Random Link: "HDR" means "high dynamic range". In photography, this refers to a higher variety of colour values. With regards to music, this refers to amplitude power. Dynamic range is usually measured in dB of average power.
Low dynamic range is the opposite, it's basically very high-throw dynamics hitting -0 dB as much as possible. This means next to no variation in the flow of music, no piano or pianissimo - all loud notes. An average dynamic range of -12 dB and more qualifies as low dynamic range. -6 dB is just plain awful pummel-in-the-face, yet modern rock/metal bands, and some pop records, manage to get there. Playing with a jackhammer might make more sense though (it has better dynamics).
High dynamic range is at least -18 dB (and lower) for an energetic piece. A symphonic piece with slow/pianissimo movements is more likely to qualify when it's more varied than -24 dB average (less is more here, really the variation ought to be better than -30 dB average).
There are two main reasons for an obsession with low dynamic range: one is low digital format resolution, another is simply higher human hearing resolution at better loudness. While the Fletcher-Munson equal loudness curves are there for anyone to see, the low digital resolution (and its companion poor transistor treble) is treacherous and subtle.
44 KHz/16-bit is the resolution of CD audio, which is hopelessly low. This is because the bits themselves are assigned gradually, across a nominal dynamic range of 96 dB, which leaves roughly 96/16=6 dB/bit. The lowest level is actually around -88 dB, which is bit 1 (-96 to -88 is bit 0, meaningless). What this means is that at -0 to -6 dB there're 2^16=65536 amplitude divisions, at -6 to -12 dB there're 2^15=32768, at -12 to -18 there're 2^14=16384 levels, etc. Basically, the lower the amplitude, the less bits there are. The chill is pretty noticeable if you ever try playing a soft synthesiser in 16-bit mode: the lower a part's output volume, the colder/hollower/blurrier it will be. Unconsciously, a mastering/mixing engineer will feel that same chill the lower his mix amplitudes are. To make it stand out, and have at least some semblance of energy, he'll have to punch the peaks all the way up.
Anything with high dynamic range will not play lively on a CD therefore, simply because low amplitudes will fall in low-bit areas. Hence the obsession with compression and low dynamic range - it's away to avoid dynamics fall into low-definition, weak-sound areas.
The trouble with low dynamic range is lack of perspective. Everything tends to be too loud rather than each instrument having its loudness contour and position according to volume in a mix.
Further trouble is the 44-KHz part of CD resolution, which again, is hopelessly low. It distorts treble: detail limit is sampling frequency/8. Therefore 44100/8=5512.5 Hz. Below 5512.5 Hz there're 8 sample points/cycle; above that limit, 7. Midrange pretty much, and very vague treble with less samples/cycle, and even less as frequency increases towards the closed space definition range above 11025 Hz (less than 4 samples/cycle). The trouble here is, a mixing engineer won't be able to correctly gauge perspective (and instrumental expression) if he doesn't get a properly defined treble. The tendency is towards flattening and making everything too loud or deformed - again, to increase contrast between instruments, which isn't there to begin with because instrument separation is defined by high-frequency content. Which is lacking definition. Saturation ("fattening") of waveforms is a workaround, but it only goes so far before everything tends to go out of control. EQ tends to add some unrealistic tonal skewing of its own: it takes some learned skill to EQ instruments naturally, and again, it also goes only so far. Mixing instruments with different volume levels (and perspective) will always require good sample/soft synth/recorded track definition.
The real solution is to record everything in at least 96-KHz resolution (which affords 12 KHz of detailed audio) and use at least 96-KHz samples. Then one might begin to learn to mix with high-throw dynamics, depth, space, and high dynamic range, just like those engineers that learned mixing on real mixing consoles and multitrack tape decks did. As soon as it is affordable, as many channels of 192-KHz recording devices as possible ought to be added as well. Otherwise there won't ever be enough space/depth/presence/expression to safely grasp and work with due to low-res digital formats' lack of definition.
Poor tuning is another problem. 440-Hz tuning is too sharp and stressful. A good equal-temper value is 435 Hz maximum, 430.53 Hz making more sense. Overly sharp tuning simply tires the mixing engineer out and makes him skew instruments out of proportion as they have too much contrast. Overly sharp tuning being too piercing, it also plays merry hell with vocals, especially female vocals, and it makes string and wind instruments screechy. Musicians also overstress and end up having trouble, down to their personal lives even.
Avoid low resolutions like hell. 16-bit is cold, hollow, and blurry masochism. 44/16 samples, even if taken of an analogue synth, can be dangerous also because compression/mixing can pull the low-res amplitude (sustain/release) parts up, noticeably chilling/hollowing the mix out. 16-bit samples tend to play vague and undefined. Anything below 96-KHz is not respecting oneself. The sound just won't lend itself to easy sculpting.
The main trick to HDR mixing is, therefore, having at least 96/24 samples/synths/tracks to begin with. The other trick is learning how to drop volume instead of increasing loudness. Starting with -12 or -18 or -20/-24 faders on each track in a mix is a good idea. There are many effects that can be stacked on a track to improve definition: Ferric TDS (dynamics/saturation), Voxengo Tube Amp (overtone/harmonic/saturation shaper), Voxengo Overtone GEQ (mid/side EQ), Naive LPF (lowpass/saturation/frequency effect) - those are just some free ones. Once a track stands out on its own and has a good definition, it can be pushed further down. Stereo imaging and stuff like analogue gear modelling plugins also help achieve good depth - stereo crossfeed is required both to ease a listener's strain and create a more natural image. Modulated reverb is another good thing here. As a simple example, a Blumlein shuffle of 2 on the Waves S1 Shuffler imaging plugin and a customised mid-side Overtone GEQ preset plus a Fusion Field reverb are a good start for any track. Simply emphasise an instrument's presence, its virtues and lackings, with the Overtone GEQ, alter its width with Shuffler, and its smoothness or echo-shoutiness/shimmer with reverb. Chorus and phase effects are good for adding presence as well - chorus makes anything stand out when done right.
The goal is to achieve maximum difference and contrast between tracks while preserving fairly low volume for each track. EQ and dynamic/stereo imaging help here, the whole mix has to be seen (listened) as such, as a natural system of musical notes, a symphony. The trick is not getting carried away so as to making individual instruments have an unnatural tone (overequalised or overcompressed, say). Again, this is where synthesiser/sample/take definition matters. Working with low-res samples/takes is similar to baron Münchhausen dragging himself by the hair out of a swamp, horse included. The strain is similar.
Baron Münchhausen dragging himself and his horse out of a deadly swamp.
In summary, low-res samples/takes tend to be flat and provoke compression, always reminding of their poor quality. High-res samples/soft synths/takes are natural and mix easily, allowing good dynamics and perspective.
Perspective is a matter of correct equalisation as well. Master EQ lifting high frequencies and lowering low frequencies (and boosting crucial midrange bands) helps. In general though, it is a matter of how close or far instruments ought to be, and of course, everything has to be correlated with reverberation (individual phase effects and reverb/delay help here). But that is a topic for another article.