Forum Home Forum Home > Other music related lounges > Tech Talk
  New Posts New Posts RSS Feed - Analogue, Digital, Vinyl, CD - tech talk
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Topic ClosedAnalogue, Digital, Vinyl, CD - tech talk

 Post Reply Post Reply
Author
Message
 Rating: Topic Rating: 1 Votes, Average 5.00  Topic Search Topic Search  Topic Options Topic Options
Dean View Drop Down
Special Collaborator
Special Collaborator
Avatar
Retired Admin and Amateur Layabout

Joined: May 13 2007
Location: Europe
Status: Offline
Points: 37575
Direct Link To This Post Topic: Analogue, Digital, Vinyl, CD - tech talk
    Posted: April 26 2015 at 05:20
I have started this thread in response to the current "Vinyl" thread that exists in this Tech Talk Lounge which has become nothing more than a chat room for a couple of vinyl enthusiasts. There is nothing inherently wrong in that, it just should not be in "Tech Talk" in my opinion if technical talk is not tolerated by those few that use the thread to display their latest vinyl purchases, even when they make subjective technical comments of their own such as "It is so nice to hear these songs in all their analog glory again...The nice big sound, deep bass..." and "...the bass and drums groovy intro in analog perfection...". I'm not going to respond to those specific points in that thread or this because they are subjective, and that is simply a question of personal preference. [not withstanding that the only copy of 'Farewell To Kings' I own was purchased in 1977 so I am unable to make any comparative judgement/assessment/comment on said album, if I could, I still wouldn't]. 

However in any discussion or commentary on a vinyl copy of an album the technical element cannot be avoided, it is inevitable that any comment on the media (that is: the music) will also contain a comment on the medium (that is: the technical format used). Because those comments are subjective the "better, best, worse, worst" argument for that medium is implied even when other mediums are not directly mentioned. Whether it was intended or not "analog glory" implies that digital is less glorious, "analog perfection" implies that digital is less perfect.

In dozens of threads in this forum I have responded to technical statements made by technical and non-technical people about all manner of technical subjects, it's what I do because the subject matter interests me, I am a professional electronics engineer who specialises in Mixed Signal components (Mixed Signal means components and circuits that contain Analogue and Digtal signals) and I am an amateur audio/recording enthusiast/'musician' who also happens to be a life-long music lover who collects, and more importantly listens to, music in all formats. 

Because I have spent a lot of time addressing misconceptions and down-right inaccurate "facts" about analogue (and thus vinyl) many have assumed that I prefer digital or that I have a downer on analogue/vinyl. This is simply because more people have made more claims about that particular topic that I wanted to answer - when people make technically inaccurate comments about digital then I comment on those too, there are just fewer of them. In my experience analogue supporters tend to be more vocal and more defensive so I tend to upset more of them.

I have never stated a preference and never will (much to the consternation of those who would like me to) - each has good points and bad points just like everything else in this world. I am interested in knowing and explaining what causes those differences in the real world from a technological point of view. That will involve physics and mathematics, it will also involve the physiology of human hearing and the psychology of how we interpret sound but hopefully we can keep this to an easily understandable level - not through "dumbing-down" but by readily accessible clarification and explanation. It's science Jim, but not as we know it.

For example when the claim is made that "vinyl is warmer" then I want to know why that is - whether it is technical or psychological, or whether it is physics  or physiology. If a claim is a subjective effect then I am interested in finding out if this is the result of an objective cause. To do that it is necessary to first ask "is vinyl warmer?" then if that is answered then ask "how is vinyl warmer?" and if that can be answered then ask "can CD be warmer?" - of course we can simply jump to the last question.

This thread is about the technical aspects of all mediums and all the technology formats used.
This thread is for technical discussion about audio technology: gear/audio/analogue/digital/CD/Vinyl.

This thread is not about which medium is better.
This thread is not about which format anyone prefers.
This thread is not about subjective comparisons between media produced on two different mediums.

However, those are not cast-in-stone restrictions - if people want to use such subjective assessments in support of the technical points they are making then that's fine. If people want to discuss conversion, compression, EQ and other subjective uses of the technology then that is fine too.

I suspect that this thread will die a death, but at least I tried.

Note: Analog vs Analogue ... I'm British, I speak and write British English and will use the British spelling of "Analogue" in exactly the same way I write "Catalogue" and not "Catalog" - the words are the same, just different spelling. I will also use the British "Valve" as a short-hand for "Thermionic Valve" and not the American English "Tube", but they refer to the same electronic component.




Edited by Dean - April 26 2015 at 05:23
What?
Back to Top
Dean View Drop Down
Special Collaborator
Special Collaborator
Avatar
Retired Admin and Amateur Layabout

Joined: May 13 2007
Location: Europe
Status: Offline
Points: 37575
Direct Link To This Post Posted: April 26 2015 at 06:06
Mr Know-it-All.

As I stated in my OP, I am an electronics engineer by profession. My interest in electronics began back in the early 70s because I was interested in music and audio. From simple beginnings in constructing basic crystal radio receivers, repairing and modifying old valve radios to making audio amplifiers I started an apprenticeship in Electronics with the British MOD. There I was taught the basics of practically every aspect of electronic such as audio processing, logic circuitry, radio, radar, computing (analogue and digital), avionics systems, remote metering and measurement (telemetry), optics and displays, simulation and emulation in both solid state and valve technology. I then went on to do a degree in Electronic Engineering that put the science-flesh to those bare bones of technical knowledge. Since then I have specialised in measurement and evaluation of complex analogue, digital and mixed-signal microelectronics, this (as you would expect) requires me to understand the science behind those technologies as well as the technologies themselves so when something does not behave as expected I can analyse and therefore explain why it doesn't.

