Suiting the Studio | Hilt Audio

by Ryan Thomas

In the early days of my recording career, I found that guitar parts that had sounded fine in my bedroom or on the stage ended up brittle, sharp and jagged sounding on a recording. I was dismayed and tried different microphones and mic placements with some success, but the problem tended to remain. My solution was to alter my tone, at recording time, to create a smoother, more rounded tone that emphasized the bell-like aspects of the guitar rather than the buzz saw. This worked fine until I bought a Fender Twin (which I still play) with the “kick-stand” bars on the side that allow the amp to be angled up toward the player’s head. I usually play this way since it allows me to better monitor my sound at lower volumes, but I began to find that at recording time, regardless of the amp, my sound had often become too smooth. The recordings began to sound bland and lifeless with a lack of energy.

I realized that my jagged, ugly tone had existed from the start — I was simply not listening to the amp head on, and so did not notice the details, which were savagely, but ineffectually attacking my shins. Once I began to look the amp in the eyes, so to speak, I could hear these details and subconsciously had adjusted my tone to something akin to what I would formerly save for recording. Then, when I would record, I would further smooth tone, creating something flaccid and impotent. It pains me to think about what my guitar might have sounded like from an audience member’s point of view years ago (for more reasons than this), but it serves as a good example of an inappropriate alteration of one’s playing to suit the studio.

Additionally, this issue is convoluted by the effect that the rest of the mix has on the part. Some parts sound great on their own, but when put into a mix, they become unsuitable, perhaps for their inablity to cut through the mix or simply because they just don't fit with the song. Here are clips of those tracks both alone, and in the context of a mix.

BlogSmoothSocialGuitarClip by HiltStudio

BlogRoughSocialGuitarClip by HiltStudio

BlogSmoothSocialMixClip by HiltStudio

BlogRoughSocialMixClip by HiltStudio

By the way, I ran the two tracks though a bus, so I was easily able to monitor their volume levels. The pairs' volumes are almost identical. Obviously the second cuts through the mix far more easily. Whether that is better is subjective, but I think in this context, it is.

Dynamics

Part of this tone smoothing effect was a compression of the part’s dynamics. The dynamics of a performance (the difference in db between the softest, quietest parts and the loudest parts) are an incredibly important issue that I will probably devote an entire article to at some point. Essentially, there are two ways to control the dynamics of a performance. You can use a compressor, expander, or an envelope/automation to electronically alter the dynamic range of a signal or file, or you can change the way you play the instrument so as to physically alter the difference in volume. In the studio, a relatively compressed dynamic range is generally desirable — you don’t want certain notes of a guitar melody to be far quieter than others on accident, for example. Plus when it comes time to mix the song, very loud peaks in a wave must be made to fit within the device’s headroom, lest horrible-sounding digital clipping occur.

It used to be important to compress one’s dynamic range simply from a technical point of view. Back in the day, when 16-bit audio was the standard, an engineer had to be careful to record a signal at the highest possible level before clipping, to maximize the amount of gain going into the analog-to-digital converters. However, in the digital realm, there are a set number of possible amplitude levels for any recorded audio wave: 16-bit audio has a far lower resolution than 24-bit audio, so it was important to utilize every possible level of amplitude in order to most accurately represent the original sound.

Engineers had ways to alleviate this problem, but they were expensive and could be detrimental to the final result. If the engineer had hardware compressors available, he or she could run a compressor between the preamp and the A/D converter, thereby maximizing the amount of average signal being recorded. The problem is that good compressors are not cheap, and anything done to the signal before conversion, cannot be undone. An engineer’s nightmare is probably to record God playing a guitar solo, only to badly compress the front end. A good engineer with a lot of experience could probably nail the settings without much effort and produce a very versatile result, but the kid with a 16 bit “studio in a box” without an inline compressor must find other ways to compress the signal. Namely, he plays his loud parts a little more quietly and his quite parts a little more loudly. For the most part, that works very well, but it takes practice and the final result might not sound as natural or vibrant as it did when he was playing live.

Thankfully, these days, almost any new interface and software combination one can find will support 24-bit recording. This means that an engineer no longer has to record audio as close to the zero-decibel mark possible. Because the 24 bit files have 256 times the amplitude resolution of 16 bit files, an engineer can record a signal that peaks a few decibels below the clipping point and worry about any dynamics processing later, where undesirable effects can be easily undone. In fact, some engineers feel that this is counter-productive and that many of these circuits sound worse when pushed to the threshold in this way. Unless the performance benefits from it, there is no compelling reason to physically compress one’s playing on modern systems.

Below, I've included two examples of recordings converted at extremely low gain levels. Both are recorded at my interface's lowest preamp gain using a Rode NT1-A, which advertises itself as being "widely recognized as the world's quietest studio microphone." While the 16-bit file has an incredibly loud noise floor, the 24-bit version is relatively clear. I believe if I had turned off my house's air conditioner and the numerous fans keeping me and my computer cool, the 24-bit version would be almost completely devoid of noise. Also, notice how the 16-bit file "ducks" during the loudest peaks, almost as if it were being compressed.

LowGainRecordingTest16bit(2) by HiltStudio 24bitVeryLowGainTest(2) by HiltStudio

Ryan Thomas records music in his air-conditioned house and plays guitar for The Machete Archive (not air-conditioned). I probably should have turned those fans off for science, but what-are-ya-gonna-do? Post your 16-bit tragedies in the comments below, or read more at hiltaudio.wordpress.com.