Home -> Magazines -> Issues -> Articles in this issue -> View
Symetrix 511 Stereo Noise Reduction | |
Article from Home & Studio Recording, October 1984 |
User Report
A single-ended noise reduction unit that provides an alternative to conventional encode/decode systems. Combining the techniques of expansion and dynamic noise filtering, the Symetrix Model 511 can be used to cleanup existing recordings. Studio owner Dave Simpson files his User Report.
In the last decade or so two major systems have been developed to combat this noise problem; 'Dolby', the brainchild of Dr Ray Dolby, and 'dbx' the invention of David Blackner. Although both are very effective, for the home recording enthusiast they have two serious drawbacks.
Firstly, they are expensive, especially Dolby A, the professional system in the Dolby stable. Admittedly Dolby B, the cheaper model is available, but I have always found this to be of fairly limited value, acting as a high filter in some cases. Dolby C, the latest development goes some way to bridging the gap, achieving excellent results on machines like the Fostex B16 but is not, as far as I am aware, available in a rack-mounting unit suitable for putting in line with a mastering machine of your choice. Dbx have such units and they are very reasonably priced. Even those though, suffer from the main disadvantage of the Dolby and dbx units; they are both encode/decode systems.
What this basically means is that the signal is processed, or encoded going on to tape, and needs to be decoded before it can be properly listened to. If a stereo master is encoded, then it can only be replayed on a machine with compatible equipment. At the professional end of the market this is fine, but the home recording enthusiast is unlikely to have access to this type of set-up. In the case of a Portastudio-based studio, where bouncing is the order of the day, noise reduction on the final mix is sometimes very necessary. It can be a pain if you then have to do a further transfer onto another cassette without noise reduction just so it can be played on any tape recorder.
Enter the Symetrix 511 noise reduction system. Unlike Dolby or dbx, it is single-ended, not requiring a decoding unit. This not only means that any tape it processes can be played on any other tape recorder without ill-effect, but it can also be used to process a previously noisy tape; for instance a hissy cassette. (Dolby and dbx only reduce tape noise as a signal is being recorded.)
The 511 utilises two basic circuit elements; a dynamic range expander and a dynamic low pass filter. The downward expander detects and is controlled by changes in input signal amplitude. Input signals which are below the level established by the control on the front panel, are subject to downward expansion, which means that for every dB the input signal falls below threshold, the output signal will fall by a greater amount. Thus the lowest level signals (which are usually noise) will be attenuated.
The dynamic filter conversly detects and is controlled by changes in input signal high frequency content. It is of the low pass type with a filter roll-off of 6dB per octave. It works to reduce noise by attenuating high frequency response in the absence of high frequency input signals. When high frequencies appear at the input the filter opens and passes these frequencies. At this point the noise is still actually present, but is masked by the high frequencies. With no input signal the filter bandwidth is around 1 kHz. As high frequency content is introduced the bandwidth increases up to 30kHz, depending on the position of the front panel controls and the high frequency content of the input signal.
The front panel controls are very simple. The unit is stereo, which can if required function as two independent mono units, selectable by depressing a switch. Each channel has an in/out switch for the downward expander, and a similar switch for the filter, both circuits being controlled by a single rotary threshold control. In the stereo mode, both channels are controlled by the left-hand set of switches.
In addition, each channel is provided with six LEDs, two relating to the filter, and four relating to the expander, the filter LEDs being marked 4kHz and 12kHz. Since the bandwidth is continuously variable, these lights display the minimum value of the filter; when the actual filter bandwidth is greater than 12kHz, no LED will light. Between 4-12kHz, the 12kHz light will register, and below 4kHz the remaining LED will register. This system provides a rough visual guide to the chosen filter bandwidth.
The expander LEDs operate in a similar way, except that the calibrations are in decibels; 3dB, 4dB, 12dB and 18dB. The rear panel has both balanced XLR, and normal jack inputs and outputs.
Operation of the unit could not be easier; you simply adjust the threshold control until the hiss and noise becomes audible, whilst at the same time ensuring that the high frequencies pass through the filter. In practice this usually means waiting until the effect of the unit is audible (whilst adjusting the threshold control) and then backing off a short way until the filter no longer muffles the high frequencies. Sometimes it is preferable to use one of the circuits in isolation; for instance, a cymbal track which needs low frequencies attenuating but the high end left intact would require the expander circuit alone. It is a question of experimenting; there is no set formula for using the unit.