Throughout that time I have had a pragmatic approach to HiFi as a hobby, I admire top-end equipment for its build quality aesthetics but have come to separate that from any technical audiophile claims made for it when those claims are not supported by the rigours of scientific and technical investigation. I love HiFi but I'm not blindly infatuated by it, I can never call myself an audiophile because my interest in HiFi requires me to understand why something happens and not just accept that it does.

As I said before, I am also an amateur audio/recording enthusiast - I am not an expert in that field and what little I do know is self-taught and learnt by experience through the application of the knowledge I have in other fields of electronics. A professional sound engineer (of which there are some here on this forum) can (and do) wipe the floor with me on this subject at every level - I welcome this because I am interested in learning from them.

What I won't do is say which bit if kit is better than any other, nor will I make specific recommendations.

So...


All this wealth of knowledge is locked away in my head. If anyone would like something explained that they don't understand I am willing to have a go at explaining it. If people have questions I will answer them to the best of my ability. And if people think I am wrong then please put me right.


Edited by Dean - April 26 2015 at 06:12
What?
Back to Top
Dean View Drop Down
Special Collaborator
Special Collaborator
Avatar
Retired Admin and Amateur Layabout

Joined: May 13 2007
Location: Europe
Status: Offline
Points: 37575
Direct Link To This Post Posted: April 26 2015 at 11:29
Sound is not Analogue

An often stated remark is that sound is analogue. This is used to stress that analogue signals are in some way more authentic than digital signals. This fits into our visualisation of analogue signals as smooth wavy lines on a graph whereas digital signals are represented by angular stepped staircase wave forms and for the unwary this all makes perfect sense. Except that sound is not analogue, it is sound: a three dimensional compression of air molecules that we call a pressure wave - it does not wave up and down like the line of a graph so it is not the smooth wavy line on a graph or something that can be seen as a squiggly line on an oscilloscope. Those are representations of the pressure wave that are comparable to the real pressure wave, or in other words: they are an analogy of the sound. 

The voltage value at any given moment in time is analogous to the air-pressure value caused by the sound. This is where the word "analogue" in this usage comes from - it is analogous to the sound, it is not the sound and therefore the sound is not "analogue", it is just sound. By the same argument, digital signals are also analogous to the sound pressure wave - each code value at any given moment in time is an analogous representation of the air-pressure value caused by the sound - digital signals are analogue representations of sound. We use the words "analogue" and "digital" to differentiate between the technologies used, neither are "authentic" representations of the sound itself.

Another way to think of this is with a clock. A clock is not time, it is the representation of time and so it is a device that we use to measure the passage of time. In the past we have used other analogies to measure time - the movement of a shadow caused by the passage of the sun across the sky, the time to burn a candle or time it takes for water to drain from a leaky bucket. Then we invented a more accurate measure and this was the time taken in the unwinding of a spring as measured by the one-second swing of pendulum - the count of those seconds was represented by the analogy of hands moving around a dial. The movement of the hands is comparable to the passage of time "in the real world", therefore it is analogous to the passage of time. We did not need to call these wind-up clocks Analogue Clocks, we just called them Clocks, but they measured time in discrete time intervals of one second. tick-tick-tick-tick. When clocks were invented that displayed this count in numbers we called them Digital Clocks because the numbers were numerical digits and to differentiate them from clocks with hands that we now call Analogue Clocks. Yet both are the analogue representation of time and both measure time in discrete one-second intervals, we just use the words "analogue" and "digital" to denote the display method.

As I said in the opening paragraph sound is a three-dimensional compression of air molecules. When something makes a noise it radiates this pressure wave out in three dimensions but this radiating wave is not uniform in all directions (the technical term is "anisotropic" as opposed to "isotropic"). As we move around in the three-dimensions of the real world the sound we hear changes as we pick-up the differing sounds of this anisotropic radiating wave. It also reflects off hard surfaces and is absorbed by soft surfaces, this also affects the sound we hear as the sound wave interacts with itself as it arrives at our ears through the different paths it takes. This means we cannot fully capture the sound at any given time when we try to record it - the recorded "analogue" is not an accurate representation of the sound even when the sound recording engineer uses multiple microphones to record that sound. How well something is recorded and converted into analogue voltage signals determines how close the recorded sound is to the original sound but it can never be an accurate representation.

What people mean when they say that "Sound is Analogue" is that the sound is a continuous wave just as the voltage trace on an oscilloscope is a continuous wave. However what we see on the oscilloscope is the result of the conversion of that pressure into a voltage and how accurate that trace is will be determined by the resolution of the microphone. 

We are used to the term "resolution" when talking about digital signals because we say that a 16-bit digital audio signal has 16-bit resolution... what we mean by that is the smallest step the ADC converter can resolve is 1-bit or 1/65535th of the full-scale value [If the full-scale swing was 1 volt then that 1-bit would represent 15.26µV or 15 millionth of a volt]. However, we don't normally think of a microphone having a resolution in the same way that an ADC does, but it has - the moving parts of the microphone have mass and therefore inertia so there are sounds that are too weak to move the diaphragm and this gives us a threshold below which no sounds can be recorded and that would be equivalent to 1-bit in the digital domain. If that weak signal is superimposed on a stronger signal we would not be able to resolve the weaker signal - by the same reasoning we cannot resolve a small change in the stronger signal that is equivalent to the same pressure-change represented by the weak signal. Therefore the continuous movement of the oscilloscope trace is not an accurate representation of the sound. 

