• Welcome to the all-new HUG. All content has been converted from the old system, and over the next days we will re-style HUG in a more familiar way.

What is 'dynamic range' and how does it relate to amplifier power and the listening experience?

HUG-1

Moderator
Dynamic range means the relationship between the quietest and the loudest sounds in an audio waveform. It is exactly the same concept of the dynamic range of a camera sensor or film stock between the detail in the blackest possible black that can be captured and reproduced and details in the whitest of whites. Between these extremes are the grey scales of vision, and we can think of the middling loudness of an audio recording in the same way. Audio and vision are exactly comparable: the loudness or dynamics of sound is exactly the same issue as the whiteness (or blackness) of vision.

Basic grey scale chart is here.

The importance of reproducing the dynamic range of an optical scene is of such importance that numerous half-tone (meaning, between black and white extremes) charts have been developed for specific purposes, such as here.

The steps between darkness and light are very carefully executed. From page 2 of this data sheet we can see the luminance ramp, the energy difference between each step.
 

A.S.

Administrator
Staff member
Grey scales and down on the farm

Grey scales and down on the farm

Dynamic range means the relationship between the quietest and the loudest sounds in an audio waveform. It is exactly the same concept of the dynamic range of a camera sensor or film stock between the detail in the blackest possible black that can be captured and reproduced and details in the whitest of whites...The steps between darkness and light are very carefully executed...
Now, what is a pity is that in the audio world we do not routinely examine the dynamic range of recording/replay equipment. Unlike our colleagues in the optical world, we are using words to describe our listening experience when they are using standardised measurement processes. They are right, and we are wrong. Cameras and lenses would not be bought by professionals on the basis of glowing prose, user stories and myth; they'd be procured after careful study of their faithfulness to the original light image passing through them.

Conventional loudspeakers, being electromechanical and with heavy moving parts, simply cannot accelerate from rest to some peak loudness (velocity) and back to rest in zero time. The inertia of the moving cones implies that the really fast moving elements of the music signal, those with a wide dynamic range, will not - cannot - be truly faithfully reproduced by any electro-mechanical loudspeaker. That's not really a problem because most music is not at the two extremes of the audio brightness chart: all in the blacks or all in the peak whites. Most classical/jazz/acoustic music rumbles along somewhere in the middle with occasional bursts to a high loudness or drops to almost silence. That's great news for the audio enthusiast because it means that the listening experience at home, using those speakers with cones that have inertia and hence cannot start and stop dead, is going to be OK. It's going to sound reasonably life-like unless we try and reproduce the canons on the 1812 Overture or a solo performance of tiny bells recorded at the back of the hall and nearly buried by recording hiss - the two opposite ends of the dynamic range.

We have to recognise then that speakers do what speakers do, and that we just have to accept that they are not truly linear devices. What we can do is to look elsewhere in the recording chain to see how well it captures the dynamics, the half-tones of our music, the (almost) infinite steps between black silence and blasting canons of peak white sound. Let's assume that the recording microphones, mixing console, mastering and CD pressing process is adequately linear (44kHz, 16 bit is more than good enough to capture the entire black to white range for the human ear). Even if it were not totally linear, there is not a damn thing we can do about it. We can't create shades of grey in the performance which have been deleted during the recording/delivery process to our homes: we have to work with what we are given. So what part of our system can we identify as a potential culprit for disturbing the nice, incremental grey-ramp of our recording's black-to-white loudness range? It's the amplifier. We can assume with good reason that the CD player, cables, room, stands, interconnects and all the rest are beyond reproach; our attention has to be to that of considering the amplifier's capability.

It should hopefully be clear by now that the amplifier, and the amplifier alone, is the engine that pumps the speakers with energy. It is the source of joules that cause the cones to begin to move, to accelerate to a terminal velocity, stop, reverse and return to rest, thousands of times a second. No joules, no sound. More joules, more sound, to the point at which the cones simply cannot stretch any further from their rest position because their (rubber) surrounds are fully taught. It doesn't matter how many more joules are available to the speaker beyond a certain level on its sonic grey-scale ramp once the surrounds are stretched tight; we will hear chronic distortion. It's pretty obvious from listening that the speakers are approaching that peak capacity and are under distress and most listeners would keep the volume setting below that point; their children may be less sensitive.

