I’m not as knowledgeable as many here, so I would be interested to hear everyone’s thoughts. Dynamic Range DB. To give some context, this all started when I was trying to figure out why I dislike listening to Coldplay on my stereo. A friend of mine and I were talking about albums we enjoy, but don’t like listening to on a decent setup. At this time we were both experiementing with digital and archiving our music in FLAC. We started looking at the files in garage band and noticed all of the albums that we disliked listening to got a DR buzz cut. We didn’t look at our collection extensively, but Coldplay was a repeat offender. I eventually arrived at the above link. There is an article which escapes me that talks about inaccuracies when measuring DR on Vinyl, so if the article is correct, those scores are inflated. I know DR isn’t the end all for if the recording is good or not, but know it is a factor. I’d like to hear others’ opinions on the subject. Are there other factors that you find more important to the quality of the recording? Are any of your own highly ranked albums lower on the DR scale?
This is a complex, and subtle area.
For one thing, most listeners, on most systems, in level-matched playback, will tend to prefer a version of a track that has had its dynamic range compressed vs. the full-range version. This reality is what started the “loudness wars” in the first place.
Anyone with reasonable gear (I’m talking entry-level, but competent, stuff), will be able to tell the difference between full-range and dynamically compressed versions of the same track or album. But which they prefer is going to be quite personal.
A commonly cited factor with dynamically-compressed recordings, where the listener is just doing blind comparisons, is that the lower-DR version will have more evident detail. This is typically because otherwise low-level elements in a full range recording get brought up in level significantly and are much easier to hear as a result. (MQA content often exhibits this effect, even when it is provably lower in final DR than a non-MQA version of the same thing).
For another, not all low-dynamic-range music has been compressed. It may simply have been created that way, so no version with greater range may exist. In which case, it is what it is, and you either like it or don’t.
Some music simply works better with limited dynamic range.
Then there’s the question of listening environment. A full-range dynamic range presentation, in a quiet room (say 30 dB ambient) is can be hitting peaks of 120 dB or else the quieter parts will be below the noise floor of the room. If listening in a car the ambient noise-level is probably closer to 60 dB, so dynamically-compressed music often sounds better in that environment.
Note that I am not arguing in favor of dynamically compressed releases/mastering/re-mastering.
In general where I have a choice of a copy of an album that hasn’t been squashed vs. one that has, I will personally almost always choose the full(er) range version (and if I’m buying it, that’ll be what I’ll default to).
One of my favorite examples, here, is Paul Simon’s Graceland, where the “High Resolution” version is bloody awful compared to the original 1986 CD release. It hasn’t been squashed that hard (from an average DR of “14” to “9”), but it kills the life of the whole album.
Contrast that with, say, Adele’s “25” only has a DR of “6” to begin with, and at least one subsequent release that’s been squashed to just “4”. Musically quite engaging, but technically rather arsey. Would it be more enjoyable if it had greater dynamic range? Maybe. But I am not focused on that when listening to it (or such things).
Again, that often depends on whether a more full-range version exists and, of course, whether I’ve heard it.
In other words … where fidelity is always harmed by DR-compression, enjoyability is much more subjective.
I would not compare dynamic range from the DR-database between vinyl and digital versions.
Comparing vinyl-to-vinyl is fine, and digital-to-digital, but there are artifacts and considerations at work that mean they’re not useful to compare cross-format.
Other thoughts …
A technically good, high-dynamic range, production won’t improve my enjoyment of music that I just don’t like in the first place.
There are definitely cases where I have found a subsequent, dynamically compressed, version of an album preferable to the original (they’re not common, but it happens). Though usually this is because the range was so extreme, important parts were lost/suppressed unless listening with very high peak levels.
This is a fascinating thread. I know so little. I would have thought the higher the DR the better. It goes to show more doesn’t always mean better. It’s a complex subject it would seem.
From a technical, or fidelity, perspective, higher DR, all other things being equal, is better.