Similarly, there are also sounds that are so loud that they move the diaphragm so far that it can move no further and this is equivalent to full-scale in the digital domain (e.g., 65535 bits). We call the ratio of quietest sound to loudest the Dynamic Range and both analogue and digital signals have a dynamic range. Analogue signals produced by microphones (and by inference, turntable cartridges and all other electromechanical transducers) do not have unlimited dynamic range and they do not have infinite resolution. We can calculate the equivalent number of bits that this dynamic range represents, so for example a 98dB dynamic range will be the equivalent to a 16-bit ADC and vice versa.

The reason why we visualise an analogue signal as a continuous smooth wave and a digital signal as an angular stepped staircase wave is simply a matter of convention. That's how people draw a digital wave to describe how it works - it is not how we see it on an oscilloscope. 

For example if I wanted to explain a digital wave I would draw a triangle-wave with (say) 16-discrete voltage levels so everyone could visualise how the analogue voltage is quantised into 16 different voltage steps - the problem with that is 16 steps is the equivalent to a 4-bit digital signal, and real life digital signals are not 4-bits. To draw a 16-bit wave I would need to draw the wave using 65535 steps... if I took a sheet of paper and my favourite 0.5mm Pentel pencil, and drew two parallel lines 100mm apart then tried to draw a 65535 step wave between those lines then each step would be 100mm/65535 = 0.001526mm ... or over 300 times thinner than the width of my pencil line so the resulting triangle-wave I drew would look identical to the equivalent analogue wave. To be able to show the 1-bit resolution of a 16-bit digital signal using a 0.5mm pencil I would need a sheet of paper at least 32m (105 feet) tall.

An oscilloscope has the same limitation - if the screen is (say) 100mm high and the trace is 0.5mm thick - then it is impossible to resolve 65535 discrete voltage steps using that 'scope.

Of course in the real world we could never see discrete steps of the digital waveform because those steps are filtered off when the digital signal is converted back into an analogue one. That filtering does not affect the information encoded into the digital signal at the beginning of the process, it merely removes the ultra-high frequencies introduced into the signal by the quantisation process itself.



Edited by Dean - April 26 2015 at 12:00
What?
Back to Top
Kati View Drop Down
Forum Senior Member
Forum Senior Member
Avatar

Joined: September 10 2010
Location: Earth
Status: Offline
Points: 6253
Direct Link To This Post Posted: April 27 2015 at 07:29

analyzing band releases inc. from the 70's and to date the vinyl mostly do have more dynamic range, however digital can, if mixed properly and not compressed to hell, produce a much more dynamic sound, i.e. LP very heavy deep bass notes would make the needle jump WinkBig smileHug

Back to Top
Gerinski View Drop Down
Prog Reviewer
Prog Reviewer
Avatar

Joined: February 10 2010
Location: Barcelona Spain
Status: Offline
Points: 5087
Direct Link To This Post Posted: April 27 2015 at 09:22
Interesting thread Dean.
About "Sound is not analogue", I guess that this sentence could lead to controversy.

Of course ultimately everything is discrete, if we should define "analogue" as "having infinite resolution" then nothing would be truely analogue. The air itself has a resolution in the pressure wave caused by the size of its constituent molecules and the density with which they fill up space, our hearing organs have also a limited resolution based on their physiology and material, and the firing of our neurons is also discrete. But I guess that everybody would conventionally say that the air pressure wave itself and the stimulation of the hearing nerves produced by our hearing organs are "analogue processes".

Rather, I would say that "digital music" is also analogue when it reaches our ears, since it has been converted to another pressure wave by the speaker.

Indeed as you say, what we mean with "analogue" or "digital" in this context refers only to the technology used in the recording and reproduction process and should not be confused with the properties of the sound itself.
Back to Top
Dean View Drop Down
Special Collaborator
Special Collaborator
Avatar
Retired Admin and Amateur Layabout

Joined: May 13 2007
Location: Europe
Status: Offline
Points: 37575
Direct Link To This Post Posted: April 27 2015 at 10:48
Originally posted by Gerinski Gerinski wrote:

Interesting thread Dean.
About "Sound is not analogue", I guess that this sentence could lead to controversy.
I hope not, though I expect it could prompt contentious debate. 

The counter phrase "Sound is analogue" is often delivered as a coup de gras comment when what the speaker means is "Sound is a continuously variable physical quantity". It's pedantry but it's pedantry that carries more than just being smart with the meaning of words.

Both sound pressure waves and analogue electrical signals are continuously variable physical quantities where the latter is an analogue representation of the former, this does not mean that "Sound is analogue"

Originally posted by Gerinski Gerinski wrote:

Of course ultimately everything is discrete, if we should define "analogue" as "having infinite resolution" then nothing would be truely analogue. The air itself has a resolution in the pressure wave caused by the size of its constituent molecules and the density with which they fill up space, our hearing organs have also a limited resolution based on their physiology and material, and the firing of our neurons is also discrete. But I guess that everybody would conventionally say that the air pressure wave itself and the stimulation of the hearing nerves produced by our hearing organs are "analogue processes".
Correct, nothing has infinite resolution and it is convention to say that anything that appears to be a continuously variable quantity is "analogue". 