So much for the theory. The question is, does it work? Well, the short answer to that one is - yes! At times it was almost too effective, but more of that later. First I shall detail some of the possible applications for the unit, and make a few comments.
I found three main areas in which the 511 proved useful; individual tracks, bouncing, and stereo mixdown. The cleaning up of individual tracks as they were recorded on to tape depended largely on the type of single being processed. Bass guitar was perhaps the best candidate for noise reduction. Many passive basses have fairly low outputs, and since budget mixing desks tend towards noisy input stages, when the bass is boosted to a decent level it can be accompanied by some background noise, sometimes quite audible.
Noise gates are of fairly limited value in dealing with this kind of problem because they only highlight the noise 'halo' every time the bass is played. With the Symetrix, however, you could adjust the threshold control until the filter cut out much of the unwanted top end hiss, and the expander attenuated much of the rumble. In between notes the unit squeezed the noise levels and acted as almost a noise gate. This was helpful because I usually use some compression on the bass to tighten it up, and in the absence of any signal the compressor boosts the noise levels to sometimes quite startling levels (depending on the amount of compression).
Using the Symetrix on other instruments as they were recorded met with rather limited success. That is not to say that the unit did not work; rather that the amount of noise reduction (due mainly to the lack of noise) was not equal to the effort in setting up the patch. On the face of it, noisy polysynths seemed to be a likely candidate (and there are lots of noisy polys!), but I found a gate was more effective, perhaps because the noise was most audible in the breaks between the signals. The Symetrix did go some way to acting as a gate though, and if you did not possess a noise gate you could get by with only the 511.
It was with the second and third areas of application that the unit really came into its own though. Even with a sixteen track recorder, I tend to do a fair amount of bouncing; for instance several percussion tracks might be recorded and then bounced down to a stereo pair. During this process several things contribute to noise levels. There is the background noise from any track already on tape. There is any noise the mixing desk adds during the bounce stages, due either to input stages or EQ. There is the background noise of the tracks being bounced onto, and there might be any reverb noise used during the bounce itself. Altogether you might agree, a fairly noisy process!
Patching the Symetrix in between the desk outputs and the recorder inputs though, elicited an almost magical transformation. Monitoring off tape I observed no audible reverb noise, no audible track noise and no audible desk noise. Listening back to the bounced tracks the only noise I could hear was that of the two tracks recorded onto, which the Symetrix did not have to deal with. I must add though, that the tracks were fairly clean to begin with.
With noisier tracks, the Symetrix will still handle them, but you might have to adjust the threshold control so that the high pass filter rounds off a bit of the top end signal as well. You personally must decide whether this is preferable to noise. There is a way round this, however, and I will discuss it later but it costs money (doesn't it always?).
The third application is that of using the noise reduction unit between the master outputs on the desk and the master tape recorder. Basically the principle is the same as bouncing; you adjust the threshold control until you hear the top end being cut into by the filter, and then back off a bit. You are then left with the maximum amount of noise reduction available without compromising the signal.
It seems to be on the multiple signal applications that the Symetrix works best, and I am sure that it was for this purpose that it was designed. In a situation involving a wide range of frequencies, some of which appear at infrequent intervals (such as cymbal crashes), the unit acts as a continuous hand on a filter, hovering at all times just above the frequency of the highest signal. The advantage of this is particularly noticeable where, for instance, the full band stops, leaving the drums playing alone. Even the fractional gaps in between the drum beats are clean and noise free. This applies to the spaces before and after tracks as well; it can save splicing if you are in a hurry!
Now for some points to watch out for. The first one surprised me, but when I considered the signal chain it seems entirely logical. My studio patchbay is wired so that the signal comes out of the right and left outputs on the mixing desk, is passed through any signal processors such as compressor or noise reduction (or both), and then proceeds direct to the master ½ track.