The myths surrounding the fabled audio quality of small amplifiers seems to fly in the face of elementary physics. The physics clearly demonstrates the certain, fixed, non-negotiable, actual, factual relationship between the power needs of a loudspeaker and the sound loudness that they will generate. There is no substitute for joules of energy. If the amplifier cannot deliver the energy the speakers are trying to pull from the amp then the speaker cannot accurately reproduce the black to white dynamic range of the original music waveform. The energy reservoir in the amp is the power supply capacitors and, just like milking a cow (which I watched with interest only yesterday), once those udders are sucked dry there is no more milk available, no matter how much the calf suckles.

The fact is that a 3W, 7W, 10W, 15W amplifier cannot behave as if it is a 25, 50 or 100W amp. It is the number of joules available in the amp's power supply reservoir that defines the amps potential, not the number of components, the size of the case, whether it is tube or sold state or any other factor. There is no substitute for joules, no substitute for watts.

So what happens when a tiny amplifier is commanded to reproduce the dynamic range of a black-to-white audio signal with all the (almost) infinite grey-level steps between those extremes? That depends on one critical factor. Whether or not the speakers are connected! If the speakers are not actually plugged-in, then the flea amp is perfectly capable of generating a wide-dynamic signal on its output sockets that is an exact mirror, but bigger, of the input signal. That's easy. It's easy because the power reserve in the amp is not actually being consumed because there is no electrical load - the speakers are not connected. An oscilloscope which presents no-load, clipped across the output terminals would show a marvellous signal, the full dynamic range from blackest black to whitest white. It would be impressive how even a tiny amp could reproduce all the greys of the music. But that's actually irrelevant. In the real world we need to hear the music and we need to plug-in those speakers with their heavy cones and 1% electrical/acoustic efficiency (that's normal for modern speakers). What then? Well, we turn up the volume until we hear a nice life-like sound. But the speaker, just like the milkmaids vacuum suckers, is sucking hard at the amp and within seconds the amp is struggling, it's out of juice. And then? Then it does not have enough joules to get up the audio brightness scale. It can reproduce the middle greys and the dark greys (because these don't draw much from the amp's PSU reservoir) but the fast, bright whites just cannot be reproduced.

Our little amp has altered the dynamic range of the sound. We've taken the incoming wide-dynamic range and chopped-off (clipped) the whites. What remains are the blacks and the greys of the recording. The inadequate power reserve has altered the contrast of the music, the shading between black and white. That's fine if we desire an 'inoffensive' non-challenging background music system with a compressed dynamic range and/or we listen at an extremely low level where we can be sure that the hungry speakers will never suck more from the amp than the amp has available in its electronic udder. But surely accurate reproduction of recorded dynamics is a basic requirement of a high fidelity reproduction system isn't it? Why deliberately starve the calf?

Examples to follow.
 

A.S.

Administrator
Staff member
Dynamic range: the audio equivalent of the grey-scal ramp

Dynamic range: the audio equivalent of the grey-scal ramp

Now, what is a pity is that in the audio world we do not routinely examine the dynamic range of recording/replay equipment. Unlike our colleagues in the optical world, we are using words to describe our listening experience when they are using standardised measurement processes. They are right, and we are wrong. Cameras and lenses would ... be procured after careful study of their faithfulness to the original light image passing through them....
I showed the grey-scale luminance tone chart here which those working with light use to evaluate graduations of brightness between jet black and peak white. In audio, we can generate exactly the same dynamic range scale between silence and peak loudness that the audio delivery medium (CD etc.) can reproduce. I've generated some pink noise and starting at -60dB (that is, 60 decibels below the absolute loudest signal a 16bit system can encode), in five second intervals I've increased the level by 10dB from the starting point of -60dB so that after six steps up the audio grey-chart we are at peak white. We cannot reproduce a louder signal with our 16 bits available but we can, in theory, encode a quieter one at an even lower loudness than the -60dB one. I'll return to that interesting point later.