The question is then - at what point does the discrete (quantised) interval denoted by the resolution change to being a continuously variable quantity? That is, when do we observe this quantised value as a continuously variable one.

This is answered by the instrument you are using to make the observation. If the instrument has a lower resolution than the discrete quantity levels being measured then we will not be able to see the increments. In my previous example the 0.5mm pencil was the resolution of the instrument and it would be impossible to draw a 1.5µm step with that pencil so the digitised waveform will look continuous.

In the case of the microphone used to convert the sound into an electrical signal 1 air molecule will be unable to move the diaphragm so we cannot resolve those individual air molecules. An output voltage will only be produced when there is sufficient pressure to move the diaphragm. We also have to contend with the issue of noise as well, but that's a different discussion.

Originally posted by Gerinski Gerinski wrote:

Rather, I would say that "digital music" is also analogue when it reaches our ears, since it has been converted to another pressure wave by the speaker.
All electrical signals have to be converted to a pressure wave by the speaker.
Originally posted by Gerinski Gerinski wrote:

Indeed as you say, what we mean with "analogue" or "digital" in this context refers only to the technology used in the recording and reproduction process and should not be confused with the properties of the sound itself.
Thumbs Up
What?
Back to Top
Padraic View Drop Down
Special Collaborator
Special Collaborator
Avatar
Honorary Collaborator

Joined: February 16 2006
Location: Pennsylvania
Status: Offline
Points: 31165
Direct Link To This Post Posted: April 27 2015 at 11:21
This EE supports this thread.  Thumbs Up
Back to Top
Dean View Drop Down
Special Collaborator
Special Collaborator
Avatar
Retired Admin and Amateur Layabout

Joined: May 13 2007
Location: Europe
Status: Offline
Points: 37575
Direct Link To This Post Posted: April 27 2015 at 12:42
Originally posted by Kati Kati wrote:

analyzing band releases inc. from the 70's and to date the vinyl mostly do have more dynamic range, however digital can, if mixed properly and not compressed to hell, produce a much more dynamic sound, i.e. LP very heavy deep bass notes would make the needle jump WinkBig smileHug

All recorded music is compressed even on vinyl, uncompressed music is too quiet and, counter-intuitively, lacks punch that we would normally call "dynamic". This is to do with how loud the music seems to be rather than how loud it actually is. Compression techniques were first developed for vinyl so that more music could be squeezed onto a single side of a disc. If you've a single album that has more than (say) 12 minutes of music on each side then it would have been compressed - if it has 25 to 30 minutes of music on each side then it will have been compressed to b*ggery. The technology back then was pretty crude so heavily compressed vinyl sounds truly awful and nothing like the over-compression we associate with the "loudness wars". 

Loudness is a psychoacoustic effect (i.e., it is a combination of psychological effect of sound and the physiology of hearing) where things sound louder if we reduce the mid-ranges because it sounds closer to us in the physical world. The converse of this is when we turn down the bass and treble the music it not only sounds quieter, it also sounds further away. Loudness is a combination of compression and EQ and has nothing to do with the technology involved. 

The association between over-compression and digital is coincidental and changing the dynamics of a piece of music has nothing to do with the dynamic range of the technology used. 16-bit digital audio has a far wider dynamic range (96dB) than vinyl (<50dB). [A word of warning here: dB's are notoriously poorly understood critters - the difference of 46dB between them is not twice as much but 200 times more because it is a logarithmic quantity].


Bass is another less-well understood thing. Many assume that bass is very low frequency sound and in one example a person (who much to my annoyance and consternation has since been banned from here) tried desperately to convince me that the bass notes that you can feel in your chest were subsonic. 

The lowest note a standard bass guitar can play, E1 (E in the Contra octave), has a frequency of 41.2Hz. To put that into perspective the mains hum that you and I hear is 50Hz and that is far from being subsonic. A bass drum is also tuned to somewhere in this region (between 40 and 80Hz). These are not subsonic and for the sake of argument we can broadly say that subsonics are bad - we cannot hear them and they use up lots of valuable space in a signal (digital and analogue) where we can put music that we can hear, so they are filtered off long before they are committed to tape or disc or CD. Neither tape, vinyl nor CD can reproduce signals below 20Hz simply because we don't record them.

The thump or kick of a bass note and the low frequency vibration you can feel (rather than hear) is not subsonic, it is just the shape of the 41.2Hz note as it is played and that is preserved through the recording process as it is a variation in level of the 41.2Hz tone.

Also, what differentiates a bass guitar from a kick drum or Rickenbacker bass guitar from a Fender Jazz bass is not the note itself but the timbre and/or tone of the note played and that is determined by the harmonics those instruments create of that basic note. Because these are harmonics they are subsequently higher frequencies (ie 82.4Hz, 164.8Hz, 329.6Hz etc) and it is those higher frequency harmonics that make a bass note sound good, and therefore make a bass sound good and thus make the bass tones of a piece of music sound good. 

There is another psychoacoustic effect called Missing Fundamental that comes into play here, this is where a note that has lots of harmonics can have its basic note removed (in this case the 41.2Hz) and the brain will not notice that it is missing - this means that the really low frequencies are not as important as they first appear or at least not as important as the higher frequency harmonics.

[In music recording no one would go to the extremes of removing those sub-80Hz tones completely but it does mean that for example that when recording a bass guitar some of the low frequencies that clash with the kick drum can be reduced without affecting how the bass guitar sounds in the recording and we get to hear the kick drum with a little more clarity.]

