Bandwidth needed for flat frequency responce from vocal chords of singer to listeners brain and Pleiades amplifiers

Using a good quality microphone produces a very small flat bandwidth to listeners brain.

The frequency responce curve perceived by ear - brain is one that starts increasing below 1KHz and increasing above 5-10 KHz. (See the next after next euroelectron post).

This is a bathtub curve.

Bandwidth defined as the distance in frequency between the 3dB frequency points is very low although the microphone can reproduce from 50Hz to 18KHz for example.

Flat bandwidth may be even less than 8KHz. The fact that it is an inverse bathtub curve makes matters even worse.

As of writing this the FM radio plays a Toscanini 30s recording. It sounds with emphasis on mid range. 60s recordings sound natural in general. And some modern recordings may sound bathtub like possibly because we have gone to extreme surpassing the flat correct frequency responce need by our brain perception.

There are many factors that explain this. This article is for voice. Voice and few single instruments can sound great on 30s recordings (Rachmaninoff chime self being recorded on the piano for example with cello). This may have to do with Fletcher Manson curves. These effects are important when there is difference between reproduced and recorded acoustic intensity. When recording a full orchestra the reproduced intensity is lower that the original. This gives and inverse bathtub curve on our brain perception.

But let's start with proximity effect first. To see half of a bathtub curve at frequencies lower than 1KHz one can refer to any frequency responce of a directional microphone at small distances from source. Of course we can listen to this by our ears as a muddy and heavy mid, mid bass, bassy heavy responce. Or a heavy responce with no high frequencies as they are dominated and buried by the bass predominance.

Also the pressure doubling effect at high frequency on pressure microphones increases the high frequency content unnaturally. They are best to be used at 45 to 90 degrees, grazing incidence.

As starting to explain above there are other psychoacoustic factors too that exaggerate the bathtub curve. Fletcher Munson curves and voice effort curves explain this.

So an added bathtub frequency responce is added by our brain itself.

This is because the original recorded acoustic intensity, (human voice for example) and the reproduced acoustic intensity (a voice matching a full orchestra) do not match. The reproduced acoustic voice is much louder.

So listening to something louder that it in reality creates another bathtub curve exolained by the Fletcher Munson curves.

The simplest perhaps way to describe the Fletcher Munson curves may be the following. The more acoustic sound intensity arrives to our ears the more more bass and treble heavy the sound is to our brain. And the corollary, the smaller the acoustic intensity the more the brain focuses on mid range. A reason may be survival itself.

An example. A crying babe (see also voice effort curves) can be heared in a dessert at great distance and be rescued.

Now on to the voice effort curves.

Softly speaking which is nice and expressive creates another bathtub curve explained by the voice spectrum effort curves. When we speak or sing loudly, the spectrum peaks at mid range, like a pyramid (inverse bathtub curve). When we speak or sing softly, the opposite ie increase at low and high frequencies, is bathtub frequency spectrum curve. We may try this ourselves even outside of a lab. All we need is our ears and our voice.

So we at least 2-3 bathtub curves to deal with and compensate before we can make a beautiful speech level singing voice sound correct at reproduction to the listeners brain. Fletcher Mansin curves explain what good engineers know and do, for example add a few dB above 2.5KHz to restore presence to voice lost by higher acoustic intensity.

These effects may explain why a flat microphone, flat preamplifier to a flat power amplifier to flat loudspeakers may sound so unnatural.

Great engineers have done great and successful efforts to increase the frequency responce of microphones and loudspeakers to cope with the most demanding transients of a piano.

A piano for example.

This great effort is not to be underestimated. Transient responce is very important.

It is just that when we consider the system as a whole, somewhere in the chain when recording speech level singing voice the bathtub curve need to be compensated with its inverse.

The preamplifier may be a good place.

It is nice to get the signal correct as soon as possible. The closer to source the ways kitty of components dictates the performance of the whole system. (An exemption may be not doing all the low cut need on ribbon microphones on the input transformer to get as much signal to the first tube stage. Other exception are deliberate high frequency preemphasis as used by FM with dee oasis to round any instantaneously clipped peaks in the waveform and make the sound loud due to a increase in average to peak ratio. This is very useful too before going to a analog to digital converter.

But the moment let us assume the signal correct as soon as possible. The closer to the source and the quality of components dictate the performance of the whole system.)

Pultec inductor capacitor equalizers are great tools for this purpose but why not getting the signal sounding right as soon as possible after the microphone. Why not attempting to reduce the number of equipment in the signal chain, reducing the weight, considering economy (including power consumption) reducing the complexity and increasing the quality of the overall system's signal path.

So the interesting fact is that a preamplifier to have an inverse bathtub curve it must have corresponding small inverse bathtub bandwidth to make the voice natural from vocal chords to brain.

Electron tubes on their own FETs and BJTs have an inherent bandwidth from DC to MHz. Surrounding components, (capacitors and inductors), have a reactance that varies with frequency and limit the bandwidth.

Pleiades (open source schematics so that anyone can perfect them) amplifiers use as few active devices as possible. Sometimes as low as just one.

For certain amplications such as voice production the MHz bandwidth is deliberately reduced to a few KHz . This us done by the surounding inductors, including transformers, and capacitors to restore flat frequency responce from vocal chords to listener's mind.

Notes:

Added advantage is the in the increase of signal to noise ratio,

On ribbon microphones it may be best not to do all the low cut compensation need at the inout transformer in order to get as much signal to the first stage. Future research will be useful.

Pleiades preamplifiers have a very low noise performance anyway.

This is because the anode voltage is as low as a few volts from one battery supplying the heaters too.

Cathode is underheted deliberately to keep electron emission cloud at smaller thermal movement and reducing the negative charge by electron cloud being induced to the nearby grid.

It may be stated that the negative bias that we have learnt to apply externally already exists inside a vacuum electron tube. At low plate operation voltage this is of paramount importance and blocks the operation of the tube.

The self assumed remaining negative charge on the grid is discharged by a high Megohm resistor from anode or Vb to grid. A typical value can be 4-8MΩ when Vb is 6 Volts. This frees electron and makes traveling to the anode very easy at very small fields created by a small voltage. So they can carry and convey as much information of a small signal amplified and arriving at its pristine glory at the anode.

Reference: Sound picture recording and reproducing Characteristics - Loye, Morgan - 1939 spring meeting at Hollywood - JMPSE