Dynamic Range in audio is the range, difference, or ratio between the loudest and quietest parts of a piece of audio. For music, the 'silence' between songs or during pauses in the music are typically not counted. So the dynamic range in music is typically the ratio between the parts played most quietly and the parts played the most loudly.
When unamplified acoustic music is played, it has an initial unmodified dynamic range based on the music being played, the performers' interpretation of that music, and the acoustics of the room where it is being played. When that music is recorded and produced, it may then be modified to fit particular needs, and this often involves 'compression' which reduces the dynamic range.
One of the 'particular needs' that has been prominent in the music industry in the last two or three decades has been to 'maximize loudness'. It is a characteristic of human sound perception to think that louder sounds better - at least at first. So based on that, many artists and producers want their music to be 'louder' than the others. This started what is known as the "Loudness Wars". If you play song A and compare it to song B (without changing the volume knob), and if you like the tunes equally, which one sounds 'better' at initial listen? Usually, it will be the louder one.
So how does a producer make one song louder than another? You might think it's as simple as turning up the volume during the production process. But it's not. CD's have a 'maximum volume' limit, also known as "Full Scale Digital". And the volumes that producers are putting on CDs are bumping up against that maximum limit. Trying to turn up the volume past the maximum limit will not work, as shown in the following diagram. The section of audio above the maximum limit will be heavily distorted if this is attempted.
If the song is already bumping up against the maximum loudness at its loudest parts, and the artists/producers want to make a song louder, they have to 'compress' it. This means turning up the quiet and medium volume parts, while turning DOWN the loudest parts. Then, to make this an 'upwards compression', the volume of the compressed audio is turned up so that its loudest part is just under the CD maximum limit. When you do this, the 'average' volume level of the song is louder, but nothing gets any louder than the maximum limit.
The big problem with this approach is that you lose the peaks and valleys of the volume landscape. Dynamics become 'flat' You had to turn those peaks down to avoid going over the maximum limit imposed by the recording technology. Typical peaks are drum hits or crescendos. When they are 'compressed' and reduced, they lose their impact - both technically and emotionally for the listener. Similarly for quiet parts that are turned up by compression. They lose their delicacy and intimacy. And you lose the impact that you normally would experience during dynamic volume changes, because everything is in a smaller volume range. ppp<->fff becomes p<->f or worse.
Dynamic Range and Compression in Music in the Past and Today
Audio loudness is typically measured in decibels (dB). This includes dynamic range measurements. (Note: To humans, a 10 dB difference sounds about like a doubling of the volume level. 20dB difference sounds like 4X the volume level, etc.) For live unamplified very dynamic orchestral music, dynamic range can approach 30 dB, while 18-24 dB is probably more typical. However, for a string ensemble or cello/harpsichord sonata, the natural dynamic range is typically a lot smaller, such as 7-10 dB, because the music doesn't vary in volume nearly as much. When this is recorded and produced, it can be compressed or it may not be. And if it is compressed, the amount and type of compression can vary.
For studio-recorded rock and pop music, there is no real 'live unamplified' version of that to use as a reference. However, compression in rock music has been common for decades. It is often used on individual tracks and voices to balance disparate sounds. For example, vocals are often compressed so that they are in a consistently small range of loudness levels regardless of the note being sung. This way they don't get lost in the mix. And drums are often compressed to bring up the level of the drum sound that occurs between transient hits to make the drums sound louder and more present. However, where compression has really been overused lately is in the final mix. This is known as mastering compression. And in the 'loudness wars', it is this compression that provides the overall CD loudness that various artists and producers have been seeking. This is the compression that does what you see in the previous figure.
In that figure above, you can see that the 'average level' is higher on the right. And the entire mix is compressed (both the loud and quiet parts). For a typical rock song, a relatively uncompressed version ("Full Dynamic Range") as shown on the left of the image above may have 15 dB of dynamic range with an average level of 10 dB below the max, while the compressed version on the right may only have 7 or 8 dB with the average level only 4 dB below the max. So the average level of the heavily compressed song is 6 dB louder than the uncompressed or conservatively-compressed song. At first listen, it sounds better, because it sounds louder. But as you continue to listen, you'll probably adjust the volume level (of either song) to what you find comfortable. You might turn the uncompressed song up, or turn the compressed song down. At this point, the playback in your room or headphones on either song will probably have roughly the same average level. And at this point, which recording do you think sounds better?
When average volumes are adjusted to be the same, in almost every case the song that is uncompressed or not overly compressed will sound better. The minimal/no compression song will sound lively and exciting, while the overly-compressed song will sound flat, and may grate on the nerves.
And listeners will adjust the volume to what they find comfortable. Radio stations do the same - they adjust the levels of the songs they play to have a similar average volume level.
Similarly, many media players available today that play CDs, MP3s, and FLAC files have a feature called "ReplayGain" which automatically adjusts the 'average' level for the listener, so that all songs have roughly the same average level. ReplayGain will automatically do the level adjustment as shown in the figure above. And in that case, which song will sound better? The answer for most people will be that the one with the fullest dynamics sounds best - the one on the left.
So given that average levels are almost always adjusted to the listener's preferred volume, does it really make sense to compress audio to make your song sound louder than someone else's song? In most cases when this is done, the heavily compressed version actually won't sound as good.
Compression and Dynamic Range at Lucid Recording Sciences
When Lucid records classical and acoustic music, we tend to use a minimum amount of compression, if any. We want to provide the 'Full Dynamic Range' experience that allows for full enjoyment of the music. The loud parts are LOUD as they are supposed to be. And the quiet parts are quiet and delicate as they were meant to be. When listening under good conditions, this provides the ideal scenario. When recording pop and rock, we do use compression, but we use it conservatively. We want to maintain a good dynamic range that allows the music to have impact, to breathe and live.
In order to provide this full dynamic range for dynamic classical recordings, where the peaks have room to grow before they hit the CD Maximum Limit, we have to put the songs onto the CD at a lower average level than just about any pop or jazz CD. A typical 'average' volume level for a classical song on our CDs may be 18-20 dB less than maximum (sounds 1/4 as loud as the maximum). By comparison, well-mastered Rock/Pop CDs may have an average level of 12 dB below max, whereas the Rock/Pop CDs that have been competing in the Loudness Wars could have average levels as high as 4 dB below max! Now that's compressed!
By setting our average level at 20 dB below max for dynamic classical recordings, this allows the peaks of the music to be up to 4X louder than the average level. Peaks such as this are common in dynamic classical music, but often get 'compressed' by the lesser classical houses or by anyone participating in the 'loudness wars'.
So the big advantage of recording at these low levels is that we can provide recordings that breathe and are full of life. They provide the Full Dynamic Range experience, and sound much closer to being at the live event. There are a few tradeoffs we have to accept when we create CDs with high dynamic range and with an overall lower volume level:
However, Lucid's main goal is to make great sounding recordings for foreground listening under good conditions, so we think the tradeoffs are worth it. Since we have high dynamic range and relatively low average volume as compared to many CDs, if you also listen to rock and pop or other CDs with that have received significant upwards compression, you may find that you need to turn up the volume when listening to our CDs to get it to sound right.
So all of the above is essentially an explanation of why we have the following advice on some of our CDs:
To preserve Full Dynamic Range, this recording is intentionally quieter than many CDs. For full enjoyment simply
Turn It Up!
For questions, comments, or inquiries, please contact us.
Phone: (602) 224-3799.
Copyright © 2015 Lucid Recording Sciences