What this means is that when you hear a piece of music and observe that it has a really good (deep heavy warm whatever) bass you may not necessarily be commenting on the low frequency response of the recording but on how well the higher frequency harmonics of that bass sound have been preserved.

As I said in t'other thread - bass notes cannot be recorded in stereo on vinyl but that doesn't mean they cannot be recorded at all. Fortunately because of our physiology (how far our ears are apart) we cannot tell with any accuracy where a low frequency sound is coming from so this is not a huge limitation, it just means we cannot put a bass guitar in the left channel and a kick drum in the right - both have to be monaural and therefore centre-stage in the mix but since we cannot hear where they are with any precision this isn't an issue. We can however, put their higher-frequency harmonics wherever we like (should we so chose, which isn't that likely but there's nothing to say we can't).

There are other frequency-constraints with making vinyl that I kinda skimmed over before and I'll do the same here Wink


Edited by Dean - April 27 2015 at 12:50
What?
Back to Top
Dean View Drop Down
Special Collaborator
Special Collaborator
Avatar
Retired Admin and Amateur Layabout

Joined: May 13 2007
Location: Europe
Status: Offline
Points: 37575
Direct Link To This Post Posted: May 01 2015 at 07:59
Yesterday an articled on "The Loudness War" from The Times newspaper (and others) started trending in my FB newsfeed. This was prompted by this article by UK Mastering Engineer and "Loudness War" guru Ian Sheppard from a month ago. One line in the Times article made me raise a simultaneously bemused and quizzical eyebrow: "On the decibel front, the album [Taylor Switft's 1989] outscores Metallica’s The Black Album, AC/DC’s Back in Black and Motörhead’s The Ace of Spades."

The deciBel:

As I said in the previous post, deciBels "are notoriously poorly understood critters" ... they are a ratio between two values expressed in logarithmically and since they are a ratio they are essentially useless unless you know one of the values that the deciBel is a ratio of. However, there is a convention of suffixing the dB with another letter to specify the base-reference so dBW means relative to 1W, dBm mean relative to 1mW, but reading dB without a suffix only tells you the ratio.

For example an amplifier with a voltage gain of 40dB doesn't tell you anything about the amplifier other than it has a Vo/Vi gain of 100, for example you cannot tell from that how big an input signal can be before the output clips. 1V on the input will not produce 100V on the output of the amplifier if it clips at 30V... but knowing that the maximum output is 30V we can calculate that the maximum input will be 0.3V. 

The reason why dBs use a logarithmic scale is because human hearing is logarithmic - if we play two tones of the same power one after the other we can tell they are the same loudness, if we keep increasing the power of one of them until we can tell that they are different we find that that that threshold occurs when one tone is roughly twice the power as the other. 

If we then set both to that higher power and repeat the experiment we get the same result and so on, so if we started at 1W then the steps would be 2W, 4W, 8W, 16W ... if you plotted those results on a graph you would not see a straight line but a curve and that curve would be exponential. To convert that curve to a straight line we need to make the power-axis logarithmic.

So we can express those powers on a logarithmic scale just by taking the logarithms of each power as a ratio of the starting power [e.g., log (2W/1W) = 0.3Bels, which we multiply by 10 to get 3dB ... since we know the base-value is 1W we can use the convention dBW... i.e. number of dBs relative to 1W] so now that sequence becomes 3dBW, 6dBW, 9dBW, 12dBW. [One important point to note from this is a 100W amplifier is not twice as "loud" as a 50W amplifier, the difference is still only 3dB - that is, the level at which we can just tell they are different].

Dynamic Range of a system:

As we have seen, Dynamic Range is expressed in dB and this tells us the ratio between the maximum voltage a system can produce to the smallest. The problem there is if the smallest is zero then the ratio will be infinite (we cannot divide a number by zero) so we call the smallest the smallest we can detect - in a digital signal this is easy as it is 1-bit but in an analogue signal this is harder to determine so we say that it is the smallest signal that can be detected above the inherent noise of the signal (noise is always present in any analogue electrical circuit) - since when we amplify the signals that are smaller than this value we also amplify the noise by the same amount it does not matter how loud we try and make it we will only hear the noise. 

Therefore output signal from a DAC in a CD player will have a dynamic range of around 98dB and the output signal from a phono-preamp from a turntable will have a dynamic range of around 50dB even when the voltage outputs of both is set to the same value. So if they both produce a maximum output of 500mV then the smallest output the CD player can produce will be 6.2µV compared to the turntable's 1.6mV. 

These have nothing to do with "Loudness Wars" - A CD will always be capable of reproducing a song with a greater dynamic range than a vinyl can - this is an inescapable fact of life. If you rip an album from vinyl and convert it to digital it cannot (and will not) have a better dynamic range than 50dB, regardless of how many bits you use in the conversion, conversely if you take a song with 98dB dynamic range from a CD and cut it into vinyl it will also be reduced to 50dB dynamic range.

Dynamic Range of a Recording:

This is where a lot of confusion, misunderstanding and misleading occurs. [I shall ignore analogue for a moment because as I've shown it can never have a wider dynamic range than CD and just concentrate on 16-bit CD recordings (and because the "maths" is easier to follow).] 

The maximum signal that a CD can reproduce is equivalent to a code of 65535 bits and that (say) gets converted to an equivalent analogue voltage of 500mV by the DAC, which then gets amplified and fed to the loud-speaker. That will be the loudest sound we can ever hear from the CD. 

