|Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting|
Author: Wayne Saylor [Audio Engineer, Memorex Corp., Santa Clara, Calif.]
High performance tapes being introduced in today's audiophile market are characterized by such features as low distortion, high undistorted output, more output at high frequencies before saturation, more dynamic range, and more headroom.
This discussion deals with how to take advantage of those benefits; to present an understanding of how to recognize and control distortion inherent in the tape recording process so that tape recordings truly have higher fidelity.
Strictly speaking, distortion is the degree to which reproduced sound differs from original sound, and this definition, lack of fidelity, holds true regardless of whether the distortion is caused by equipment or by recording medium. Distortion includes amplifier overload and hum as well as tape "hiss" and any change in frequency response between input and output of the tape recorder.
The scope of this discussion will, however, focus on distortion as related to magnetic tape and how to optimize the record and playback levels to obtain minimum distortion. The particular types of distortion to be discussed are odd harmonics, compression and saturation, and inter modulation distortion.
These types of distortion are all forms of nonlinear distortion and occur when the output and input of a recorder depart from an ideal straight line relationship. The function of a tape machine's bias circuitry is to linearize the output -input relationship which otherwise would be highly non-linear (owing to magnetic hysteresis). Beyond having a unit's bias correctly set (usually done in the factory or service department), there are other things which must be done to record with minimum distortion, and these aspects of recording will be discussed here.
Odd Harmonics of the Input Signal
Harmonics can be described, generally, as an addition to the fundamental tone of spurious tones having frequencies that are whole number multiples of the fundamental. Thus, third harmonic distortion is a combination of the fundamental frequency and a second frequency which is three times the fundamental. The percentage level of a harmonic almost always decreases with each higher harmonic order; the fifth is much lower than the third, etc. Consequently, only the third harmonic is generally measured in the case of odd harmonics distortion, because it usually has a substantially higher percentage level.
Since all musical instruments have rather significant amounts of their sound energy in harmonics, harmonic distortion is not, in itself, unpleasant to listen to. Third harmonic distortion is very convenient to measure, however, and is used to quantify distortion more because of its convenience and correlation with other types of distortion. Of course, third harmonic distortion, inherent in tape recording is only perceptible or measurable using fundamentals up to 1/3 of upper frequency limit of the recorder. The laboratory method of measuring third harmonic distortion involves simply recording a pure tone (single frequency) and measuring the relative level of the third harmonic of that tone.
Even harmonics are not generally characteristic of magnetic tape systems because the hysteresis curve is symmetrical, even though it is nonlinear, and asymmetry is required to generate even harmonics. However, even harmonics can occur in the recording system and are caused by any nonsymmetrical polarizing magnetization. Such even harmonics are, however, equipment related, rather than tape related, and are usually caused by either magnetized heads, a d.c. current component in the heads, one-sided amplifier overload, or asymmetrical distortion of the bias signal.
Compression and Saturation
Compression of the recorded signal occurs when a given amount of increase in input level does not result in as much increase in output level.
Compression occurs at all frequencies, but is much more prevalent at high frequencies. Obviously, this alters frequency response. This happens principally at high input levels due to the inability of the oxide coating to be magnetized further and a phenomenon known as self -erasure occurring predominately at higher frequencies.
At the level compression begins, that is where no increase in output level results from increasing the input level, saturation has been reached.
Beyond saturation, further increases in input record level result in decreases in output. Use of recording input levels beyond saturation alters frequency response much more adversely than compression prior to saturation. Compression and saturation result in non -true level, rather than the introduction of new or different frequencies.
Intermodulation or IM distortion occurs throughout the spectrum of frequencies which the recorder is capable of reproducing and is the generation of new frequencies not harmonically related to the input frequency. This form of distortion results from the combination of two or more input frequencies at the same time; the output contains not only the original two frequencies, but also the sum and the difference of the original two frequencies. These are usually referred to as beats of the original frequencies. IM Distortion has no musical relationship to the original frequencies from which the distortion is derived, so IM Distortion causes the greatest subjective impairment to the music. The words harsh, dissonant, and "fuzzy" are used to describe the unpleasant subjective effects of IM distortion.
The main idea in making superb quality recordings is to use the highest record level possible without objectionable distortion. The higher the Record level, the more distortion, but a lower Record level results in less difference between the inherent tape noise and the level of music during Playback. Tape hiss gets louder as Playback level is increased, but does not get louder as Record level is increased. As Record level is increased, the recorded signal is increased, so better signal-to-noise ratio is achieved by recording at the highest level possible.