From a musical, enjoyment, or engagement, perspective, its both highly subjective and both musically and environmentally dependent.
Audiophiles, or music connoisseurs, listening in reasonable conditions with competent gear will, I expect, generally gravitate towards preferring high-dynamic range productions, but I’d be surprised if that was the case for 100% of music/situations even then.
My understanding is that DR is the primary culprit behind the mythology of vinyl “fidelity”. Essentially, that HiFi enthusiasts comparing early CD’s that had been “squished” for lower grade electronics misinterpreted the compression as a limitation of the format.
Does that track with your understanding @Torq?
Well, while I am pretty interested with DR stuffs and actually respect the effort of DR database, this term per se is quite misleading in my opinion. Indeed I’ve been fooled multiple times.
Like @Torq said, it’s better to approach this concept technically. I’d like to add the importance of definitions.
Conceptually, dynamic range isn’t difficult. It means how the loudest part differs from the quietest part. This, however, introduces A LOT OF CHALLANGES in real measurements.
“Loudest” is simpler, once ‘loud’ is well defined (which is not)
Quiet is tricky, because in this digital era, we have at least three kind of different quietness in every recording: (1) no digital signal (zero value), (2) recording environment noise floor, and (3) “effectively” lowest sound signal. All are associated to some extent, but not entirely identical.
Thus, more technically convenient definition is prevalent. Called crest factor.
Instead of using the quietest part, crest factor takes the difference from the average – specifically root mean squared (RMS) level.
DR is nearly similar with crest factor, only differing in ‘weighted loudness’ instead of ‘real signal level’. Google EBU R128 (broadcasting standard) if interested. I can’t say DR uses R128, but it evidently uses its own way to measure loudness. And in my observation, various (perceived) loudness measures are very similar.
Back to the topic, in reality, it’s very, very dangerous to interpret high DR better without full context.
First, as I explained, DR is the difference between high and average. Most low DR tracks have higher RMS level. Definitely genre-dependent.
Second, some tracks have DR reduced when remastered as dynamic compression applied in order to raise RMS. Unlike typical beliefs, this isn’t necessarily bad. What really matters is whether and how such compression affected important information. Not every piece of information is important.
Third, DR has nothing to do with perceived micro-dynamics. It is possible that large DR (others being equal) may benefit smaller changes. However, it hugely depends on context and art/skill of post-processing.
Bottom line, DR is just a tool with MANY limitations requiring TONS of caution. Without detailed information and context, it’s very misleading.
It’s certainly one factor.
Four other commonly cited ones I’ve heard are that:
Vinyl offers superior frequency range vs. CD. This is true on a technical level, but is dependent on a number of factors (including the equipment and material in use and its condition) and, even where it is true, raises obviously audibility and relevancy questions.
Vinyl offers superior temporal resolution vs. CD. In other words transient response is better due to the continuous vs. sampled wave form.
Vinyl is smoother and offers more micro-dynamic resolution vs. CD because it is truly analog* rather than being sampled and “coarsely” quantized.
And the “big one”:
- CD sounds bright and tinny compared to vinyl. There was a lot of truth to this early on. For one thing, the “pre-emphasis” bit was not always set, or interpreted, correctly, so that could result in unnecessary treble emphasis on playback. And for another, more than a few high-profile early CD transfers were done with completely different, sometimes flat out “wrong” EQ applied (e.g. the Beatles catalog).
Only the frequency response claim is really valid for a properly band-limited recording. And that’s probably the least relevant of the claims anyway.
All that said … when it comes down to it, that vinyl is not the technical/fidelity equal of properly done digital, that doesn’t necessarily correlate to digital being more enjoyable. I know that, for myself, I typically enjoy listening to an LP played on my turntable, through my speakers, more than from a digital copy or CD version through the same system.
Note, I am not making nor supporting the above claims, I am merely reciting what I’ve heard as reasons why CD isn’t as good as vinyl!