The minimum signal the CD can reproduce is equivalent to a code of 1 bit and that will get converted to an equivalent analogue voltage of 6.2µV. That will be the quietest sound we can ever hear from the CD. 

If we assume that 500mV produced a peak output of 50W then the 6.2µV will produce a peak output of 7.8nW. [Proof: 10×log(50÷7.8e-9) = 98dB] ... as you can imagine that is really quiet - it is very likely to be too weak to even move the loudspeaker at all.

It does not matter whether you play Taylor Swift or Motörhead - the maximum power that either CD can produce in this example is 50W so the loudest sound that they can make on the same hifi system will be exactly the same.

If we simply turn down the volume control we reduce the volume of both the loudest sound and the quietest sound by exactly the same amount so the dynamic range is unaffected. [Everything is ratio-metric].

So if we have a recording that has a fade-in or a fade-out on a track then it stands to reason that the output of the amplifier will go from 0W to 50W so the Dynamic Range of the recording could be said to be still the equivalent to the 98dB dynamic range of the CD. 

No amount of compression will undo those fade-ins ad fade-outs so in that respect compression and over-compression does not affect the Dynamic Range when it is expressed this way. 

So... how is Dynamic Range expressed so that compression does affect it?

Dynamic Range as Crest Factor:

One way that is used is to compare the peak value of the signal (i.e., the maximum value equivalent to code 65535 or 500mV in our example) to its average value. The average used here is the Root-Mean-Square (RMS) value which is the equivalent direct-current heating effect of an alternating signal. In other words it is the equivalent DC voltage needed to produce the same heat in a load as an AC signal - for a sine wave this is 0.707 times the peak voltage. We call the ratio between the peak and the RMS the Crest Factor and this is a term you see often in "loudness war" articles.

Since the peak is fixed to an equivalent code of 65535 bits then the Crest Factor gives us a measure of the average RMS power in a recording, and this seems to make logical sense. In our previous example the CD can peak at 50W but the average power for the duration of the song or the album will be considerably lower and a recording that has a higher average power will be louder than one with a lower average power.

So now if we compare the rms power of two different recordings we can say that one is louder than the other even though their peaks are the same. Compression changes the ratio of peak to rms so we can make the same recording appear louder by increasing the rms value with respect to the peak value. 

Except there are a few flaws in this approach.

These flaws are variations of the same "problem": the peak value is an instantaneous measurement and this only needs to occur once during the measurement interval to be counted. The rms power is an average which means it isn't possible to make an instantaneous measurement, we have to take a number of measurements over a fixed period of time and then calculate the root-mean-square average from those readings. 

For a simple continuous unvarying sine-wave this is does not change as long as the measurement period is a whole number of cycles - the rms average of one cycle is the same as two, ten or a million cycles but if we measure 1½ cycles then we will not get the magic 0.707 answer. However if the number of cycles is large then this error introduced by not measuring a whole cycle is reduced, so after 100 cycles the error is only 0.5% and so on. 

Also, if this sine-wave constantly changes in volume with time the rms value will change and will be less than 0.707. So, for example a linear fade-in will have half the rms value of a constant volume tone but the peak will be the same so the Crest Factor will be "better".

If I create a compound wave by adding together two or more sine waves of different frequencies the rms and hence the crest-factor changes. For example if I add lots of odd harmonic sine-waves together the rms value of each is 0.707 times the peak but the resulting wave is a square-wave and the rms of that is 1 times the peak. So now if I start adding in lots of even harmonics to the square-wave create a sawtooth-wave then that changes the rms to 0.577 times the peak. Therefore the Crest Factor is dependant upon the shape of the wave-form. An over-driven guitar will have a "worse" Crest Factor than a clean guitar tone because an over-driven guitar tone approximates to a square wave.

So, since music is a complex constantly varying compound wave whose rms value is determined by the kind of sound being made, the Crest Factor will be different for different pieces of music. As this music also changes with time this also affects the Crest Factor if changes during the measurement interval.

Another thing that affects this measurement is the kind of music being measured - obviously putting a microphone in front of a folk singer playing an acoustic guitar is going to have a "better" crest factor than an 80-piece orchestra playing Holst, but what if we try to compare that orchestra playing Holst's Mars with the same orchestra playing Jupiter?

So, Ms Swift is louder than Motörhead... or is she? 

Comparing the Dynamic Range of two different pieces of music is pointless and practically meaningless. So putting an arbitrary "goodness" number where everything above the line is "bad" and everything below the line is "good" doesn't mean very much. 

Where it is useful is in making a comparative measurement on the same recording before and after compression. Here it gives an indication of how much the "natural" dynamics of the recording have been affected by the compression but it doesn't tell us how "bad" it has become. As mentioned previously - Loudness is a psychoacoustic effect that is a combination of the psychological effect of sound and the physiology of hearing, so two different recordings can have the same crest-factor/dynamic range and both will seem to have similar levels of loudness, yet one will affect us more than the other. [Empirical measurements only measure the physical/physiological - not the psychological].



Edited by Dean - May 01 2015 at 08:05
What?
Back to Top
Sean Trane View Drop Down
Special Collaborator
Special Collaborator

Prog Folk

Joined: April 29 2004
Location: Heart of Europe
Status: Offline
Points: 19541
Direct Link To This Post Posted: June 07 2015 at 03:13


I am your father!!

