Since the earliest compressors were conceived and built, the ability to modify, control, and maximize the dynamic range of a musical performance has been the quest of many an audio engineer. In the early days of audio, limits on a recording’s overall dynamic range were dictated by vinyl – the state of the art in music delivery until the CD’s debut in 1982. Today, with virtually all music being recorded and massaged in a digital environment, it’s become standard operating procedure to limit, compress, or maximize the dynamic range of a track or an entire mix.
On the surface, having your track as loud or louder than the competition seems like a good thing in the world of car radio and MP3 devices. But in speaking with some top engineers from the music, broadcast, and mastering fields, maxing your levels is not a recipe for a great sounding recording. In fact, according to experts Michael Bishop, Patrick Fitzgerald, and Steve Hall, you may be doing more to harm your recordings than good through the liberal use of the compression and dynamic control programs found in every workstation across the country.
Steve Hall’s Future Disc Systems mastering studio is in McMinnville, Oregon. With more than 30 years of experience as one of the top mastering engineers in the business, Steve’s work spans the era from when vinyl was king up to today’s high-resolution digital formats. He has mastered thousands of albums for artists such as Madonna, Chris Botti, Green Day, Sisquó, and the Grateful Dead, to name a few. Of these more than one hundred have achieved gold or platinum status.
Would you say dynamics have been squashed out of music today?
They’re pretty much gone, yeah. So much of the music I hear today is totally hammered and the result is that most of the life, detail, and energy of the original performances are just gone.
How did we get to where we are today?
When vinyl was the dominant release format, artists and labels were always concerned about cutting their records as hot as possible. Factors that were limiting on vinyl were how much bottom end the disc could hold and how long a side could play. That said, most of the dynamics heard during the original recording session were pretty well preserved on the LPs. With 45s, you always wanted your record to pop when it came on the jukebox or radio, but not to the point that the performance was lost. So the artist, label, A&R guy, and mix and mastering engineers would all listen carefully and agree on some sort of compromise between maintaining the recording’s musicality and achieving the hottest level possible.
Today, with digital, there aren’t those kinds of format limits in place. There’s simply a brick wall looming at the end of the digital audio pipeline. With finalizer-type tools, and various multi-band compressors, the goal is to modify the master to constantly have everything as loud as possible, and you pay a price for this so-called “normalization.”
When did the shift occur?
You only have to go back about ten years and listen to the masters from that era, really anything before the mid-90s. Listening to those albums is a totally different experience. There’s life, energy, and detail in those releases that you seldom hear today. And the main reason is that now everything is just maxed out from the beginning of the album right to the final note.
Recordings aren’t the only place you’ll hear this problem. I just heard a holiday radio spot the other day for the Salvation Army and they used the sound of the bell ringer as a background cue. Unfortunately, the station’s compressor picked up that high frequency signal and compressed the heck out of it so it was louder than the announcer. That’s just one example of compressor abuse.
Misuse of compression is even worse on television. Especially on cable TV, things are often abusively slammed. You are constantly reaching for the volume control on your remote to deal with a situation that is made worse by compression, rather than being improved.
What can musicians and mix engineers do to try to avoid the pitfalls you’ve discussed?
The first thing I would suggest is to avoid putting any one-size-fits-all-type compressor on your final mix. Allow your music to breathe a bit and use compression selectively, not as a fix for your entire album. It’s partially because there are so many tools today that it’s very easy to go overboard.
Next up is multiple Grammy award-winning engineer Michael Bishop, of the renowned audiophile label Telarc Records. Telarc, based in Cleveland, OH, has won an amazing 40 Grammy awards for its outstanding recordings in classical music, jazz, and blues.
Michael, are dynamics dead in today’s popular recordings?
I’d say that in most non-classical recordings, the answer is yes. Load any CD into your workstation and look at the resulting waveform. You’re seeing rectangles where you used to see waveforms. There used to be dynamics there! And what’s amazing is that all the while, you look at some of the recording forums that go on, and every once in a while a thread will start up about old Frank Sinatra recordings or older rock recordings from the ‘60s and ‘70s, and these old recordings are being held in great reverence.
People are often referring to them as “the sound to go after,” all the while taking their own present-day recordings and maximizing them, hitting the “normalize” button, and cranking everything up to be maxed out all the time. If you go back to those classic recordings, compression was almost non existent. And yet they still sound loud and have tremendous depth. There’s an example where less is more.
Telarc is known for producing audiophile recordings. How do you go about recording and mastering to ensure your recordings have depth and life?
We do everything in high resolution, most using the Sony DSD [Direct Stream Digital] format. Our source masters are always DSD for all in-house engineered and produced sessions. We do have some outside-produced projects, and they are so accustomed to delivering a master with everything maxed out. So we ask them to go back and remove some of the compression because it doesn’t mesh with the rest of our releases, which tend to have a pretty wide dynamic range.
Going back a few levels of “undo” puts some life back into the recording. I don’t think it’s too much to expect the consumer to bring their volume control up a little bit. When a consumer puts a recording in a multi-disc CD player or an iPod with play lists that have everything thrown in together, levels from song to song become an important concern. In that context, if one artist’s release doesn’t hold up against another’s, that’s a problem.
How do you deal with that?
We’re releasing in two formats on many of our releases: the standard CD and the high-res SACD version, which will be in both stereo and surround. But we never produce square waves where there once was music.
We use a moderate amount of compression, listening carefully to the results as we go. We have two other engineers in-house besides myself, and we each approach things a bit differently. One of our engineers prefers to use the Sequoia system and the Waves software bundles that come with it. He often does the releases for our Heads Up label, which focuses on smooth jazz, so they will have a sound that is compatible with that genre.
