What does 'compressed' sound like?

Compression is when quiet sounds are boosted to the same level as loud sounds, and loud sounds are boosted to the maximum level of volume that a CD can hold. There is a limit to how loud you can master a CD (100db? I forget the exact number) because with 16-bits of resolution, that's the most the format can offer. Because there are digital tools that allow record companies to jack up the sound of CDs, they compel mastering engineers to maximize the volume. They do this because they know that people react more to "louder" sounds, so in today's world of ipods and the "skip" button, and CD changers, people are more inclined to stop on the loudest track and listen, it catches the ear. Same goes for radio, which has even less dynamic range than the CD to begin with, they like to pump up the volume with compression to make a song "pop" out on the radio compared to other tracks.

Compression is the enemy of dynamics. Low-db sounds should be allowed to sound like low db sounds, a finger snap should not be as loud as bass drum thwak, but sadly that's what we get these days.

With vinyl, the limitations of the format prevent that kind of artificial jacking up of the volume, so they are left uncompressed with full dynamic range info intact.

That said, not all compression is evil. Rock/pop would be unlistenable without it, it has been used since the dawn of recording. The problem strated around 10 years ago, when the so-called "loudnes war" began where record companies kept escalating the amount of compression they were adding to CDs to make them "more competitive" in the market. It's now gotten to the point where the typical modern Cd has only *a few db* of dynamic range, and all the songs are slammed into the roof in digital clip most of the time.

And yes, that Flaming Lips CD is heavily compressed.

Yeah, what markl says. Flaming Lips CDs since Soft Bulletin have been very compressed, but that's the sound Dave Fridmann and the band were after, I think. The music never gets quiet. Bass is punchy. Kind of in your face. Barely have to turn up the volume compared to some of your old (80s early 90s era that haven't been remastered) CDs. There is often a significant amount of clipping on the peaks but you wouldn't likely hear that as such unless it was very severe. Another problem is that the peaks become distorted on many CD players because of a phenomenon known as "inter-sample peaking" and how digital filters handle peaks above max level, but that's another story. I only mention it because the opportunity for distorted peaks get much worse when the music spends so much time near maximum level (0dB) instead of if it was mastered properly with the peak level set about 2dB down from max.

Don't agree with his vinyl comment though. The sound could be just as compressed on vinyl, but there just isn't as much compulsion to do it. It's usually mastered by someone else and often just a very limited run these days, sometimes by companies more in tune to the wants of the audiophile community.

Markl.

Dynamics on recorded media are measured in the opposite direction. The maximum level that can be achieved is given a value of 0db, with the noise floor persisting somewhere between -60db or -90db, with silence being given an infinite value.

When sounds are compressed the quiet sounds are lifted in volume by a varying degree, but not made equal in volume to the loudest sounds. Having experimented with this, I can vouch that no commercial recording I have ever heard has compression that severe!

Compression has not been used since the dawn of recording. It was first used, I think, in the 1950's on classical recordings so that the quietest parts would be raised above the record's surface noise.

Davey is half right. Compression is actually more evident on vinyl, as the masters that were created had to conform to the limitations of the mechanical nature of the groove.

Who remembers the fuss caused by the uncompressed Led Zep album that made the needle jump out of the groove? To play the album you had to weight your stylus with a penny.

Half right? Which half?

Yeah, vinyl has gone through the compression wars too, but what we're talking about now is something of a totally different magnitude. CD could use much less compression than vinyl since it theoretically has a higher useable dynamic range than vinyl, and doesn't have the mechanical limits to contend with, that's certainly true, but in reality the opposite is usually true in the modern era, at least with popular type recordings which are about the only ones available on vinyl anymore, except for some of the niche audiophile labels.

Heh, should have been more specific, I was referring to dawn of rock music.

I don't pay as much attention to the vinyl threads on Hoffman's site (I'm digital), but I've seen him comment several times in threads on CD compression that mastering to vinyl does not allow the kind of compression we see on modern CDs (the square-wave buzz cut effect). Again, not an expert on vinyl, but I do remember him saying that several times.

trains are bad,
I recommend hanging out at Steve Hoffman's forums. Steve is one of the most highly-regarded mastering engineers around. He does not use compression or noise reduction for the CDs he masters.