Interesting stuff here. Too much to comment on while at work, but I’d like to add something. One thing that really stood out to me on the Coldplay album was the cymbal crashing. I would look at the time stamp and view it in Garage Band. Each time the peak was chopped off. Coldplay’s waveform looks like a buzz cut, so I found it more useful to look at U2. Not as much compression, but each time the cymbal crashed, it was missing the peak. I was perceiving it as harsh before looking at the waveform. It might be a fun exercise to try. I’d like to find an album next with compression that I enjoy listening to then look at the waveform.
That could be deliberate.
Or it could be specific to the cymbal track within the overall mix.
(And for real instruments it can be a capture-time issue.)
OR it could be a result of global levels being pushed too high during overall DR compression.
Even the amateur recording, mixing, mastering and rendering tools allow you to configure warnings for such things, so usually when it happens its a choice on the part of the artist or engineer.
You can play with such things using any number of DR compression tools/plug-ins if you want to, either in real-time or rendering to a file, even for albums you already own (what you cannot do, of course, is UNDO blown levels, or clipping or recover data lost from smacking into the 0 dB ceiling).
Compression is very important during production, as it sets each instruments in the foreground or background. A complex production (e.g,. 4 piece pop/rock or greater) has specific and narrow frequency lanes for each instrument. Yet, each part must be hearable despite the inherent volume of each instrument.
In most pop music (e.g., Coldplay) the voice is front and center, with a lead instrument drifting in and out of focus (e.g., guitar, keyboard, cello, bass, violin, etc.), and percussion in the background. The trouble is some instruments sound better or require a specific volume to sound their best, and this often overpowers other elements.
- Drums are notorious for being too loud, and have even been put in a plastic cage or have had a towel stuffed inside to dampen them.
- Guitar amplifiers (e.g., Marshall) sometimes sound best when cranked to maximum, and thereby overwhelm everything else. Some classic recordings were made either with tiny little amps moved to the foreground (e.g., some Led Zeppelin) or created on a truly overpowering amp with the guitarist in the next room (e.g., Randy Rhoads on Ozzy Osbourne’s Crazy Train).
- Electric guitar parts and notes range from delicate or faint to searing, sizzling, and overwhelming. As such, a ‘compression’ pedal is routinely used to control the volume but produce one of a hundred desired tones.
- In context of the MASSIVE volume of drums and 100 watt guitar amplifiers, other instruments would be totally lost without a boost. Weaker sources relative to amplified instruments include mandolins, chimes, and the human voice.
- Sometimes heavy compression is used for a low-fi, garage rock, punk rock effect. In the Loudness Wars period this followed The Strokes “Last Night” single and most everything from The White Stripes. However, these artists were seemingly imitating cheap retro recording equipment.
All in all, it’s a delicate balance and heavy compression can sound just fine. I was surprised to find that Radiohead’s “Let Down” (from OK Computer) is actually heavily compressed – it came across to me as nuanced. The tinkle tinkle parts I guess.
That’s some very interesting information, thank you. Today is a learning day for sure.
There used to be - and I’m sure still are - dynamic range expanders. I’ve heard hardware ones, and the math is interesting, so I’ll bet there are some software alternatives.
Absolutely. I had a friend who lost 3 fingers* doing that.
3 fingers of decent Tequila
There are, but they cannot recover the lost information, they merely “stretch” what still remains.
The current range expanders are called Focal Elear and Elex…
There are many. Theoretically possible enough. Indeed my behringer can do it to some extent.
It’s like image/color processing. Everything you can imagine is possible. Not all possible changes are either meaningful or beneficial. No free lunch.
And this reads more associated with how headroom and clipping were treated, rather than compression/DR. Compression and dynamic ranges barely affect peak levels, and vise versa.
If you’re really interested with how files are measured, I suggest to use various information (not merely overall DR)
Below is one analyzed result of Beyonce’s Partition. Its dynamics IS compressed. But quite skillfully done.