Interesting read, Dean Wink
Back to Top
AEProgman View Drop Down
Forum Senior Member
Forum Senior Member
Avatar

Joined: August 11 2012
Location: Toadstool
Status: Offline
Points: 1787
Direct Link To This Post Posted: June 23 2015 at 11:35
Just now now seeing this and another EE approves of this thread. Clap

Great stuff Dean.  Have not thought about "tubes" and calculations of dBs in many decades since school (broke out the calculator and stimulated my brain after reading this LOL).  My world of those decades has been in power quality and conversion.  The "D to A" (bad pun) I deal with is converting DC to AC (inverters) and vice versa.

Most of my listening over the years have been mobile, car and portable stereo systems (I know, low quality), and have not delved much into the hi-fidelity world.  With less time on the road of late, I have begun to have an interest in it and do have a question for you (or others).  You often hear of the "master tapes" or the like when re-mastering is done.  I am curious as to the method or process that studio engineers take to get the music/sound from the musician to a media that it is recorded on and what the media is?

That may be a loaded question and most likely things have changed over the years in the process, but since I seem to see a recent upswing in the market for HD or hi-res audio, I guess a side question from that would be, is any of the "master tapes" involved in the creation of those?

PS-Hope you keep posting interesting articles and knowledge related to all this...
Back to Top
progbethyname View Drop Down
Forum Senior Member
Forum Senior Member
Avatar

Joined: July 30 2012
Location: HiFi Headmania
Status: Offline
Points: 7738
Direct Link To This Post Posted: October 13 2015 at 12:09
Originally posted by AEProgman AEProgman wrote:

Just now now seeing this and another EE approves of this thread. Clap

Great stuff Dean.  Have not thought about "tubes" and calculations of dBs in many decades since school (broke out the calculator and stimulated my brain after reading this LOL).  My world of those decades has been in power quality and conversion.  The "D to A" (bad pun) I deal with is converting DC to AC (inverters) and vice versa.

Most of my listening over the years have been mobile, car and portable stereo systems (I know, low quality), and have not delved much into the hi-fidelity world.  With less time on the road of late, I have begun to have an interest in it and do have a question for you (or others).  <span style="line-height: 1.4;">You often hear of the "master tapes" or the like when re-mastering is done. </span><span style="line-height: 1.4;"> I am curious as to the method or process that studio engineers take to get the music/sound from the musician to a media that it is recorded on and what the media is?</span>
<span style="line-height: 1.4;">
</span>
<span style="line-height: 1.4;">That may be a loaded question and most likely things have changed over the years in the process, but s</span><span style="line-height: 1.4;">ince I seem to see a recent upswing in the market for HD or hi-res audio, I guess a side question from that would be, is any of the "master tapes" involved in the creation of those?</span>
<span style="line-height: 1.4;">
</span>
<span style="line-height: 1.4;">PS-Hope you keep posting interesting articles and knowledge related to all this...</span>


If I may help...sorry to see this thread getting some very limited activity as well lately...
anyhow, the "master-tapes' obviously refers to the original source the piece of music occurred on with its original mix, recording etc. depending on the master tape it could have been recorded in 2,4,6,8,16 and even 24 multi channel recording/ tracking. Now to get the best results, especially for remix or remastering audio purposes it is always best to have the original source or recordings to achieve wonderful sonic reproduction outcomes, which is why so many artists and engineers will do just about anything to preserve or get their hands on those master tapes or original (Non-manipulated ) sources. It's not always easy because preservation of these original master tapes is terribly less than ideal.
Here's actually a great story before I tell you a little bit of what steps audio engineers take when working with master tapes. ;)
I think you are aware of the Neo Prog juggernauts "IQ". Not too long ago (2013) IQ released a total revamping of Tales from The Lush Attic sonically. Originally TFTLA was recorded for lack of a better word 'sh*t.' :(
As a band, they admitted to being rushed and given their circumstances they still released a fantastic album, but the audio quality was incredibly less than ideal so it actually took away a lot of the performance and Total appreciation for what the album was or is musically.
Main guitarists Mike Holmes set out on a personal mission to breathe life back into TFLA by remastering (polishing) and remixing the album, so of course he needed to get his hands on the original source (master tapes) and once he did get the tapes (ironically stored up in an old dusty attic for the better part of 30 years) was mortified to discover that the glue or resin that binds the tape thread had oxidized! This is a bit of a travesty because obviously you cannot remaster or remix an album with out its original source to utilize studio/professional audio equipment to its full potential. You'd be better off just re-recording the whole album. Lol.
Anyway, Holmes and Peter Nicholls had the brilliant idea to heat the tapes in an oven at a very very low heat to allow the glue to be less bonded to the tape itself, so they could get at the thread. The process worked but it took 3 days of slowly baking off the resin...a lovely TFTLA cake finally ready for studio console use!

So by IQ being able to have the original master tapes they can now have the opportunity to 'clean up' the album without having to deviate, manipulate or take away from the original source recording too much. ;)
A sound board or console that has the ability to section off or use a tracking channel system can be an audio engineer/producers best friend.

