n this article, I shall attempt to shed light on an often misunderstood concept: loudness.
Let us begin from the very basics. Three folks are playing a tenor sax, a bass and a drumkit in a room. If someone holds an SPL (Sound Pressure Level) meter in that room, a reading in dB SPL will be obtained. This is the measured loudness of the performance in that room.
Now, let us take a recording of that performance, take off the lab coat, pour a glass of Cabernet Sauvignon and enjoy life a little bit in the comfort of our own living room, a different space at a different time. Ah, the miracle of sound recording and reproduction.
But how loud is it now? Where did I put that SPL meter? Given that most domestic listening systems will offer some means of arbitrarily adjusting the listening level (commonly the volume control knob), the loudness of the reproduced sound in our living room may or may not bear any resemblance to the loudness of the original performance. The reproduction can therefore be scaled up or down, depending on our mood, preferences, and neighbors.
In the process of getting to the recording and reproduction of a performance, there are several important relationships between the various representations of amplitude of the original acoustic sound waves.
We usually start by sneaking some microphones into the room. These convert the air pressure variations (sound in the room created by the musicians) into electrical signals. This way, dB SPL becomes volts.
These volts can be used to drive an amplifier and loudspeakers, converting the volts back into dB SPL (sound). This is what happens in public address and sound reinforcement systems. But to achieve the recording of sound in a manner which would allow its reproduction at a later time, we need to convert the electrical signals generated by the microphones into something that can be stored on a practical storage medium.
In the case of a vinyl record, the storage medium is a flat disk onto which modulated grooves are cut. As such, the vinyl records in your collection do not contain air pressure variations and they do not contain electrical signals either. They only contain a spiral groove on each side. When rotating and reproduced, the phono cartridge converts the stylus motion back into an electrical signal, which can then be eventually converted by the loudspeakers back into sound.
The electrical signal coming from the playback cartridge is too weak to move the loudspeaker cones directly. It is usually first amplified from the millivolt range into the volt range by means of an audio system’s phono stage. But the volt-range signal is still not adequate to really drive the loudspeakers. This is because the loudspeakers need power to do the work it takes to produce sound. The volt-range electrical signals running through your interconnect cables are still only in the milliwatt range in terms of power. The power amplifier is tasked with converting these milliwatts to several watts of power to be able to drive the loudspeakers.
So how loud will those watts sound? This depends on a huge number of factors, including the sensitivity of your loudspeakers, given as dB SPL per watt measured at one meter from the loudspeaker (dB/W/m), in anechoic conditions. In other words, not in the typical conditions encountered in your living room, unless you’re seriously weird (honey, I’ve just built an anechoic living room for the family!). As such, the actual room acoustics play a very important role in the sound pressure level we will end up with in the living room for a given amount of electrical power.
But let us go back to records.
When cutting a record, the electrical signal (volts) get amplified into watts (power) by a cutting amplifier which drives a cutter head. The cutter head responds by moving its cutting stylus, modulating the groove as it is being cut. The volts become watts and the watts become stylus velocity. This is why reference levels on test records are given in cm/s (velocity). You could even convert it to miles per hour, but the numbers would not be as convenient. Somewhat confusingly, when the record stops spinning and becomes an inanimate object on a shelf, there can be no velocity, since nothing moves. In storage, what remains is groove excursion. But since inanimate records on shelves do not produce sound (phew..!), it is the velocity which becomes the most important component of motion. Loudness is directly proportional to velocity.
Neumann VMS-70 disc cutting lathe at SAE Mastering, Phoenix, Arizona. Courtesy of Wikipedia/VACANT FEVER.
Upon reproduction, the higher the playback stylus velocity, the higher the voltage it generates. This relationship is often given as the sensitivity of the cartridge, in mV/cm/s, or more commonly, in mV for a 5 cm/s RMS velocity at 1 kHz.
But how loud will that sound? As loud as you set your volume control knob. Unlike some professional audio systems used in mastering and broadcasting facilities, most consumer listening systems are not calibrated end-to-end. There are no absolutes there; the level is only relative. As in, adjust to taste.
So, how loud is one record compared to another record?
This is where things get complicated. Apparent loudness is often achieved at the expense of dynamics. By intentionally restricting (compressing/limiting) the dynamics, the average loudness can be pushed higher. But the transient impact lives in these dynamics. So the instantaneous (transient) loudness goes down as the average loudness goes up, within our available dynamic range. It is called apparent loudness because, for example, if you compare two records at the same volume control setting, the one that appears to be louder will be the one with the highest average velocity, not the one with the highest peak velocity. Yet, it is the one with the highest peak velocity that uses up more of the available dynamic range of the medium.
In fact, if we were to calculate the theoretical dynamic range potential of the storage medium alone, we would end up with a mindblowingly high number, which could never be achieved in practice. The same holds true with any theoretical dynamic range calculation for any medium. But unlike other formats, the vinyl record does not actually have any hard ceiling of maximum loudness. Which is why mastering records is so complicated and why it can make such a huge difference in the sound of the final product, depending on how it was done.
Since there is no clearly defined hard upper limit, designing a phono stage that can really cope with the full practically-achievable dynamic potential of the medium becomes extremely challenging.
This is no longer about loudness, but about impact, transients and dynamics. This is one of the biggest secrets to realistic sound reproduction. The commonly achievable dynamic range of the vinyl record already far exceeds what can be heard in a typical domestic listening room and even challenges professional studio systems. We could even cut a wider dynamic range than that on a record, but very few playback cartridges would be able to reproduce it.
VU meters on a Teac A-3340S four-track tape recorder.For decades, as a direct consequence of the loudness wars, a lot of technological development in disk mastering systems was concerned with increasing the apparent loudness. Yet, the ability to capture the sharp transients (instantaneous loudness), which depends more on the skill of the engineer, was there all along! Which beautifully explains why many audiophile records, known for their excellent sound, were cut using very early, bare-basics, manually operated disk mastering lathes.
Vintage Neumann Lathe with a copper DMM blank on the platter. This was not a DMM lathe and was not able to cut copper blanks. This machine predates DMM technology by about 50 years. Image courtesy of Wikimedia Commons/32bitmaschine.