Thanks to all that contributed so far. Comparing this to image editing puts it in to context.
So music with heavy distortion will probably have a higher RMS and thus lower DR? Makes sense. Additionally if the track never gets loud (say Dave Matthews Say Goodbye) it may have a low DR.
So I guess I need to read up on how DR is measured, and how to interpret all of those graphs.
Here is an interesting article related to this discussion for anyone interested. Micro Dynamics
For an interesting layperson’s perspective on dynamic range compression and the “loudness wars” have a look at this NYT piece (link at bottom) that was written earlier this year. It is titled, “They Really Don’t Make Music Like They Used To”. In the piece they compare the dynamic range of various popular songs. In particular I found this part of the article interesting:
Several years ago, Chris Johnson, an audio software developer, tested a theory, espoused by some anti-loudness activists, that the hyper-compression roiling the industry was partly to blame for shortened careers. Using a list of all-time best-selling recordings, he rearranged them by “commercial importance,” assigning each a score derived by multiplying an album’s number of platinum certifications (how many millions sold) by the number of years it had been on the market. These were records that were not merely popular — they also displayed longevity. He then used software to analyze the sound waves of each album.
His findings revealed they had a common trait: these albums, even across genres, had extraordinary dynamic range. The most commercially important albums, he wrote, featured lots of “high contrast” moments, when “the transient attacks of instruments” — very brief outbursts of high energy — were allowed to stand out against “the background space where the instruments are placed.” This was especially true for vocals and percussion (one of the more intriguing similarities, from Michael Jackson’s “Thriller” to Pink Floyd’s “The Wall,” involved what Mr. Johnson called the “hit record drum sound”). Loudness has its place, but most of us like our music to have breathing room, so that our eardrums are constantly tickled by little sonic explosions. In a tight, compressed space, music can get asphyxiated.
This is a horribly confounded thesis and I’m not buying it. Any commerce based analysis must consider the technology of the era and the changing economics of music.
[And my eyes roll when I see “Grammy Songs”! Egad, basing quality on the Granny Awards! Jethro Tull = Best Hard Rock/Heavy Metal!]
- Distribution was expensive and tightly controlled by record companies and radio stations before the loudness wars. Loudness boosts were used during TV broadcasts many decades ago (absolutely by the 1970s) to grab attention. I recall seeing my father’s old Popular Mechanics magazine or somesuch with a DIY hack for a “TV Ad Mute Box” (before remotes were common). When a musician is desperate, they will try anything to get attention. Loudness didn’t matter (couldn’t matter) when a handful of TV networks and radio channels ruled.
- Every year new releases compete with the back catalog and dilute interest in any given artist or release. There are now a 100+ genres and no one has the time to explore them all. During the pre-loudness era there was only really Classical, Country, Pop (including Urban Contemporary), or Rock. Punk received no radio play, heavy metal received no radio play. Landmark bands (today) were often known through a few big city radio stations, college radio, and word of mouth as ‘alternatives’ to the bland, easily accepted, beige, commercial middle. This persisted until the late 1980s. Examples of the early 1980s underground include REM, U2, The Clash, The Smiths, The Cure, Metallica, The Misfits, and Black Flag.
- There are plenty of artists/genres with limited dynamic range and not much background space, by their very conception (e.g., early Kinks -> Ramones -> Motorhead -> The Strokes). These genres have devout followings but fell outside the beige content push of the Radio Era.
- The style of song described seems to follow the old definition of ‘mainstream’ big pop songs. The convention follows very old classical compositions that alternate between solos and orchestras. Many songs internally shift moods, but the pauses and breathing room of the song are confounded with dynamic compression. These are very different topics.
- Until well into the 1980s, many people listened to extremely poor audio through AM radio, junky record players, single TV speakers, and 8 Tracks or cassettes. No matter the original dynamic range, many people heard just a pale image of the optimal quality.