Are "louder" CD's, re-mastered or otherwise, really "better" (1 Viewer)

[Joel Fontenot]

Hello all.
I have an issue to ask about here for all you audiophiles out there. This is a long one, so if you’re interested please hang in there .
If you listen to a new CD by some group, or even a re-issued one (whether it’s listed as a “re-master” or not, or even a “greatest hits”) and you notice that it seems to play back significantly “louder” than your old one (or, in the case of a “greatest hits”, any song plays louder than your album version of the same song on CD), is that a good thing?
I ask this because of several things I’ve discovered making my own mixed CDs for the first time after years of doing this on tape. I make mine on a PC using a digital extraction program that works very well called “Exact Audio Copy”. It’s been recommended on this forum, and I concur.
I also call attention to the thread entitled Finally! Supertramp re-masters!

I found the A&M re-mastered CD to be more punchy with more going on than the MFSL CD. The MFSL version was more reserved and clustered. However, while the A&M CD was more dynamic, it was somewhat more "digital" as well. The A&M CD brought more of a digital edge the vocals, but it was certainly not the sort of edge that peels paint. As a result, I did not have to go to Home Depot after the listening session for painting supplies. Still, there was no denying to my ears that the MFSL CD was a tad warmer in the vocals.

What caused that “digital edge” on the vocals?
As I understand it (and someone please correct me if I stray way off base here), digital sound files on a CD are stored similarly to WAV files for a PC – it’s a straight PCM file with a set number of values that can be assigned at any given sample. The number happens to correspond to quite a wide dynamic range and is the reason PCM audio has greater dynamic range than, say, consumer audio cassette tapes (I’m simplifying, I know, but bear with me ).
In a Wave file editor on a PC, the top end is usually labeled as 100% or 0dB above and below a base line. Anything above that gets “clipped” – in other words, any sound that is recorded louder than 0dB is flat-lined as a sound wave at 0dB for however long the time above that level goes for – any other frequencies that may be there above that range does not get recorded with any other values.
Now I have 2 examples to share. I extracted two versions of the song Heartless by Heart. One from the 1998 Greatest Hits CD on Sony Records and the other from the original album version on the Capital CD Magazine from the original Mushroom Records masters (which Capital acquired after Mushroom went under).
The ’98 Greatest Hits version plays significantly louder than the album version. In fact, all the songs on that CD are much louder than any album version. It’s also much louder that the earlier Greatest Hits/Live Epic CD that used digital re-masters of the original recordings for the original double LP release in 1980 (as the liner notes stated on the LP, and later CD, – re-recorded using the Sony PCM 1600).
This Greatest Hits CD does seem to have more “punch”, but also has that “digital edge” that KeithH experienced in his listening of that Supertramp remastered CD. I, in fact, noticed quite a bit of digital distortion going on. Compared to the Magazine version which sounded much “warmer” and “cleaner”, with no distortion in the vocals (and remember that Ann Wilson has got quite a set of pipes in her throat). And this from a CD released back in 1985.
So, I did the extractions on the two versions and loaded then up in a wave editor to see what the physical difference is in the sound wave.
I’m linking to a webshots album page as that’s the only way I can post images, but, take a look for yourself:
Full wave file of the Album version of “Heartless”
Full wave file of the ’98 Greatest Hits version of “Heartless”
Closeup of the opening note of “Heartless”, Album version.
Closeup of the opening note of “Heartless”, ’98 Greatest Hits version.
As you can see, massive clipping occurs from the first beat, and every beat after is clipped in the Greatest Hits version. Frequencies that were carried in those louder ranges in the album version are gone in the Greatest Hits version. In some cases, up to 40 samples are topped out (or bottomed out as the case may be). That means that digital clipping is occurring for long enough time that it becomes noticeable to the hearing.
I believe that this is the “digital edge” we are hearing.
We are losing the true dynamic top end of a song when this is done. There is no reason that Sony had to do that to their Greatest Hits CD. All they had to do was find the song with the loudest passage, use that as the bases to normalize the other songs such that no one song gets clipped.
I’ve noticed this on other CD’s.
In Pat Benatar’s Synchronistic Wanderings compilation, every song is clipped to some degree and loses the “real” dynamic range in many songs in the process compared to the original CD songs which reveal a wider sonic range - even if you do have to turn it up a bit more, it makes the louder punches on the original CD’s come through that much more.
Duran Duran’s Greatest compilation is another offender compared to the earlier Decade compilation. The original dynamics of A View To A Kill take a serious beating in Greatest.
Paul McCartney and Wings Band on the Run. This one was brought up before when someone asked about the differences in the re-masters that were done - the Steve Hoffman DCC Gold Disc, and Capital’s own 25th Anniversary Re-master. Several posters thought the 25th Anniversary version was better, some saying because it was “louder”. I posted that I thought the DCC version was cleaner and less distorted in some of the louder passages – especially in the final crash at the end of Nineteen Hundred and Eighty Five. I later found out why. The 25th Anniversary Edition is indeed louder – to the point of clipping quite often. The DCC version also has a wee bit of clipping, but much less often and for much shorter period of samples. It makes you wonder what you are missing in the 25th Anniversary Edition. Those frequencies (sounds, tones, notes,) are gone.
I like the idea of re-masters. Many sound just fine, and many do much better than some initial CD releases that used poor masters to begin with. But I’m not happy about the idea of punching up the volume, topping out the range and calling it a “re-master” simply as a selling point. It seems that “Greatest Hits” and some soundtrack compilations are also big offenders. The soundtrack to “Buffy the Vampire Slayer” is another one, by the way.
Anyone else have any thoughts, musings, rebuttals, corrections to my rantings?
Thanks for getting this far,
Joel

