11 Jun 2008

Fairlight CMI Series IIx - First impression

Ever since I was a child (I grew up in the 80's) I had an unhealthy fascination with the Fairlight. Buying every record that had "Fairlight programmed by..." in the liner notes and trying to find all the info and sounds of this mythical instrument.

Finally I have a Series IIx in my possession and I've been busy learning it and using it.

Here are my first impressions:

- This instrument really has a sound of its own. Whatever you sample in it will come out sounding, well... like a Fairlight. All sounds have a breathy quality to them. When you transpose a sound on a conventional sampler, the sampler will add or remove samples from the original sound to change the pitch of it. Not the Fairlight, it is one of the rarer samplers that alters its sample clock and always plays all samples of a sound. Just like when you change the speed of a tape recorder. Transposing the sound down will give it lots of bass.

Here's an example of a sound sampled in the IIx:
This is the original sound. A singing indian female voice (Wav 152KB)
This is the sound sampled and looped in the IIx at a sample frequency of 27600Hz (Wav 180KB)
The sampled sound played on the IIx through a Dimension C pedal (Wav 5.1MB)

- Each voice (there are 8 of them) has a whopping 16KB of sample RAM and an analogue filter. You might think that 16KB isn't much but it's actually quite sufficient. Just don't expect to sample long loops in it.

- There is no ADSR for the VCA, only Attack and 2 different release settings called damping. You can choose any of the 2 damping values by assigning a switch to them. There's also a single LFO that's routed to the pitch.

- The analogue filter is only used as an anti aliasing filter. There is no resonance setting or an ADSR to control it. What you can do however is assign key velocity or one of the sliders to the filter.

- The filter has a fixed key tracking value. This means the higher you play on the keyboard the more the filter will open up.

Here's is our sample played on the IIx with the anti aliasing filter wide open. (Wav 1.3MB)

- If you retrigger a voice on the same key the sound will cut off as if you were playing a monophonic instrument. Giving you the ni-ni-ni-nineteen effect. However when you play chords each note will get assigned to a different voice as you would expect from a polyphonic synth/sampler.

- There are 2 ways of creating a sound. Mode 1 and Mode 4. I don't know what happened to Mode 2 & 3 but I'm sure the Fairlight guys had a good reason to call the 2 modes 1 & 4. Mode 4 is sampling as we know it (Jim). Mode 1 lets you use the sample memory as a wave table with additive synthesis. I had high expectations of this but in the end it sounds disappointing. You can convert samples to Mode 1 but the results are pretty useless. It sounds actually the same as when you load a sample in Sounddiver and send it as a wavetable to a Waldorf Microwave. As long as the sound is a continuos tone like in our first example the results will be satisfying. Here is our sample converted to Mode 1 (Wav 172KB).

- It doesn't have sample accurate looping. You have in fact 128 segments that each contain 128 samples. You can only select a segment as a starting point and a segment as an end points for a loop. So, if you want to make good loops your sample frequency needs to be in relationship to the frequency of the sound you sample. That's why Page D (Display) is so useful. This allows you to view the FFT waveform of your sample and judge if your sample frequency was correct. A good sample, like the one I used as an example, will show you an even waveform as can be seen in the picture on the right (click to enlarge).

That's all for now. More to come in the following days.

Labels: , ,

3 Comments:

Blogger Javier said...

Nice report! The K2000 also use that transposition method...

7:42 am  
Anonymous Jesper said...

K2000...are you sure? I tought that the K2000 uses some "advanced" interpolate algoritms to prevent the aliasing when transposing, or im i wrong?

As David has demonstrated - The Fairlight using "drop sampling", no interpolate algoritms at all. The fairlight use an analog filter to "tame/reduce" the aliasing (the nice thing is that you can turn this of)

10:33 pm  
Anonymous Anonymous said...

The Kurzweil K2000 used smooth interpolation at a fixed sample rate. The K250 used variable sample rate transposition.

12:14 am  

Post a Comment

Links to this post:

Create a Link

<< Home