Noise Reduction System
There are more sources of noise in a studio than recording tape -and a noise reduction unit ought to be able to cope with all of it, right? Vic Lennard thinks he's found one that does.
One of the most useful pieces of equipment in a small studio is one that can be put to many uses - when those uses involve the elimination of noise, the SNR1 is invaluable.
THE NAME OF dbx is synonymous with tape noise reduction, as anyone with a Tascam recorder will tell you. However, the use of compression on recording followed by expansion at playback is only suitable for reducing tape noise at the recording stage. If the program content is high in noise, especially at the top end, then this will show up on the final master, and while the use of noise gates will keep quiet passages under control, they are not program conscious.
The alternative approaches to this system lie firstly in downward expansion - which increase the gating effect as the program level decreases, so reducing noise when the program content is low - and secondly in dynamic filtering - which rolls off the top end once the input level has fallen below a certain threshold, so masking the high-frequency noise present throughout the audio input.
Units employing these operations do exist, the most well-known one being the Symetrix 511 which incorporates both. This still changes hands for £3OO-£4OO secondhand, while the newer version, the 511A (Reviewed MT, February '89), retails at £569. Consequently, there is room in the market for a cheaper model, and this has now been introduced by dbx in the form of the dbx SNR1.
THE SNR1 WAS originally intended to be a hi-fi add-on, hence the two sets of phono sockets on the rear. One of these is for interfacing to the tape output/monitor return of your amplifier, and the other for connection to a tape deck. This means that you can either use the noise reduction on recording or playback, and use the Pre/Post buttons on the front to determine whether you are effecting the signal pre-tape deck or post-tape deck. A bypass button allows you to hear the signal on input to the SNR1 or after noise reduction has been applied, and two toggle buttons for Tape and Source make the most of any recorder with three heads for monitoring off tape.
A Threshold rotary control sets the level at which dynamic filtering commences, and ranges from -7OdBV to -3OdBV. This, along with the permissible input voltage, makes this unit suitable for the semi-professional market with a system level of -1OdBm.
The manner in which the unit is set up is of paramount importance if good results are to be obtained with it, especially from an input level point of view. The Threshold control on the SNR1 dictates when the noise reduction comes into effect and is related to the high frequency content of the incoming signal. To monitor this input, there is a nine-segment level indicator with a central green LED sandwiched between four each of yellow and red. Each LED represents 10dB of signal level and if the red LEDs light, then the level is above the threshold and no filtering will occur. It would appear that, as the unit is stereo, it responds to the higher level shown on either the right or left input channel.
The other meter on the front panel shows the frequency range at which filtering is taking place. It does this with the help of seven red LEDs representing from 800Hz up to 20kHz. These LEDs light up from the upper frequency downwards.
FROM THE ABOVE, it is clear that the Threshold setting and the high frequency content of the incoming program are the factors which govern the cutoff frequency of the SNR1. The lower the threshold, the more difficult it is to induce noise reduction - for example, if the Threshold is set to the maximum of -3OdBV, any signal below that level will reduce the cutoff frequency from its maximum of 20kHz down to a value directly proportional to the amount by which the input is less than the threshold.
Consequently, there are two different ways to use the unit: set the Threshold low enough to leave the majority of the signal unaffected, so that the filtering occurs in areas where the program level drops, or ensure that the noise reduction is adequate at the quietest point of the music. The former method may mean that the degree of noise reduction is insufficient, while the latter method may well entail a degree of filtering is taking place throughout. Your choice of which approach to adopt should be dictated by the noise content of the audio program, although it must be said that poor engineering will inevitably lead to poor results, no matter what equipment is used subsequently.
"In a moment of despair I patched the SNR1 into a noisy Roland Super JX and sat back to listen - I have to admit to being impressed with the result."
The review SNR1 was first tested on some rather noisy pre-recorded cassettes to see how much of the inherent copying noise could be extracted. This was attempted using both of the above techniques. Setting a high Threshold certainly caused audible rolling-off of the high frequencies while a low threshold didn't really handle hiss in the quietest parts. Removing dolby B - which these cassettes were supposed to be encoded with - and so increasing the inherent high-frequency component of the music gave totally different results. Happily, the SNR1 seemed to be more at home. I would have expected to hear a degree of "pumping", but found it to be minimal and very much subject to the Threshold setting. Now I'm not suggesting for a moment that you turn off dolby noise reduction when using the SNR1, but I have found that the quality of the "chrome" cassettes sold in the pre-recorded market often leaves much to be desired, and the ferric ones are quite atrocious in many cases.
The next situation the review SNR1 found itself in was during a mixdown with a Roland MKS70 Super JX. The MKS70 has a great sound, but a noisy output (which has been the bane of my studio for the past year or so). In a moment of despair I patched the SNR1 into the module, set quite a low Threshold value and sat back to listen. The sound in use was one with a lot of high harmonic content and I have to admit to being impressed with the result.
Continuing this line of investigation, the SNR1 was set up to be compared with the Symetrix 511 in masking the output noise from an Axxeman guitar processor. In Clean mode, the SNR1 scored over the Symetrix unit in hiding the high-frequency hiss, because its 800Hz frequency cutoff point is lower than that of the Symetrix. However, in Lead mode (we're talking distortion here, for anyone not familiar with widdly-widdly terminology) the downward expander on the 511 was necessary because the noise exists at a far lower frequency and a dynamic filter cannot handle this on its own.
How about using the SNR1 across the outputs of the mixing desk on mixdown - lots of top end there? The results obtained were pretty good except during a controlled fade out, where the lack of high frequencies was particularly apparent. Pressing the Bypass button gave an audible click, but rotating the Threshold control to minimum didn't produce any unwanted audio side-effects and defeated the SNR1. Inherent noise from the unit itself was nil and I could not detect any shifting of the stereo image.
Having got me into the mood for eliminating noise from problem sound sources, the review SNR1 found itself in the company of a Yamaha SPX90 - which has to be one of the noisiest digital effects units of all time. The pitch-change programs in particular often demand the top end be boosted on the effects return on the mixing desk, yet the signal noise makes this rather unsavoury. The SNR1 provided a good answer as the top-end boost on the desk no longer accentuated the SPX90's high end output noise.
Another area of experimentation the SNR1 was subjected to involved samplers. I sampled a voice from a Fostex E16 into an Akai S950 with the intention of repeating it further on in a song with a sequencer. I first tried sampling through the SNR1 to remove any tape hiss but found the noise generated by the S950 to be greater than that generated by the E16. So, having sampled the voice straight from tape, I then replayed it through the dynamic filter. This vocal section had a couple of quiet patches with a little headphone breakthrough, and the SNR1 was most useful in hiding this. I also tried gating the section, but found this to be less natural. In this context, the SNR1 was found to have a very fast reaction time, which is a good reason for having limited the front panel controls to threshold only.
Finally, the overworked review SNR1 was interfaced between the mixer outputs and the input of an aural exciter. These tend to dislike a lot of high-frequency noise and consequently function in a less than satisfactory manner unless this type of noise is removed. In the time I spent using the SNR1, I found that the aural exciter lost the harshness which had previously stopped me using it across a full mix.
AS YOU MAY have gathered from the review, I was more than impressed with the SNR1. Performance-wise, it is very difficult to knock it in any respect except that it's a shame that it can't be used as two separate mono units - but this would raise the retail price of £279 to unreasonable heights. This aside, the SNR1 is one unit which can provide many of the tools required for reducing the noise and hence improving the quality of recordings made in a small studio.
Price £279.90 including VAT
Review by Vic Lennard
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!