 

[Ken_McAlinden]

Louder is neither good nor bad. It all depends on why it is louder. If it is louder because the dynamic range has been digitally compressed to the point of distortion so that the average overall level could be raised, that is bad.

If it is louder because the old master used dynamic range settings for a modern orchestral recording of the 1812 Overture for a pop song and basically wasted some of the alloted 16 bits per sample, there is nothing wrong with that.

Regards,

 

[Mike Broadman]

It seems to me that what is happening with most CD releases is analogous to edge enhancement for video: both sacrifice clarity and detail to provide the illusion of being brigher or louder.

And it sucks.

 

[Kevin C Brown]

Compression. Compression sucks.
I 1st noticed when this comparing lps to CDs years ago.
The irony is that CDs have the *capability* of greater dynamic range, s/n, lower noise floor, etc, than vinyl. But nobody in the pop/rock world takes advantage...
"It's 11. It's one louder."

 

[Kevin C Brown]

Double post...

 

[Jeff Ulmer]

All they had to do was find the song with the loudest passage, use that as the bases to normalize the other songs such that no one song gets clipped.

That is not how mastering works, that is simply what common folk have come to do with programs. Mastering is done on a song by song basis, and is done to make the album consistent in tone and levels.

Mastering involves many things, one of which is (at least these days) limiting, which is getting the tracks as loud as possible without clipping. Unfortunately, by compacting much of the musical content into the upper bits of the digital scale, cheaper D/A converters will not be able to tolerate this and will crap out. At best, it can be harsh, at worst it clips, and is dependent on how many consecutive full scale samples occur.

Older recordings weren't mastered as hot to CD because digital limiters which can brick wall at just under full scale weren't available back then. Today, record companies are demanding the hottest CDs, which leads to this problem.

 

[Joel Fontenot]

Older recordings weren't mastered as hot to CD because digital limiters which can brick wall at just under full scale weren't available back then. Today, record companies are demanding the hottest CDs, which leads to this problem.

That's what I'm finding out.

In the case of older CD's, as long as the initial release used a good master (that is, a proper studio master instead of the LP pre-equalized master as it sometimes happened in the past) - even if the overall volume was lower - it still sounded better and more "full" than a newer re-master that had that limiting applied and was boosted in volume.

Not that some of those earlier releases couldn't have been mastered a little louder, because I have quite a few older CD's that still have plenty of headroom in the sampling range.

I just don't like the idea of limiting the original recording.

At least I now know that I'm not going crazy.

Loud is fine as long as nothing is getting cut off and the original dynamics are properly preserved. As Kevin posted, the noise floor is so low on the lowly CD spec, that there's just no reason to crush it at the top. I won't mind a few wasted possible samples.

Joel

 

[John Watson]

This thread is going to be very interesting. Haven't read it all yet, but I'll go out on a limb now and say - many disks are made louder because BIG IS BETTER. Sofas, houses, office towers, cars and SUVs, TV sets - everything has to get bigger to impress that mindset.

I had a couple of the earlier Sade disks, thought they would fit nicely on the carousel with the Greatest Hits cd of a few years ago, even if there was some song duplication. Instead, the greatest hits sound level was so wildly discordant with the early albums (ie, twice as loud) that it was impossible to put the 3 disks on for a relaxing listening session.

Where does stupidity like this stop?

 

[Philip Hamm]

Three words:

Dynamic Range Compression

It sucks. I recently got Bruce Cockburn's retrospective, with two new songs and a collection of songs from the last 25 or so years. The two new songs are so compressed they're practically unlistenable, and so are some of the older songs. I hate it. It kills music.

 

[Steve Owen]

I wouldn't always blame it on compression. I've ripped CDs only to find that the the peak amplitude doesn't come anywhere near to the maximum 16-bit level. That seems incredibly stupid... it's like mastering the CD at 10 or 12 bits when 16 is available. Makes no sense.