On the other hand, I like to use the Sadie system and the TC 6000 for more of the straight ahead jazz, blues and classical releases that come out on Telarc. Regardless of the artist or release, we’re never going to get in a level war. About the worst thing we could do is to start comparing apples and oranges, instead we just make each release sound as good as it can for the specific format it will be listened to on.
You have to keep in mind that too much intrusion into the dynamic range of any music is taking out the life and taking out the fun of listening to it. It also makes it very fatiguing to listen to, no matter what the format.
Even in high-res, there are peaks that eat up the headroom that I don’t need. DSD is absolutely unmatched in being able to capture the fast transient response, say of a loud snare hit. But that transient may be 20 dB louder than the meat of the main snare sound that I want to focus on, so I don’t always want it to be a total literal recording. I’m looking for the overall musicality of the track and to see that it serves the music, getting across what the artists wants. I can’t only take a purist approach and let it detract from the material, but it has to be in a transparent way and not beat up the music. Unfortunately, anyone who’s listening carefully can point too many examples of music that’s been mangled on the way to CD.
There’s a great fallacy out there that if you make the CD really loud, it’s going to come across much better on radio. It’s actually exactly the opposite.
Any time you listen to a jazz station, and you listen to a modern recording that’s been pretty maxed out and then you have a classic track come on – say Louis Armstrong or Ella Fitzgerald – that track just jumps out of the radio in comparison. That’s because it was not squashed when it was mastered and released, which once again demonstrates that less is more when it comes to using compression.
Do you encourage the musicians to take dynamic control of their music during the recording process?
Quite often, yes. If I’m doing a straight jazz date, we’ll do a sound check then I hand it off to the musicians. For a jazz artist like John Pizzarelli, the recording is done live straight to two-track and surround, so once levels are set, the performers determine the balances. In the case of the orchestra, the conductor will determine balances.
Over the past 20 years, Patrick has mixed thousands of short and long-form soundtracks including the famous MTV animated station IDs, hit video games, a number of feature films, DVD releases, and many thousands of commercials. He’s faced with a wide array of delivery requirements as he works with an increasingly diverse range of clients who may or may not have time to learn about the intricacies of how audio is delivered to consumers via the many paths available.
From your perspective mixing for broadcast or film, are dynamics dead?
I wouldn’t say they’re dead, but it is a daily struggle to make sure that there are some. Working with music is becoming more of a struggle, because what I am now getting is so compressed to start with. I think that’s because many musicians think, “My music mix is the end of the line, and I’m used to the levels I see on commercial CDs, so let’s make it like that.” This isn’t a good place to start because it limits my options with the music track. You look at the waveforms of many of the music tracks that are sent to me and everything is tucked right up next to digital zero and all the peaks are just cropped right off!
It’s ironic when you look back at the history of recording, how the value of dynamics in music has changed. Consider the evolution in the middle of the 20th century, from really poor quality tape recording to really good tape recording then to vinyl. During those years, there was always this quest to answer the question: How do we get more dynamics into the recording process? And now that we have a system where you can have absolutely incredible dynamic range, everybody throws it away.
What about other media, such as video games?
With video games I feel there’s definitely room for keeping the dynamic range in your music because with some of the production values going into video games these days, including surround sound, you can really have an incredible music experience as part of the game. For me personally, if the music does have dynamics in it, it’s more alive. As opposed to something that’s hugely compressed, you’re just going to turn it down to a level that’s comfortable and it’s going to stay there. It will never blow you away. So if there’s a section of the game where the characters are dialoguing and the music is underneath, and then you progress to an action sequence or a run up to a big event, that music is never going to get any louder. And that’s a problem.
Do you use compression?
I use compression all the time, but I’m careful to never let the compressor mix for me. I probably spend the most work evening out dialogue. Twenty years ago we worked primarily with professional actors and voice artists, people who spent years training to project and control their voices. They knew about breath support. You can hear anyone who’s had that training right away, because their levels come out much more even. A lot of the dialogue or voice over we get today is voiced by people that don’t have that training, so there’s a lot to do in order to make it all intelligible.
What I’ll try to do is to level the track out a bit before it hits the compressor. I think of it as having a gain stage before the compressor that I use as a dialogue pre-mix. That’s different than the approach many people take, which is to put the dialogue track through the compressor. Then the loud things are going to get audibly compressed, where it sounds like it’s being squashed, and the low level audio still might drop off.
What compressors do you like?
Most often I reach for the Waves bundle, because it’s so practical. It doesn’t have a specific sound the way that some of the plug-ins do that are designed to emulate a classic compressor like an LA-2A, or the functions of a classic SSL or Neve compressor.
I don’t set my compressor by how far it’s taking the level down, but I do look at the metering. Let me take it back a step, because I usually work with high compression ratios, usually 50 to one or infinity to one. But it’s not kicking in until a very high threshold, so that for the most part, the signal is passing through the compressor unaffected. And on those peaks where it does pull it down, I’m seeing -3 to -6dB reduction. If you use a lower compression ratio, say 4:1 or 2:1, then the compressor will affect most of your dialogue.
Any final tips for musicians or composers writing music for various media?
If you want to squash the music, go ahead, if that’s the sound you are going for. But keep in mind a couple of things before you do that.
First, is what you’re doing to your mix the final step? Are you really the end of the line? Is your mix the one that will play on radio or TV? If so, then fine, use up all the level you want, and it should sound like what you heard. But if it’s going to be used as part of another program, a soundtrack, anything where it may be mixed further – then don’t suck all the life out of it, because you can never put it back. The same applies if you are going to send it to a mastering engineer.
Second, if you do want to do a mix with lots of compression, then provide an alternate mix that keeps the original dynamics in the performance. That will allow the next person down the line to use what will work best for the whole project.