Clip 1: pink noise stepped up from an initial -60dB to -50, -40, -30, -20, -10 and finally 0dB, absolute maximum loudness. See attached picture.

I'd guess that when you hit the play button you assumed that something was wrong because you couldn't hear anything but without adjusting your volume, by the end of the clip I imagine that you were feeling that the roar was rather uncomfortable. Right? We could write another whole piece on our human reaction to this dynamic range, but that's for another day. It's 2210 already. Obviously, these seven loudness levels are very crude jumps in loudness - technically we'd say that my example has very poor level resolution. Imagine if a musical performance only had seven loudness levels and the music hopped from one to another, rather that sweetly swelling and receding as acoustic music does: the shades of texture that makes for great music would be lost. I'm going to make another example but with 5dB increments between levels. I hope that you get the idea that music, just like light has many graduations of tone.

In fact, there are about 64,000 graduations of sonic loudness between silence and full level on a CD - best exemplified with this finely graduate luminance chart here.

To follow.
 
Attachments only viewable to members

A.S.

Administrator
Staff member
More on loudness and dynamics

More on loudness and dynamics

There has been a side-discussion of relevance here concerning the use of small amplifiers here. The conclusion at this point in the dicussion is that small amp = small sound (logically) and that if you want to listen at around conversation level (which is not a natural replay level for music) then a tiny amplifier may well be sufficient. But serious music listeners yearn to try and reproduce not only the frequency range of live music, but the excitement and dynamics of live sound, together the holg grail of high fidelity at home.

Having dedicated my working life to creating speakers that can play reasonably loud and, at that loudness, capture the essence of the spectrum and dynamics of real sound, I do not understand why the listener would intentionally limit the replay loudness of his system to that of a meagre capability of a tiny amplifier unless music is a background activity, in which case, why not invest in a high-end portable/bedside radio. When we arrive at certain recommended minimum power amplifier wattage for our speakers, these figures I consider will give something approximating to live sound, not loudness of the real thing of course, but a fulfilling quasi-real experience. That needs a certain minimum number of joules - or watts - and 5,7,10w or so doesn't meet that criteria. It's as simple as that.

However, we know that there are very deeply held views, nobviously not by mechanical/acoustical engineers, but amongst audiophile marketeers that considering all amplifier types and technologies available, that there is something unique, something marvellous, something deeply insightful about the sound reproduced by the use of a small amplifier (let's say, <25w). Let's just side step all considerations of sonic quality, and look at the one issue that won't go away, the elephant in the room: the admitted fact that small amps cannot (logically) produce a bigger sound oomph than their tiny power reserve can deliver. We can't kid ourselves that 10w amp can deliver anything more than 10w, and even that quoted figure may be optimistic under real world music signals. The truth might be nearer to 6w, both channels being driven hard by dynamic music into modern low impedance/low efficiency speakers and even at 6w, there may be 20%, 50% or more distortion thanks to current limited clipping, which is, as I've shown by audio clip, not always identifiable and may even be preferred by the listener.

So let's concentrate on the common ground, which is indesputible: small amps are only capable of generating a small amount of measureable loudness at the listeenr's ears. I can't imagine that could be disputed. Logically then, given two amplifiers, A having an output of say, 10W and B having an output of 50W, to make a proper valid real-world comparison of A v. B, B would have to have its volume knob turned down so that its greater power advantage was not allowed to generate a louder sound at the listener's ears, which would make a valid comparison impossible. Once we had set equivalence of loudness being within the capabilities of A and very well withing the capabilities of B, we could start to do some serious listening and any opinions we reach, assuming that the loudnesses were very carefully matched (to <0.3dB in my experience), would probably stand independent scrutiny by other independent observers.

