top of page

Squish It Good: Hardware vs Software Compression

Hello Emmett, I hope you’re doing well.
I’ve been thinking about adding a hardware compressor to my chain. I know you use a lot of hardware. My questions are: 1. Does hardware sound that much better? 2. Is that the best way to prevent clipping during recording?
No rush getting back to me. Whenever you have time is fine. Thanks for all you do, friend.


Hey Robert, good to hear from you again!

Let’s get right down to business. 

1) No. 

2) No.

Simple, right?

You know I’m not letting you off that easy.

Yes, it’s no secret that I love analog hardware and have a good collection of hardware pieces in my studio. It’s important to remember that I am, first and foremost, an audio engineer. I’d also consider myself somewhat of a connoisseur of fine audio gear. I can’t tell you what makes a $200 bottle of wine different than a $2,000 bottle, but I can tell you a lot about the differences between a $200 mic and a $2,000 mic. 

When it comes to those listening to recordings, most people are similar to the way I am with wine. Above a certain point, they just don’t care about the minuscule differences that add an extra zero to the price tag. This is true of mics, preamps, and all forms of processing. With audio, once it’s good enough to pass the threshold into ultra-fine details, it becomes almost completely subjective. The differences, if they’re even audible, between hardware and software, are subjective and mostly unimportant. But there’s also an experiential element, and many people would agree, the experience of dialing in hardware can feel more rewarding than clicking a mouse. 

I like knobs and tiny details. I like seeing the faceplates and learning about the ingenuity that has gone into analog circuits. Just as someone who appreciates fine wine might enjoy learning about the growing season and climate for a particular vintage, so goes the background of a unique piece of analog gear.

No one can tell you what you like. The subjectivity part does have some value for certain buyers. But when it comes to the sound, I can’t honestly say my hardware sounds objectively better than software processing. Conversely, I can say software processing usually sounds objectively better than inexpensive hardware. 

People often forget about a really important detail, when discussing boutique or vintage gear. Those circuits were designed to be as close to perfect as possible. Especially with the vintage gear, had the creators been able to access quality digital processing, most of that expensive vintage gear wouldn’t exist. In the process of trying to capture clean, transparent audio, the flaws and limitations of some of the gear became desirable. That fact, however, does not change the desirability of pure, clean, transparent audio. People often associate expense with desirable sound, but that’s a flawed way to look at it. Both software and hardware have similar upfront expenses. Both require significant design and testing. Hardware, however, requires costly physical components, as well as the labor to build each unit, making hardware inherently much more expensive. But that doesn’t make it better.

For the second part of your question, the best way to prevent clipping during recording, is with your gain knob. I’m in the minority of VO engineers, in that I suggest, except under extreme circumstances, you find a safe spot for your gain and leave it there. I do not advocate riding the gain during a session, or even a single change to the gain during a session, unless it’s absolutely necessary. On the receiving end, gain adjustments are annoying for several reasons:

First, when you adjust the gain, it forces me to adjust processing. If I’ve dialed in a compressor, expander, de-esser, or noise reduction, and you adjust the gain, I have to figure out each of those all over again, or I won’t have a consistent result.

Second, if you’re using headphones, you’ll change your performance with each gain change. You will adjust your acting, based on what you hear, and the result will be less authentic. 

Third, my mix will have to adjust with levels. I will spend extra time trying to approximate where your voice would have been, had you not made the adjustment.

Fourth, pickups. If you don’t know exactly where your gain was set when you recorded, getting a pickup line to match the context is more difficult. It’s not impossible, but this also circles back to my second point, which is, if your gain isn’t the same, you’ll have a more difficult time matching your energy and performance.

Fifth, dirty pots. Analog potentiometers, such as analog input gain knobs, collect fine dust particles over time. This can create a crackling noise when the knob is adjusted. It’s not uncommon, nor harmful, but it’s annoying on the receiving end.

Sixth, room tone. The very low-level sound you hear when you stop making noise. Even very quiet rooms have room tone. When you adjust your gain, the room tone adjusts with it. Depending on the nature of the work, this can create unpleasant fluctuations. It’s much more natural if you get louder and softer, but your room tone doesn’t change.

Finally, preamp circuitry and sensitivity. Some preamps add more or less color, depending on where the gain is set. The preamp and mic work in tandem to create your overall sound. When you adjust the gain, you’re changing the nature of the circuit, by changing the relationship between the mic and preamp. Constant adjustments to your preamp will cause constant changes to the sound.

The average human has around 50-60dB of range, from a whisper to a yell.  Virtually all modern interfaces can provide a dynamic range of 110dB or more. Combine that with 24-bit (or higher) recordings, which allow 144dB of dynamic range, and it’s pretty simple to avoid both clipping, and equipment noise, just by staging your gain with care.

Now, with all of that said, I do think there’s a practical limit. If you do encounter a session that will involve both whispering and yelling, they are unlikely to appear next to each other. Realistic range for a typical dynamic performance is typically only about 25dB, which is simple to handle without touching the gain knob.

Back to your question, placing a compressor in the chain reduces dynamic range and adds noise. You risk compromising the quality of the entire recording, for very little benefit. Your overall noise floor will increase, and you’ll still have to take extra care when staging your gain. 

I could go on and on about the exceptions to the rules, and all the times my advice here may not prove to be the best course of action, but for the most part, setting your preamp gain correctly, is the straightest path to a good outcome.




bottom of page