What happens though, is that if you operate the master faders at the end of a track, the expander on the noise reduction unit starts cutting in. When a certain point is reached in the fade, the unit assumes that the signal is so low that it is noise, and cuts it out entirely. This can do two things; it can result in a linear fade up to a certain point, and then a sharp drop as the expander cuts in. Alternatively, it can cut off a reverb tail (which is usually fairly low in volume compared to the original signal) leaving a dry finish, which in some cases might not be acceptable.
This side-effect depends, of course, upon the level of noise reduction selected. I find though, that I usually end up using quite a large amount (for a reason which will be made clear in just a moment). There are two possible answers to the problem (three if you include buying a mixing desk with insert points before the master fader).
The first is to fade the track on the ¼" tape recorder. This I found effective, but on the Revox PR99 which I use, there is no way of ganging both input level controls together to produce a smooth fade simultaneously on both channels. What I ended up doing, was to use a pair of noise gates in the chain after the noise reduction, and to use the fade control on those. (Luckily my gates have a fade time of up to 30 seconds.) Thus I can get a perfectly linear fade, and since it follows the Symetrix in the chain, the expander does not detect it. In the case of a reverb tail following a sudden stop, I simply set the gates to follow the fade of the reverb.
I mentioned before that to achieve really effective noise reduction over a noisy tape, you have to set the filter on the unit so it just encroaches into the highest frequencies of the signal. I also added that I tend to do this as a matter of course. How do I get away with it?
This isn't going to be much help to many of you I'm afraid; I put an Aphex Aural Exciter after the Symetrix in the chain (but before the noise gates). This has the effect of synthesising a whole new top end from the remaining frequencies, creating a sparkling noise-free track.
Having said this, however, the Aphex is by no means necessary for the vast majority of home studios; I run a professional studio setup and am in the business of producing master recordings. The same quality would not be required of a budget set-up. I notice, however, that Vesta-Fire have produced a device they call a Driving Exciter, for about half the cost of the Aphex, which appears to do the same job. Might be worth investigating?
So much for the tips of using the Symetrix. Now for the niggles (yes - they had to come!). Only small ones though.
Firstly, I should have liked to see a simple on-off switch on the front panel of the unit (indeed, anywhere on the unit). Since all my rack units are powered from the same source, to switch off the 511 I have to switch off the whole rack. This is annoying, as it means that even if I know in advance that I will not have to use it during a session, those six LEDs remain alight and unwinking at me. The MXR digital reverb I have is the same; why are manufacturers starting to do this? Surely not to save money!
Secondly, it would have been helpful to see a couple more LEDs relating to the filter. The 4kHz and 12kHz ones were helpful, but certainly a 16kHz and possibly an 8kHz LED would have made it easier to see where the unit was cutting in. I take the point though, that the LEDs are there as a guide only; your ears should be the final judge. It is just that if you are listening on a pair of Auratones, for example, you might cut into the higher frequencies with the filter and not realise it.
The last niggle concerns the lack of a separate threshold control for both the expander and the filter. Although each is switchable in or out, when both are in, they are controlled by the same knob. I have found certain applications in which, say, a heavy amount of expansion is required with rather less filtering, and vice versa (the fade problem for instance). Perhaps a double pot, with a smaller knob sitting proud of a larger one yet still being able to be ganged together, as found on some hi-fi equipment, would be the answer.
The Symetrix 511 noise reduction unit at £499 is expensive. There is no doubt of that. It is also very good. I personally have no doubt of that. It all comes down to priorities. It is not the kind of unit you buy before you get a reverb unit. Rather (and this may seem obvious) you buy a unit like this when noise on your recordings starts to become a problem to you.
For some, this point may never be reached. For others, however, the more you become interested in recording, the more critical you get as to the final product. Suffice it to say that this unit does everything it is supposed to, and does it well. It is the highest quality single-ended unit I have used, which makes it useful in cleaning up old noisy tapes and is so versatile, it can even be used to quieten noisy reverb outputs. If you have reached the point where a unit of this type is a requirement, and can afford the price tag, then I would suggest you buy it!
Symetrix 511 £499 inc VAT.
Details from Atlantex Ltd, (Contact Details).
Review by Dave Simpson
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!
New issues that have been donated or scanned for us this month.
All donations and support are gratefully appreciated - thank you.
Do you have any of these magazine issues?
If so, and you can donate, lend or scan them to help complete our archive, please get in touch via the Contribute page - thanks!