Noise Reduction

So, as I learned during my recordings, church acoustics apply just as much to the heating system as it does to choirs and organs, which means that I have to put the actual creation of my sample patches off, and focus on getting as clean a signal as possible without reducing the fidelity of the organ. Which leads me to look into the process of noise reduction.

I’m going to preface this by saying, I don’t fully understand this process, these algorithms, or half of what I’m going to be trying to talk about in this post.

So, the process of noise reduction can be done in many different ways, but the first thing to understand is the two types of noise reduction algorithms: Fixed and Adaptive.

Fixed Noise Reduction

Fixed noise reduction, or spectral noise gating uses Fourier analysis of sample sections of noise to create a spectrum graph, which is used to gate the audio, reducing the level for any sound that isn’t about the threshold. This threshold varies at different frequency bands in order to sufficiently remove the background sound during sections that are above the threshold.

XNoise – Waves Plugin

This works best for samples which have a consistent noise throughout the entire sample, as the filter is static.

This gating is often combined with other processes such as frequency smoothing and time smoothing, which both are baffling to me for the time being. From what I can gather, these processes help to make it so that the effects of the noise gating don’t degrade the quality of the frequencies that are above the threshold at any given time, but again, I’m still not 100% on what they are or what they do, and definitely not how they work. [1]

Adaptive Noise Reduction

Adaptive filtering is a process which models the relationship between the input and output of a filter throughout the duration of it’s use. This means that it adjusts, or adapts, to the changing signal and as such, adaptive noise reduction is typically used for samples with noise that varies over the duration of the recording.

where x(n) is the input signal to a linear filter

            y(n) is the corresponding output signal

            d(n) is an additional input signal to the adaptive filter

            e(n) is the error signal that denotes the difference between d(n) and y(n). [2]

The process is similar to fixed reduction, but instead of using a Fourier analysis of the noise, the filter is created with a Least Mean Square algorithm. This takes a fixed filter like the Fourier example used above and changes the variables of the filter (ie. the threshold in each frequency band) in response to the input signal.


The quality of both of these types of filtering depend on the specific algorithm used. And quality noise reduction software isn’t cheap. Waves noise reduction plugins range between $200 and $600, iZotope’s RX3 runs $1200 for a full version copy, and CEDAR’s noise reduction hardware units (like the DNS1500) run upwards of $5000.

And it makes sense, in a world where audio is often second (or third) fiddle, it’s not always possible to get the best recordings, so noise reduction is a necessity.

Further Reading


Going Forward

As of now, it’s 1:30 am on April first.

I realize this blog hasn’t been as active as it could have been, but, in the coming days I will be taking the time to edit and post a backlog of research, annotations, and process logs. They will likely be out of order and might occasionally come off as half-baked.

By the end of this week, I will be posting links to downloads of the first three patches, including zips of the raw samples (cut and uncut).


Let’s rock this.

Decompression: A Series of Post-Recording Thoughts

  1. Organs are quieter than I expected.
  2. Church acoustics apply to the HVAC hiss just as much as the choir or the organ.
  3. It’s really easy to zone out when recording samples. Like, really, really easy.
  4. My feet are too big to work the pedals easily.

All told I got something like 850 samples, capturing four more stops than I expected on the manuals, and one more on the pedals.

I’ll have to look into noise reduction now, but I’m pretty happy with the results.

Choosing a Sampler

Tonight I go out to record my samples. I’ve already made the important decisions of which stops to record, and done a comparison of mic techniques. So the next question is, which sampler do I spend my time working with? What are the benefits of each one? How much do they cost?

Well, going through a list of features on these samplers, I am confronted with a list of features that I don’t need. In essence, to fulfill my mandate of being as transparent as possible I will only require the ability to internally loop the sample. That is to loop a section of it while maintaining the natural decay of the room. There are obviously other controls which I’d like, such as volume, but those aren’t nearly as important as the looping. At least not in the patch itself.

My initial instinct was to go with something like Native Instrument’s Kontakt, a stong, powerful, commercial sampler. The 400$ price tag was a bit scary, but it seemed like this would be a tool that was easy and fun to use. However, I started to question the nature of the project, the spirit of the thing, if you will, and I realized that if I were to choose a sampler like Kontakt (or AirMusic’s Structure), that I’d be limiting how this product could be used. That’s when I did some digging and found the SFZ file format. It’s an open-source patch format that’s already become somewhat of a standard in smaller samplers (at least according to it’s website). So if I were to work in this format I wouldn’t be limiting my audience to Kontakt, Structure, EXS24, or HALion, although it seems I would be alienating these audiences by using a format not supported by these major players.

So here’s my dilemma: Who do I build this for? How easy is it to transcode these patches? Should I even be worrying about this?