiters. cliveb, what do you think?

** Just to add: The least dynamic track on the Oasis' album is the closer, Champagne Supernova. Replaygain: -12.63 dB, Dynamic Range: DR3 (wow!). But also in this song Wave Repair finds no clippings **


If you ran Wave Repair's clip detection "out of the box", then you'll be running the default settings, which looks for 4 or more consecutive samples at 0dB. Many overcompressed CDs have clipping below 0dB, and so those won't be found. (Incidentally, the version of "Rush of Blood" I have access to has 815 such clips on track 1 ("Politik"), so I'm guessing you have a different mastering).

I suggest you try setting different clip detection values. Invoke File > Options, go to the Operations tab, and as a starting point choose the "Overcompressed CDs" preset. Then re-run the clipping detection to see what you get on the Oasis track. (Using that setting when checking Politik on Rush of Blood gives me 3027 clips!). If this still doesn't find many clips, try setting the clipping level lower (eg. -2dB) and the flat-top variance higher (eg. 1.0). These settings are there specifically so you can find things like analogue clipping and/or clipping that has subsequently been normalised below 0dB - perhaps to defeat casual attempts to detect clipping.

After the clip detection is run, zoom in on some of the markers to take a close look at the waveform to get an idea of what the clipping is like. Note that some clipping looks like genuine flat tops at normal levels, but if you switch off the "Maximise Amplitude" setting (View menu) you can then use the scroll bars on the right hand side to zoom in the vertical scale to see the fine detail. Some clips are not perfect flat-tops, but have a slightly ragged appearance.

A new discussion on over-compressed CD titles

Reply #18
Just because a waveform visibly clips doesn't automatically mean it sounds bad...after all, this is HA, right?

Clipping, and its potential audibility, is always a matter of degree, whether in the "classic" sense of hitting the input to an A/D hard enough to produce overs, or in the more "modern" sense of utilizing a digital brickwall limiter.  I've heard plenty of albums that have visible clipping throughout, but sound great (Aimee Mann's "Lost in Space" is a favorite example of mine).  It's all a matter of the decisions made by the artist/producer/mixer/mastering engineer.  Just because something *can* be misused and abused, and frequently is, doesn't mean it *has* to be.

It's interesting that Oasis's "Morning Glory" album was mentioned - I absolutely love the sound of that album.  I think it's the perfect representation of *that* music being played by *that* band.  Would the same production and mixing have worked for, say, Radiohead?  Of course not, and vice versa.  The sound on "Morning Glory" strikes me as a great example of Olde Skool analog production methods and equipment being used to great effect.  It always makes me think of standing in a small rehearsal space with a good rock band going full tilt.  It's just this huge mass of sound - DR0! - and the visceral impact is literally breathtaking. 

Oh, and regarding heavily-clipped tracks - 3027 clips on a single track?  Try The Chemical Brothers' "Setting Sun" - scanning it with the "Overcompressed CD" preset gives 88,805 clips...and it sounds like it.
"Not sure what the question is, but the answer is probably no."

A new discussion on over-compressed CD titles

Reply #19
I rescanned 'Champagne Supernova' with the 'overcompressed CD' setting and found 6545 clippings. I will look to those in more detail later. Thanks for the tip, cliveb.

Just because a waveform visibly clips doesn't automatically mean it sounds bad...after all, this is HA, right?

Clipping, and its potential audibility, is always a matter of degree, whether in the "classic" sense of hitting the input to an A/D hard enough to produce overs, or in the more "modern" sense of utilizing a digital brickwall limiter.  I've heard plenty of albums that have visible clipping throughout, but sound great (Aimee Mann's "Lost in Space" is a favorite example of mine).  It's all a matter of the decisions made by the artist/producer/mixer/mastering engineer.  Just because something *can* be misused and abused, and frequently is, doesn't mean it *has* to be.

Agreed, and that's what good limiters try to achieve.

Quote
It's interesting that Oasis's "Morning Glory" album was mentioned - I absolutely love the sound of that album.  I think it's the perfect representation of *that* music being played by *that* band.  Would the same production and mixing have worked for, say, Radiohead?  Of course not, and vice versa.  The sound on "Morning Glory" strikes me as a great example of Olde Skool analog production methods and equipment being used to great effect.  It always makes me think of standing in a small rehearsal space with a good rock band going full tilt.  It's just this huge mass of sound - DR0! - and the visceral impact is literally breathtaking.

I can imagine people liking this sound, but I don't. As you can guess from my username I was 15 years old when this album was released. I had no clue about sound, mixing, mastering and how CDs work back then, but I remember that was the first time I heard that something was 'wrong' with the sound, even though I liked the songs. I can't help imaginging how it would sound with 2-3 dB's more dynamic range. It will probably be way more in your face, powerful and punchy than it is now.