Here is a link to an interview with him that explains what good and bad mastering is: http://www.netassoc.net/dougspage/Interviews.htm

Here is a link to his forums, there is an incredible knowledge base there where you can find out which version of which album sounds best on CD. They also often post waveforms of originals vs. remasters so you can visually see how bad some remasters are clipping:

http://www.stevehoffman.tv/forums/

You bring up two very relevant (and alarming) points:

1. Many younger people have never heard an uncompressed song or CD. There's a generation out there that has never heard a full-range recording. For them, it's not the compressed CD that sounds "bad" or "wrong" or "different", but the uncompressed one that is the odball.

2. Given that 95% of of modern CD releases of new music are heavily compressed, what happens after the loudness race ends and people get fed up with the crappy sound of their music (assuming that such a fine sunny day would ever occur, yeah right )? Will future Steve Hoffmans, whose stock in trade is remaining true to the original release and true to what we can only surmise was the artist's original intention, have to squash and slam every single future remaster of the music of the late 90's-00s to match these now-current CD abominations? Or will it be "right" to uncompress it, to let the music breathe, even though that's not what the original sounded like? I know what I would rather hear, but it's still an interesting dilemma. I don't *think* the Lips want their CD to be lacking in dynamic range, most artists are not audiophiles, and have been playing insanely loud music for decades so their hearing is shot, and are not always the best judges of how a CD should sound. Plus there is pressure from the record companies to make them louder and louder so they will be played on the radio. They want to make a living, and probably have almost no real formulated opinion on what good or bad mastering sounds like (just like the majority of the population) so they allow or just aren't aware of the over-use of compression on their recordings.

There's a third factor here as well-- Pro Tools. Often you have the music slammed and compressed long before it reaches the mastering engineer. It is highly likely that many of the master versions of most modern albums can never ever be "fixed" or restored to full-range, because they were never recorded or mixed that way to begin with. There is simply no dynamic range to restore...

Here is some further reading:

http://www.audioholics.com/techtips/...ngtrendsP1.php

http://www.digido.com/portal/pmodule...er_page_id=93/

Using google I couldn't find the better web sites on the subject but those above are a good start maybe others can chime in with other links.

The explanations of compression on here are a bit oversimplified.

In compression, a threshhold point is set at some level, such as -10 dBFS. Whenever the wave is quieter than -10, nothing happens. Whenever the wave hits or exceeds that point, every increase above that point is reduced proportionally by some specified amount. For example, you could do 2:1 compression at -10, which means that a -11 dB signal would stay at -11, but a -2 dB signal (8 dB over the threshhold) would be reduced to -6. (4 dB over the threshhold). The result would be that the loudest possible sound (0 dBFS) would be reduced to -5 dBFS, meaning the whole signal could then be raised by 5 dBFS without ever clipping. This is called "makeup gain."

"Limiting" is basically compression that's set to a very high ratio, like 5:1 or more. Then there's limiting done at such a high ratio that it practically puts a volume ceiling right at the threshhold point, and this is sometimes called "brickwall limiting."

Compression is not bad. You'll find compression on just about every pop song ever recorded ("pop" meaning anything that's not classical) and no studio is complete without one, or many compressors. The bad thing is when compression gets too aggressive (high ratios, low thresshholds, lots of makeup gain) and the dynamics of the recording become so crunched that they start to sound awkward and unnatural. Another bad thing which some sh*tty engineers do is set their levels so high that the waveform on the CD hits or blows past its maximum level (0 dBFS) which causes clipping. Digital is not like analog, where you can oversaturate the tape and get a musical-sounding distortion out of it. When you "oversaturate" a digital waveform, the tips of all the waves get "shaved" off and it doesn't sound musical in the slightest.

Overaggressive compression does not cause clipping, nor the other way around. Both are symptoms of the effort to release the loudest, hottest-mixed record. Sometimes this is an artistic decision, and sometimes this is done by the mastering engineer.

You're right about the fact that sometimes a signal is crunched before it even gets to the mastering process, but this has always happened and it has absolutely nothing to do with the supposed evils of modern studio techniques. Pro Tools can record at 24-bit and does not introduce any compression you don't deliberately throw at it, and adding compression in Pro Tools is not necessarily any easier than sending an analog signal through your console's compression circuit.