Since today's superior performance tapes achieve their performance through high record level and output capability, it is necessary to know how to record at high levels and still control distortion in order to utilize the full performance capability of the tape.
Tape noise or tape "hiss" is present in all recording tape and stems from such factors as the size of the oxide particles, irregular distribution or dispersion within the coating, surface roughness, and poor orientation of the particles. Within basic categories of tape (such as high performance audiophile tape), the noise is usually quite similar. The predominant factor in determining the relative signal-to-noise ratio of competitive tapes is maximum output signal level rather than how low is the noise.
Optimum Recording Procedure
Optimum record level settings can be determined by the recorder user by utilizing some of the tape recorder's features to conduct an A-B comparison test as commonly used by audio engineers and other recording professionals. The easiest way to conduct an A-B test is to use a three -head tape machine with a Source-Tape switch. Most open -reel recorders sold today have separate erase, record and playback heads. Using the operating instructions for the tape deck, connect the recorder to a sound source such as a phonograph record player.
While in Record mode you can then switch repeatedly from source to tape, back to source, etc. In this way, you can compare the source and the tape playback. This comparison should include not only careful listening, but observing and making note of the level indicator reading on the loudest passages. There will be a slight delay in music between the Source and Tape, but this fraction of a second does not detract significantly from the A -B comparison. The delay is dependent upon tape speed and the distance between the record and play heads. Note that in the Record mode on a three -head machine, the level indicator shows Record level in Source and Playback level in Tape. For each A-B test, the playback level should be adjusted to match the loudness when listening to Tape to the loudness in Source. This is important because a difference in loudness may mask a difference in distortion.
If your machine has only two heads (erase and record/play), you can still conduct an A -B test, but it is more difficult and time consuming. The difficulty stems from the fact that you can't listen to tape playback during the same pass as record, necessitating rewinding and playing later. Such a machine doesn't need a Source-Tape switch since the level indicator reads Source during the record pass and Tape during the playback pass. To A-B compare the source and tape in this case, the phonograph record source and tape recorder playback should be connected to their respective inputs of the stereo system preamplifier. The A-B comparison is then made by switching the system pre -amp between Phono and Tape. It is, however, necessary to closely synchronize the phonograph and the tape during playback.
Once your tape machine is set up in the A -B mode previously described, you are ready to experiment. Initially adjust your recording level so that the level indicator consistently reads far into the red zone so that the recording is purposely very distorted. Match playback levels and switch between the original source and the now distorted recording. Listen to the distorted recording and compare it with the original source. Now that you know how distortion sounds, start reducing the recording level (compensating for any drop in volume by increasing the playback level) until you reach a point where you no longer notice any distortion in the recorded program relative to the original program. By observing the level indicators at this point, you have identified the point on your meters where distortion becomes unnoticeable. Signals swinging further into the red zone on your meter will give noticeable distortion.
Ideally, you should adjust your record level to as high a level as possible without inducing distortion. And you should recognize that recordings from different program sources made at the same recording level setting may differ in distortion because they differ in dynamic range and average level. Thus, we should repeat the above experiment for various types of music. Remember that with some tape machines, you may be able to operate well in the red with no noticeable distortion. With others, you may have to keep the indication below the red even for sudden peaks.
The single most valuable tool you have in controlling distortion is the level indicator. The recordist must learn from the equipment manufacturer and from his own experience how to use his particular indicator to control distortion.
There are great differences among level indicators of which there are several types. The most well known is the VU meter which averages the signal, ignoring sudden peaks. There are also bar -graph indicators, LEDs, and peak reading meters. One type of peak reading meter measures "average" loudness much the same as VU meters except that it has a fast-attack, slow-decay ballistic response so it responds to shorter transient bursts of sound. It measures the flat response input signal when metering source.
The other type of peak metering has the same peak response but meters the record signal in Source AFTER record equalization so that the actual level applied to tape is indicated.
(High frequencies are recorded higher relative to low and mid -frequencies in what is known as Record equalization or Record pre -emphasis.) With some tapes you can record at very high levels, but with some you can't. Thus, you should determine what sounds best to you, and then operate on that basis. This test will also allow you to explore different brands of open -reel tape, giving you the opportunity to determine if one brand offers better performance on your machines than another. One brand of tape may have more "headroom" than another, and this might allow the level indicator to read further into the red zone. On the other hand, if the tape also had greater sensitivity, it would give a higher undistorted output for the same reading on the level indicator. Remember that high undistorted output is the key requirement in obtaining a stronger, more dynamic recording with less hiss.