So, I can perceive of instance where a remastered CD might sound "louder" because they're making a concious effort to use all of the available CD resolution where they might not have in the past.

-Steve

 

[Joel Fontenot]

I've ripped CDs only to find that the the peak amplitude doesn't come anywhere near to the maximum 16-bit level... it's like mastering the CD at 10 or 12 bits when 16 is available. Makes no sense.

Well, first of all, the 16-bit in the CD format has nothing to do with the amplitude of the sound-wave being stored. It has everything to do with resolution of the possible frequencies (or amplitude of the frequency) in a given sample. The 16-bit is the "word" size per track that the digital converter is able assign in each sample which happens at 44,100 times a second. 16-bits allows 65,536 possible values to be assigned for any given sound amplitude recorded in that sample, +32,768 above (0-dB at line level, I assume) and -32,768 below (also 0-dB) that middle line you see in any wave editor. Some wave editors convert that sample value on the side scales of the graph to a percentage of 32,768 above and below that middle line. This is your amplitude, and yes, I've come across some CD's where the loudest passage in the whole thing reaches only 68% of the possible range the CD can store. Could those CD's be mastered a bit louder? Sure, and that I wouldn't mind.
Did any of that make sense?
If not, it's getting late, and maybe some else can clarify or even correct me.
"Dynamic Compression" is exactly the term I was looking for - and, no, that I don't like.
Maybe someone could start a petition
I'm about ready to get one of Blondie's new re-mastered CD's - Parallel Lines. I have all their original CD releases and this is one of the quieter ones. I think Capital has taken over all of Chrysalis' releases and Capital has been doing a lot of "Dynamic Compression" and pegging out the range lately.
So, we'll see.
Joel

 

[Vince Maskeeper]

First of all, we have to be careful about generalizations. The use of terms like compression (or "Dynamic Rage Compression" which is actually a consumer dolby digital term, which I've never heard used in music preparation) and the association with "bad" is incorrect.

The vast majority of recordings- whether tracked in pro tools last week, or tracked at Motown four decades ago contain some forms of "compression". Often the "analog" sound people enjoyed from older recording comes from butchered dynamics caused by "tape saturation"- which is a form of compression. Tube pre amps are often the rage, and again provide a subtle compression simply from the slope and response- and of course are often driven to saturation specifically for the "warm" tube sound (which is often little more than midrange smearing from harmonic distrotion resulting in a compressed signal). Most of the "punchy" drum tracks that big prog rock fans droll over are 100% the product of creative use of raw compression.

Basically- compression is a tool, like any other tool- and it can be used in wonderful ways to help recreate what our ears would hear in a real environment.


The issue of post production limiting and digital stage sweetening is a separate issue from the tool of compression. While compression can be used as a tool in this stage as well- it isn;t mutually exclusive, nor the prime scalpel in the surgery. Often this finalizing type boost is taking place in the digital realm, and involves applying brickwall filters and goosing the level for all that it is worth.

The ironic part is, you will find the majority of mastering engineers don't like the concept any more than you do. However, they will claim their hand is forced by the artists or the producer, and the buck will be passed on up the line. The bottom line is that louder discs are seen as better- even in all knowing "audiophile" circles. "Remaster" versions of vintage recordings (or worse, multichannel remixes) are "ooooh'd" and "ahhh'd" over every day on this forum- and often the biggest difference is the application of dynamic controls to push the overall level.

But this is not at all "news". I would suggest to Mr. Fontenot that his hard found research is akin to discovering America in 1982. These processes and basic of digital effects on transient response are as old as digital audio technology itself. The ideas of how modern recording are created are not at all a secret- and the waveform example are almost literally page one of digital recording knowledge.

What you're seeing is not necessarily "clipping" in the traditional sense- as the signal is not authored (or attempted to be authored) in excess of the maximum available headroom. Instead filtering is applied to create a limited of the dynamic response. This does not really result in a "loss in freq" as you first post suggests- just a loss of the original dynamic difference between loud and soft. This, in turn, often pulls details which were obscured by the extreme in dynamics to the front a bit more- resulting in a "different" sound- I would never argue better or worse, as I am unaware of the original intent.

Anyway- I'm babbling. Point is that if you look at a physical rip of the majority of modern recordings, you find sharp limited resulting in albums which really only exploit the top 4-5db of the Dynamic Range of CD. This is a circle that feeds on itself, as engineers and producers compare their finished product to others- and they feel theirs sound whimpy. The push to limit their record to really hit harder, and the cycle repeats.

Ironically, the recent release of LOTR on DVD will probably push the DVD film audio into similar directions. BY reducing the dynamics on LOTR, Mi Casa created a mix which is essentially "louder" in average level-- and it's getting rave reviews in the "I've got a big subwoofer" world of HT audio.