The key point to remember is that when objectively, impartially, professionally, comparing any audio equipemnt or event it is the maximum loudness potential of the lower of A or B which sets the maximum loudness for the test. It's irrelvant that the 100w amp is capable of 20dB or more of loudness at our ears if the 10w can't produce that loudness. We have to turn down the 100W amp until it produces no more loudness than the 10W amp. Then we're ready to do some forensic listening, and not before. I trust this is absolutely accepted.

Earlier on in this thread I talked about the tools the photographer uses to set the peak whiteness of his image just before 'burn-out' and the blackest black before all detail is lost. He calls that his optical dynamic range and would calibrate his equipment using a grey-scale chart as I showed. It's interesting to be on a TV set; before transmission the vision director will ask for all the cameras to be calibrated and a grey and colour calibration card will be held up to each cameras's lens (all of which may be of different makes) and back in the gallery the 'tone controls' on the vision mixing console are adjusted so that all the pictures we see, regardless of camera angle, have exactly the same brightness and colour balance. Have you noticed how well matched the cameras are on TV programmes? Once the camera tone controls are locked set, it's essential that the lighting rig is not altered at all. There cannot be any more illumination of the set or any less. If the lighting director has a change of mind, then the cameras must be recalibrated before transmission or recording. The reason is that if different parts of the set are stronger or weaker lit, the individual camera sensors pointing across the set will be operating at different steps on the grey-scale tone chart; their calibration is only optimisable at one specific light level or luminance. The consequence of changing the light but not the calibration is that, to the human eye, the underlit camera angles will look dark, murky and dingy: details will disappear into the blackness of the image and technically, the dynammic range of the image will be reduced. And that brings us neatly to audio dynamic range and the ears, not optical dynamic range and the eyes.
 

A.S.

Administrator
Staff member
On the TV set .... illumination & volume

On the TV set .... illumination & volume

Let's compare what we'd see at home if two TV studio cameras were broadcasting our favourite weekly multi-camera show, but a last minute intervention by the big star's agent resulted in an unplanned change in studio lighting to be more flattering or whatever. 'Just turn the lights down a shade' says the agent. So they oblige, but there's no time to recalibrate the cameras. The lighting department has done to the luminance (brightness) on-set exactly what we do when we turn down the replay volume (loudness) of our hi-fi system at home.

First, I should say that what you will see on the attached picture will slightly depend upon whether you are viewing on an Apple product, a PC or Android. That's because Apple use a different grey-scale representation on screen; their screens are brighter and with (technically) a different gamma. The blacks will look somewhat brigher- none of this matters, we're just observing a general point.

The attached shows the perfectly graduated luminance grey-scale between peak white and black black when cameras A and B are optimised for whatever studio lighting there was in rehearsals. You can see that all the way up and down the scale there are clear, but subtle, graduations in lightness. Just as it should be.That tells us that regardless of whatever shade the camera registers, it will correctly reproduct its brightness to us at home,. Everything on the TV set, and especially the faces of the actors, will be rendered just as if we were there. The last strip shows what happened when the agent insisted in the light over his actor be turned down - the lighting director didn't have time to recalibrate the camera for the lower luminance. Let's analyse the uncalibrated image from camera B.

There are two very important observations we can make if the 'volume' of sound or the level of illumination is decreased, as with the uncalibrated camera. In both cases, this is an issue with the human ear/eye and may not necessarily apply to all creatures - we'll never know. We have to work with this sensory reality, not be in denial about it. We note that the uncalibrated grey-scale chart has slid downwards a little compared to the calibrated ones. That means that the overall illumination has dropped, and now the peak illumination as seen by the uncalibrated camera is only as bright as the second peak level for the calibrated camera. We've lost the ability to reproduce that ultimate state of whitness, the peak loudness of the image as it were. That's not ideal, but we'd probably not notice the loss. What is much more of a problem is that below a certain illumination - a certain loudness - all the subtle shades of brightness, the shading and texture of hair, the folds and creases and patterns in cloth, the details in the shadows of the set have all gone. They've all been fused together into a black, amorphus mush, exactly the way that hiss is the enemy of high fidelity because it erases and coagulates microtonal detail into the audio blackness under and around musical detail.

