Over the last few months I’ve become aware of several cases in which people produced audio using Levelator and SoundSoap together. The results were not particularly good. More significantly, in some cases, people charged for their audio production work.
Let’s examine the tools and what they do.
Levelator is a free tool created to simplify the process of increasing audio levels for recordings that captured very little signal (the foreground audio you want to hear such as a person speaking or a musical instrument). It was created by the The Conversations Network (TCN) to make life easier for the average podcaster by eliminating the need to learn how to produce audio. TCN recognized that most people who get into amateur podcasting don’t have an interest in understanding compressors, limiters and normalizers — much less do they have the time to learn how to use them. They just want to publish their podcast.
Simplicity is the key to the Levelator. A user drags a file to the application and it decides the best way to increase the audio levels. The problem is that when Levelator amplifies an audio signal, it also amplifies audio noise, some of which may not be obvious in the source recording. So, if you have a refrigerator running in the room in which you delivered a masterful monologue, the listener is likely to hear as much from your fridge as they will from you.
For the average podcaster, that’s fine and I respect that. Someone who bills for their audio production services should be delivering value to their client by using the right tools for the job. This is not about someone “whose eyes have never stung from the sweet sweat of a hard day’s work” (thank you Michael Bluth), this is about being honest and realistic about the work you do and the necessary tools to do it. Sometimes, audio level problems can be improved by tweaking the EQ; perhaps some compression or limiting could be applied. Using a paint roller to do quarter-inch detail work is the wrong approach. Levelator takes away the guess-work and the flexibility and the results can demonstrate that.
As the name suggests, SoundSoap is for cleaning up your audio. It’s been designed to help reduce (maybe even remove) unwanted noise (such as hums, hisses, rumbles…), clicks and crackles from a recording. SoundSoap has also been designed for simplicity. The user plays a section of audio that has the noise print with no signal (e.g. the ambient noise of the room when noone is speaking) and SoundSoap determines how to best reduce that noise from the recording, even when someone begins to speak. Then, presto, you get a better signal-to-noise ratio in your audio.
I use SoundSoap on occassion and find that in many circustances it does a fantastic job — though I often find myself dialing back the amount of noise reduction and threshold it auto-selects for me, and tweaking some other settings.
So, consider a situation where you have a recording noisey enough to merit SoundSoap and quiet enough to need something like Levelator. Imagine what Levelator will do to the recording, including constantly fluctuating the ratios of signal and noise. Think about which order these two tools could be applied and the effects of either order. While both strive to increase signal over noise, these tools are not necessarily complementary and anyone who produces audio for money should not put that much trust in automation.
When I think about the monetization discussions that have been taking place on the Canadian Podcast Buffet over the last few weeks, this stands out as another example of someone who asked the question “how can I make money from podcasting?” without asking themselves if they have something unique to offer and the capability to deliver it.

Of course, the best option is to have as noiseless a recording as possible in the first place, and if it’s not possible, to embrace the noise and aim for clarity over perfection. Our last Episode of Inside Home Recording was an example. As usual, Dave and I recorded our intro and outro with good mics in quiet studios. Our voices needed little processing.
But my interview with Paulo Pietropaolo took place in a noisy coffee bar, and you can sure tell! I made quite a few attempts to get it to sound better, and I think I succeeded, but not by eliminating noise: I just made sure that both he and I were easily audible above the din.
And I inserted a couple of very quick bits using the mic on my laptop lid. It was tolerable, but still a bit noisy. In one case I disguised that with background music; in another I used it as a transition between less- and more-noisy segments.
With luck, everything is clear enough to hear, and not too distracting. We used the Levelator in some places, and tweaked things manually in others. But if you want the least noisy speech with good levels, get that at the mic in the first place if you can.
Comment by Derek K. Miller — January 23, 2009 @ 3:27 am
The problems that SoundSoap and The Levelator address are different because The Levelator doesn’t try to increase signal over noise. If there is noise during quiet passages, it will be brought up along with the signal, so you get the best results by doing any noise reduction before applying The Levelator. Although I haven’t used SoundSoap specifically I have achieved many good recordings by combining noise reduction with levelating.
Comment by Bruce Sharpe — January 25, 2009 @ 12:31 am