I'm actually amazed daily at how little the average person, even "enthusiasts" understand about recording and the audio technology. I would suggest to anyone serious about music listening to understand the production element- and how the entire atmosphere of a recording facility is designed to suck the life out of music. Subscribe to a magazine like Mix or Recording, or heck- subscribe to TAPE OP (it's free).

-Vince

 

[John Watson]

I guess the old saying that some people's taste is all in their mouths might apply? In this case, they are overdosed on aural candy.
The Academy of Ancient Music tried to perform the work of the masters as they were played in Mozart's (or whoever's) time. But by putting these performances out as recordings, the Academy was defeating its purpose?

 

[Joel Fontenot]

Mr. Maskeeper,
I understand everything you’ve written, and it makes perfect sense to me, believe it or not.
Perhaps I should start searching around more on this issue.
Listening to music is, after all, just a fun hobby of mine that I’ve been fiddling with since the mid 70’s when I was recording music from the radio onto a portable tape recorder with the built-in microphone slapped up against one of the stereo speakers. Those were my first mixed tapes – at 10 years old.
In later years, friends of mine would wonder why I wouldn’t peg the VU meters deep in the red when recording albums on cassette tape decks the way they often did (I refused to buy pre-recorded cassettes).
I know that I’ve only just begun to understand what many others have known for years now that I’m just getting into the “common folk” home digital recording area. But, this whole “louder” issue is something that I’ve seen crop up a lot in various places lately – including this forum – without any real understanding of the consequences, if any at all.
I see quite often comments like: “I like this new release because it’s (fill in the number) dB louder than this other one”. Maybe the “other” one had the room to go louder, maybe not.
I only have a few MFSL gold CD’s, and one very early MFSL disc before they went gold (I, Robot, from The Alan Parsons Project). Those and my one Sony “MasterSound” gold CD (Pink Floyd’s Wish You Were Here) all are mastered with some headroom left in the possible amplitude range. The several DCC gold CD’s I have are a bit “hotter” in that the peaks do tend to top out quite often, but the mass of the sound is kept lower. And the “flat-lines”, for ignorance of a better term, at the top end aren’t for nearly as long as these newer releases.
I brought this up so that someone who might understand what is really going on in the recording industry could at least help clarify – whether it’s for the good, the bad or the ugly.
Sounds like it’s for the ugly – people pushing other people into doing something they don’t really want to do.
I would still rather a CD try to remain faithful to the original recording artist’s and engineer’s intentions – sonically, dynamically, whatever... as best a CD can. However, it is hard to know what that original intention would be on today’s new releases, because that “punched up” (if it was done) release would be the only standard available to go by.
And as you say, Vince, maybe that is what the artist wanted.

That is not how mastering works, that is simply what common folk have come to do with programs. Mastering is done on a song by song basis, and is done to make the album consistent in tone and levels.

I realize that, but I was referring to the “Greatest Hits” CD, which uses songs from several different albums. I would assume that the engineers would have to compile those songs in a somewhat similar fashion as those of us making our own compilations. Epic Records had no problem doing it with Heart’s 1980 Greatest Hits/Live album and later CD. That CD has every song sounding much more like the original album versions and still staying in line with each other, except for Straight On which used a slightly different mix.
Thanks for all the comments. It is a learning experience, and I’ve come here to learn.
Joel

 

[Steve Owen]

Joel, you said exactly what I meant... I over-generalized.

I see very little reason to NOT use all 65536 steps. Just for giggles, I've taken a look at some (what are supposed to be) audiophile CDs of "quiet" music like violin concertos and the like only to find that they're using only 20-25% of the available resolution (peak range somewhere around step 8200 or so... when normalized to the same output, that's the equivalent of about 13 bits or so). To me, that seems to be a complete waste of resolution.

What's the point in going after 24 bit resolution when a lot of CD mastering doesn't bother to use the 16 that they already have.

-Steve

 

[Dave Morton]

Vince,
Great explanation! I'm getting into the world of home recording and am trying to learn as much as I can. I must have learned something because I understood your comments. Also, what do you think of the magazine Electronic Musician?
As for as 24 bit is concerned, cd players cannot play back at this resolution. Thus, the mix must be dithered down to 16 bit to maintain the dynamic range. I believe most new recordings are all 24 bit if they are recorded digitally. Actually, you can maintain the 24 bit off of old analogue tapes too.
I think this is correct. But someone will correct if I'm wrong.

 

[Ken_McAlinden]

Besides just being ready for the future, another reason 24-bit masters with higher sampling frequencies are used is so that when engineers use their digital tools for various things, the artifacts are less severe when dithered back down to 16-bit.

Regards,

 

[Luke Pacholski]

Well, if something *peaks* at that level, yes, that's dumb. Your peak level should always be close to 100%. Anything lower is, as you say, a waste.
On the other hand, if the peak *is* around 100% but the "average" volume is around 20%, well, anything you do to "fix" that will result in reduced dynamic range.