How does this relate to music? Optical illumination is analogous to musical volume. As Peter Walker of QUAD said, think of the volume control as a focus control, bringing you closer to the performers. What he implied is that if the replay volume setting is too low, and it will be with a small amp trying to passably reproduce real life, the musical dynamic range detected by the human ear is going to be compressed and all the really quiet elements in the music will slide into the inaudible blackness of no-sound. In which case, why invest in a good hi-fi system? There needs to be a certain loudness to give the reproduced music enough energy to get it evenly lit, so everything to fall into place throughout the performance and between all instruments.

For anything approaching high-resolution, high-fidelity sound, there needs to be sufficient loudness to make the best possible use of the ear's auditory potential.

>
 
Attachments only viewable to members

A.S.

Administrator
Staff member
A tiny tube amp

A tiny tube amp

Absolute bargain! A 3.5W + 3.5W (that's not much more than a headphone amplifier) for a mere $2999 on sale at an internet auction.

Clearly quality will out.

Wait! Update! Let's find all about these tiny tube amps. Let's lay down some cash, and see what we find. Note the ambiguity about the power rating - how 40w becomes 16 real, rms watts. I bought it! And it has tone controls! I'm going to put this issue to bed once and for all.
 

mhennessy

Member
All tube?

All tube?

I'm impressed with your commitment to this!

I'm not sure what you'll find in that box - looking at the valves it gives the impression of being single-ended - in which case 16 watts doesn't seem at all likely. Perhaps it's a hybrid with a solid-state output stage? I note the fan on the rear panel with interest...

If nothing else, it's a nice box, especially with that "Nostalgic VU- level meter". Look forward to the report :)

Mark
 

A.S.

Administrator
Staff member
I'm impressed with your commitment to this!

I'm not sure what you'll find in that box - looking at the valves it gives the impression of being single-ended - in which case 16 watts doesn't seem at all likely. Perhaps it's a hybrid with a solid-state output stage? I note the fan on the rear panel with interest...

If nothing else, it's a nice box, especially with that "Nostalgic VU- level meter". Look forward to the report :)

Mark
So headline '40w' becomes '16w rms' and now, what, 5w? Are there any engineering truths in audiophile amps at all? Even one? There has to be one, surely. Or is it all smoke and mirrors? Don't answer. The day Peter Walker retired was the day engineering pragmatism and truth exited. What a charade we're seemingly left with now.
 

Pharos

Member
Component layout may be critical

Component layout may be critical

In my apprenticeship I, as did many of us from ''66 to '69, built numerous and varied amplifiers, almost of which were valve. From formal layouts using component boards as shown in that example, with long connecting wires, we gradually adopted 'point to point' wiring, because it was effectively delineating, shorter, and simpler in execution.

It did not however have the neat visual appearance which the rows of parallel components convey. But by wiring point to point fewer problems occurred and better results were obtained.

In my last year a student from two years later asked me (because of my reputation with such), to build his ELL80 amplifier for him, which I did. Though it worked fine when I had finished, he didn't like the point to point approach, and tore it apart to then rebuild it in the component board layout.

When retested it had an 8Mhz oscillation which he could not get rid of, and he asked me to rebuild it for him, which I declined to do. The amplifier on Ebay is wired in this way, using component boards with the components next to, and parallel with, each other. Look at those lengths of twisted wire, and some parallel with other ones! Asking for problems.

All of our amplifiers were of the standard 'sensible configurations'; EL84 and EL34 push pull ultra linear O/Ps, with a few KT66s as well. Partridge O/P transformers were regarded as the best at the time. No SE stuff.
 
Last edited:
Top