A new discussion on over-compressed CD titles

Reply #20
Glad people here found the Owen Morris interview of interest! I was lucky to get in contact with him and he was very generous in providing info for my site on the recording, mixing and mastering of Oasis's early albums.

The Dynamic Range Database features several relevant entries on the subject of the mastering, which show the possibility of more dynamic masterings of the Oasis catalogue. Here's one example... the very rare Stop the Clocks instrumental promo CD: http://www.dr.loudness-war.info/details.php?id=9057 (The b-side The Masterplan is DR5 on the mastered mix but DR13 here. Some Might Say is DR4 mastered, DR12 unmastered).

If you search by Oasis and click to order the results by descending DR there are a number of other similar entries relating to their earlier recordings from 1994-6. Also some of the earlier singles are less heavily compressed than the album versions of the same tracks. The Supersonic and Some Might Say singles spring to mind (the latter was mastered at Abbey Road and has greater dynamics than the album version).

A new discussion on over-compressed CD titles

Reply #21
Glad people here found the Owen Morris interview of interest! I was lucky to get in contact with him and he was very generous in providing info for my site on the recording, mixing and mastering of Oasis's early albums.

The Dynamic Range Database features several relevant entries on the subject of the mastering, which show the possibility of more dynamic masterings of the Oasis catalogue. Here's one example... the very rare Stop the Clocks instrumental promo CD: http://www.dr.loudness-war.info/details.php?id=9057
Wow - Some Might Say on there is a revelation. I bet with that, and some rock band stems, you can make a decent mix? Though maybe the SACD stereo fold-down is a decent mix (still haven heard it).

Quote
If you search by Oasis and click to order the results by descending DR there are a number of other similar entries relating to their earlier recordings from 1994-6.
http://www.dr.loudness-war.info/index.php?...&order=desc Someone has had great fun uploading data to that list.

Quote
Also some of the earlier singles are less heavily compressed than the album versions of the same tracks. The Supersonic and Some Might Say singles spring to mind (the latter was mastered at Abbey Road and has greater dynamics than the album version).


As you quote on your site, Owen isn't really a fan though... "The single version was mastered at Abbey Road and sounded appalling. My mastered version of the song for the Morning Glory album is good." ...!

Cheers,
David.

A new discussion on over-compressed CD titles

Reply #22
Quote
If you search by Oasis and click to order the results by descending DR there are a number of other similar entries relating to their earlier recordings from 1994-6.
http://www.dr.loudness-war.info/index.php?...&order=desc Someone has had great fun uploading data to that list.

It's good that a site like this exists, although IMHO the quality ratings are too strict, I think. To me, some modern music styles sound fine with a DR7 value (which, according to this website is 'bad'). In fact, those styles will sound thin/weak/not in your face/etc with DR14+ values, which according to this site are "good" values.

See this table (source: http://dynamicrange.de/node/1)


A new discussion on over-compressed CD titles

Reply #23
Quote
It's good that a site like this exists, although IMHO the quality ratings are too strict, I think. To me, some modern music styles sound fine with a DR7 value (which, according to this website is 'bad'). In fact, those styles will sound thin/weak/not in your face/etc with DR14+ values, which according to this site are "good" values.


I find it REALLY difficult to tolerate anything at or below DR9.

A new discussion on over-compressed CD titles

Reply #24
If I find something not loud enough, I turn up the volume on my speakers. In extreme cases (such as Sade's classic 80s song Smooth Operator), I normalize the file to 0 dB without compression or clipping. I don't feel dynamic music is "weak", in fact the loud parts, especially drums, kick a lot more.

Extreme compression destroys the bass and the kick of the sound, as demonstrated in this video https://www.youtube.com/watch?v=3Gmex_4hreQ . I even experimented myself with an original 1980s song, and my own overcompressed "2010s style" variant. After Replaygaining the 2 files and ABXing them, the overcompressed clipped one (I destroyed it roughly to the same level as the later Red Hot Chilli Peppers albums) had signifactly less kick than the original one. Clipression does not add kick to music, it just distorts it.

Today I heard a song in the morning in our car's radio which was obviously clipping beyond belief - it was the same loudness as the other songs yet it had distortion noise all over it. It was so strong that even me being half asleep (before anyone asks, I was not the driver) listening to a song on the car stereo it still sounded really awful. FM radio has dynamic compression, but it is an utilitarian one that increases signal area reception, not optimized for extra loudness. Death Magnetic CD version will sound just as loud on FM radio as does Sade - Smooth Operator, only DM will sound awfully distorted while Sade will sound very good (I am not stating this as some 80s music snob, I actually love Death Magnetic, but the Guitar Hero version [which has nice dynamics to go with the awesome songs]).

Overcompressed, clipping music does not sound good as ingame music, it does not sound good on car stereos, it does not sound good on good hi-fi systems, it does not sound good on bad hi-fi systems, it sounds bad on good computer speakers and it sounds bad on horrible computer speakers. There is simply no excuse for the loudness war.