I guess this brings me finally to your question. The use of compressors, limiters, faders etc are put into play to achieve their maximum threshold settings without trying to distort or manipulating ideal frequency responses with in any given section or tracking you are focusing on. I believe TFTLA was recorded originally as a 16 audio tracking format, meaning 16 different sections or microphone placements were used to record the album.
Now where a good audio engineer comes in, is as if he/she can manipulate certain frequencies with in a given section of music he/she chooses to focus on using the right sound equipment, however you need the original source to have the best success at this. Think of a master tape like having an original blue print or reference point to which later on you can successfully add on or subtract to make tastes more suitable or in the musical sense "cleaner" or more natural sounding.
All in all, mixing tolerances with in a song are greatly explored to achieve a more ideal, different and sometimes a flat out better sound during mastering processes. It's time consuming and finicky but I feel TFTLA is a really good example where remastering and remixing principals are very valuable in the restorative audio preservation world.
They do serve a purpose and the bending and manipulation of sound frequencies in a section or sub section Of an audio recording is a rendering process that is greatly necessary. TFTLA is a great example of this.
Obviously there are many alternatives to the view of remixing and remastering albums and a 'negative' response can be as such. Some audio engineers are not very good and obviously the overuse of compression and limiters and poor mixing can result in an audio disaster no matter if it's a 16 or 24 bit depth recording.
It's easy to get carried away, but I think someone like Steven Wilson is another good example. He remastered and remixed several classic Prog albums without deviating too much from the original master tape sources he has to work with. I think checking out Jethro Tull's "passion Play" is a prime example to show how an audio engineer can make an older album recording sounding new and fresh without killing it by making "louder" or sounding harsh because the treble has been pushed way out of wack to sound cleaner.
Cleaner or clearer is not always better. If you overly compress a source you will kill a ton of the natural tonality of what a particular instrument is supposed to sound like. This is more commonly referred to as having "digital edge.'
Steven Wilson doesn't do this to his remastered versions. He remains very faithful to the original recording and just usually changes the mix of the album very sparingly to have it sound more fluent and balanced.

All I can say is that I love the subject behind the art and science of audio recording but I never would want to do it for a living. I just don't have the social skills for it. Lol.

Hope this answered some of your questions. It's a bit difficult to explain on a nitty gritty technical level, but the right sound console(board) to work with the original recording is the basis and nowadays it's all done on computer anyways. There is some very sophisticated computer software out there these days. KONTACT is one of them.
Overall I'd say a lot of people can play the role of being an audio engineer, but to say that mostly can do it well is another story indeed.
Gimmie my headphones now!!! 🎧🤣
Back to Top
Roxbrough View Drop Down
Forum Senior Member
Forum Senior Member
Avatar

Joined: December 02 2012
Location: Yorkshire
Status: Offline
Points: 100
Direct Link To This Post Posted: November 12 2015 at 11:09
To add my two penneth I have just bought myself a valve amplifier and now reaslise I've been wasting years listening to solid state. Both on CD and LP music is a quantum leap better with valves. Has anyone else found this to be true? 
Live Long and Prosper
Back to Top
Dean View Drop Down
Special Collaborator
Special Collaborator
Avatar
Retired Admin and Amateur Layabout

Joined: May 13 2007
Location: Europe
Status: Offline
Points: 37575
Direct Link To This Post Posted: November 12 2015 at 11:25
Nope. Different but not "better". Subjectively "I like it more" but objectively not "better". Wink
What?
Back to Top
Catcher10 View Drop Down
Forum Senior Member
Forum Senior Member
Avatar
VIP Member

Joined: December 23 2009
Location: Emerald City
Status: Offline
Points: 17451
Direct Link To This Post Posted: November 12 2015 at 15:59
Originally posted by Roxbrough Roxbrough wrote:

To add my two penneth I have just bought myself a valve amplifier and now reaslise I've been wasting years listening to solid state. Both on CD and LP music is a quantum leap better with valves. Has anyone else found this to be true? 

Cool! What amp did you get? 
My DAC has a tube buffer stage as well as SS out and Digital out, I much prefer the tube buffer stage output when listening to my digital files.
Back to Top
progbethyname View Drop Down
Forum Senior Member
Forum Senior Member
Avatar

Joined: July 30 2012
Location: HiFi Headmania
Status: Offline
Points: 7738
Direct Link To This Post Posted: November 12 2015 at 17:18
Originally posted by Roxbrough Roxbrough wrote:

To add my two penneth I have just bought myself a valve amplifier and now reaslise I've been wasting years listening to solid state. Both on CD and LP music is a quantum leap better with valves. Has anyone else found this to be true? 


Gotta agree with Dean on this one. It's apples and oranges. 'Different, not better.'
You can find yourself in a situation where 'sonically' you may just crave different things.
I personally love neutrality, dead accurate and analytical sound preferences/settings so I opt for solid state with a CLASS A circuitry design. However, with tube amps sound reproduction is usually less neutral and more coloured in some areas and maybe heavier on the bass side of things. Does that mean that tube amps because they are usually less neutral sounding are worse interns of SQ? Nope. Absolutely not. It's just different.
Like (catcher 10) I'd love to hear with tube amp you got. :)
I was thinking about getting the WOO AUDIO FIREFLY with the tube PSU but that amp just didn't have enough juice cause the VRMS was too low and couldn't give the headphones I really love the proper power they needed. Nice amp and actually quite neutral sounding surprisingly but the firefly just couldn't push my headphones. :(
Anyway, love to hear about the amp. ;)

Prog on and enjoy!!
Gimmie my headphones now!!! 🎧🤣
Back to Top
 Post Reply Post Reply

Forum Jump Forum Permissions View Drop Down



This page was generated in 0.168 seconds.
Donate monthly and keep PA fast-loading and ad-free forever.