Of course it sounds that way because it is supposed to. The question that needs to be asked is, why is it "supposed" to? If the answer is simply to make it not sound like "the CD that is too soft compared to other CDs", then the overall sound quality may be suffering (potentially a lot) because of it.

You do still hear quiet sounds in the background, when that track of the mix is supposed to be quiet. The explanation given was rather poor. It's more like, everything just sounds overly controlled. You don't have sudden impacts where the musician popped the bass drum a lot harder at this moment or stepped closer to the mic and his voice got really loud... it's only as loud as the guy at the mixing board says it is. You need some control like that, or you might end up with CDs where everything is soft and the occasional noise dominates the mix (try whispering into a cheap mic and recording it for an example), but it can be taken to the opposite extreme where everything just loses its edge and kind of sounds bland.

Run - do not walk - to a good record store, and score some early Depeche Mode CDs to get a good idea of how good light compression sounds. Not the remasters! I have Black Celebration and Music For The Masses, and both are phenomenally dynamic and punchy. (BC replaygains to -1.03dB and MFTM to -1.33db.) Apparantly the remastered CDs (starting with Singles and continuing to the album releases) are far more compressed and are a good comparison point to how "modern" mastering sounds.

Compressed sounds like the Killer's Hot Fuss, or any recent RHCP offerings.

IMO compression sounds pretty bad on any hi-end setup that is capable of reproducing full dynamic range. But when played back on a setup incapable of high dynamic range or stereo seperation, etc (i.e. FM radio), it actually may sound better. You can think of it the same way as watching movies. If you have a theater screen that is capable of playing back widescreen anamorphic blah blah to its full capacity, you will probably prefer those widescreen movies over watching a 'fullscreen' version that is cut off or pan and scann'd.

True compression or saturation can be 'artistic', but the reality is the main motivation is to move all the 'details' that wouldn't show anyways on lowest common denominator stereos, and shoving them all into a range where they are audible by even cruddy speakers playing back FM. On a hi-end setup, it becomes more glaringly obvious that there is just a lot of non-utilized dynamic range, etc. Also IMO this colors the way people voice their gear. If the majority of their music is highly compressed, they may disregard a headphone fully capable of reproducing full dynamic range in a balanced manner as being 'boring' and prefer a more unbalanced headphone with more peaks and dips in its response, because they color the end result in making what is normally a flat dull compressed recording more 'dynamic' and in essence acts as a dynamic expander in a way.

Factor, I agree, but I did point this out earlier if you read my posts in full. Compression in and of itself is not automatically EVIL. It's the over-use that's the problem.

I often find myself at odds with some of the more hard-core Hoffmanite purists who insist all compression is bad. I seem to have a slightly higher tolerance for its use in modern masterings than some there. IMO, compression is just one of many factors that determine how good a particular CD or remaster sounds relative to the original. EQ, no-noise, resolution, etc. also plays a role.

FWIW, I'm not on record as saying that all older versions of CDs sound better as some Hoffmanites *seem* to. IMO, there are current remasters that are *significantly* improved over older versions, despite the *judicious* use of some compression.

The "problem" with many older recordings on CD, and yes I'll say it, is that many are not loud enough. Many actually left some resolution on the table by mastering too low, way way below the threshold of digital clipping.

Also, concurrent with the loudness race, DA converters underwent a major shift and improvement. DA converters of the late 90s to today are vastly improved over what we had for many original issues in the 80s, leading to the *potential* for vastly improved sound. Also, many early CDs did not use the original master tapes but second or third generation copies or even tapes EQ-ed for vinyl not CD. That lead to less than ideal CDs.

Sad but true, but if you want to see/hear well-mastered music or sound, it is usually in movie soundtracks or DVD's, etc such as DTS/AC3. For a combination of reasons, the dynamic range on movies are typically well preserved. In fact so much so that people themselves choose to *add* DTS/AC3 dynamic compression if their listening environment is not dead silent, etc.

A bit OT, but is it possible for one part of a mix to be compressed while other parts aren't? (Well, I guess I know that it is possible, but is it done?) I ask because there are a few CDs I have where the vocals sound very compressed, yet everything else sounds dynamic... It could just be my mind playing tricks on me though... Or maybe the relative loudness of the voice is causing clipping which makes the compression more noticeable?