While you can, of course, play it safe and always avoid distortion by recording signals at a low level, you would be recording too close to the noise floor, so you would notice more "hiss" than the recording would have if the Record level were set to its highest "distortion -free" position.
Compression is most noticeable in the high frequency end of the spectrum. Therefore, certain percussion instruments are best for observing where compression takes place. Cymbals and some tambourine sounds are good examples of this. An example of records very good for experiencing what compression really is and how to control it is I've Got The Music In Me on Sheffield records, Rough Trade-Live on Umbrella (through Audio-Technica cartridge dealers), or Direct Disco by Crystal Clear Records.
These records are particularly clear, free of distortion, and have great high level high frequencies. A record that has great dynamic range from very soft passages to very loud ones is Saint Saens Organ Symphony in C Minor, Phila. Orchestra, Columbia MS 6469.
For A -B comparison, however, the most useful type of record music is one with lots of high frequencies at high level on a sustained or repetitive basis. Other than that, you should choose music of the kind you like most and listen to most frequently.
Ways to Control Distortion
Although wise use of your level meter and experimentation is a primary way of controlling distortion, there are other techniques to be considered. They fall into two general categories, machine adjustments and tape selection.
As mentioned earlier, the sole purpose of bias in audio tape recording is to promote a linear relationship between reproduced output and recording input. Ideally, this should be done for each different type of tape to minimize distortion. Too little bias causes non-linear distortion and increases high frequency output. (This is due to bias field strength gradient throughout the oxide coating thickness, the strength being higher nearer the surface where high frequencies are recorded.) Too much bias effects the frequency response, decreasing the high frequencies more than the lows.
Very few tapes decks provide a convenient means of bias adjustment by the user. Unless your deck provides such a means, it is unwise to tamper with bias if you are not equipped with proper external test equipment to measure the degree of adjustment. If you purchase a new machine or change brands of tape, you might find a worthwhile improvement by having a service agent re-bias your machine.
Many new machines are not properly biased for the brand of tape being used. Some machines have a switch with more than one fixed bias setting, labeled Standard/High, Normal /LN, etc. You should use the position recommended by the recorder manufacturer or tape manufacturer.
The higher the tape speed, the easier it is to get good signal output versus noise, without getting into distortion.
The slower speeds place the most demands upon the tape performance, and it is at these speeds where the differences among tapes are the most apparent.
Dirt on the record head separates the tape from intimate contact with the head and causes distortion because the separation effectively lowers the bias as well as the record signal. A dirty playback head can cause distortion mainly through recordist confusion if you erroneously increase record drive to make up for lower playback level caused by head to tape separation.
Magnetized heads cause even harmonic distortion and increased noise level, so periodic degaussing is recommended.
For best results you must select the ape which will provide the greatest performance parameters with minimal distortion. Even among premium tapes, it is important to try out various brands and experiment with record levels in order to find out which offers the best performance on your tape deck. Generally speaking the important characteristics to consider include: Distortion. If your tape has relatively low distortion, you will be able to record higher levels while still maintaining high quality reproduction.
Signal-to-noise ratio. The higher you can record above the tape noise, the less "hiss" you will hear even on soft, quiet music since the output level setting can be lower relative to the record level setting.
Saturation. A high output before onset of saturation results in a greater dynamic range and, most important, provides assurance against over recording high frequencies to the point where higher input results in lowered output.
Sensitivity. A tape with higher sensitivity is one which has higher output than another tape for the same record level. While the M.O.L. (Maximum Output Level, usually quoted relative to a certain percentage distortion) is the more important parameter, sensitivity is ordinarily a reliable indicator of relative M.O.L. among tapes of similar bias requirement. A more sensitive tape usually has higher M.O.L., especially at low or mid -frequencies.
To check this out yourself, record some music which has a sustained level from a phonograph record at about -5 to -10 on your level meter. With the same record level settings, check which brand of tape plays back at the highest level. You may find 2 or 3 dB difference between major premium brands. More sensitive tape usually has lower relative distortion as well.
This lets you capture all signals at a higher level relative to the noise.
In conclusion, it should be evident from this discussion that any recording enthusiast can control distortion if, first, he has a good basic understanding of the mechanics and limitations of his tape recorder and its meters; and secondly, if he understands the performance characteristics of the tape and how it interacts with the recorder. Both conditions necessitate a certain amount of recording experience and experimentation, but once this is acquired you have the tools you need to control distortion.
(Source: Audio magazine, Apr. 1977)
= = = =
Prev. | Next