• Welkom bij Audiofreaksforum, Audio, Hifi, Luidsprekers, Buizen. Graag inloggen of registreren.
 

Interessant artikel over cd's, compressie etc. (Let op; Engels)

Gestart door Sytse Fliege, 5 mei 2022, 11:09:31

Vorige topic - Volgende topic

Sytse Fliege

Bron: https://beocentral.com/beomic/whats-happened-to-cd

What's happened to CD?
17th January 2010 by Tim Jarman

Do you remember the first time you heard Compact Disc? Compared to the cheap record players and misadjusted tape decks that formed the bulk of the audio equipment in use up until the end of the 1980s it was a revelation, with its clean bright treble, absent background noise and the way that loud passages of music seemed to leap out of the loudspeakers.

One undeniable advantage that CD has over the various analogue audio formats is its dynamic range: the difference between the quietest possible sound that can be discerned above the medium's natural background noise and the loudest sound that can be recorded without excessive distortion. CD can in theory manage 96dB, in comparison to the best cassettes at 70dB and LPs at around 50dB.

The dB (decibel) logarithmic scale is used by engineers to express the difference between two quantities that vary over a wide range. When discussing the magnitude of audio signals all that you have to remember is that the signal halves (or doubles) with every 6dB of change. The dB is a relative unit, so one needs a reference value to make any sense of it. In terms of CD players 0dB is the level at which all 16 bits in the digital code on the disc are at "1", the highest value that can be represented and therefore the maximum output level of the DAC (digital to analogue converter) in the player. In CD terms, 0dB (strictly speaking 0dBfs, or "full scale") is a "hard limit" that it is not possible to go beyond, in contrast the 0dB of a cassette recorder is an arbitrary signal level which long ago represented the point at which, with a certain type of tape, the amount of distortion present was 2%. With modern tapes and heads one can easily record beyond 0dB without generating 2% distortion. +4dB is not unusual when using metal tape and Bang & Olufsen's own HX Pro recording system.

The way that distortion occurs is another difference between analogue and digital audio systems. An analogue system in theory suffers from less distortion the smaller the signal it has to handle. Factors such as noise and the non-linearity of some types of amplifier work against this simple rule, but it is still valid as a general case. Digital systems are the opposite, because the resolution of the system falls as fewer of the available bits are used to represent the signal. Thus it follows that distortion becomes worse, in fact performance improves the more the signal is expanded, right up until all the bits are at "1" when suddenly it becomes horrendous. So bad is this effect that for years CDs were recorded at a fraction of their maximum possible level, under the strict guidance of the few CD pressing organisations.

Initially the advice was that the bulk of the music should be at a level no higher than -18dBfs, that is, using only 13 of the 16 bits of digital capacity. The reason for this was to allow transients, like the crash of a cymbal or the snap of a snare drum, to be reproduced accurately and completely without distortion. The trade off was of course that the overall level of distortion present in the rest of the signal was worse than necessary, giving some early CDs their characteristic steely, edgy sound. The average level was soon revised upwards to -12dBfs, doubling the available digital resolution for the bulk of the programme content but still leaving a possible fourfold increase in signal level for the loud bits.

Under these rules, the best use was made of the strengths of the compact disc system, music was crisp and clear with brilliant, breathtaking transients which no other format could match. If you look at the specifications of an early Bang & Olufsen CD playing system like the Beosystem 5000 or 5500 you will find that even though the CD player and the cassette recorder are both line level sources the cassette recorder typically produces an output signal of around 500mV RMS at 0dB where as the CD player produces 2V RMS at 0dBfs. This is not an accident, it simply reflects the fact that most of the time the CD player should ideally be playing material that has an average level of -12dBfs, a quarter of the maximum value of 2V, which is of course 500mV. Therefore when changing between the two sources the listener would for most of the time hear no change in volume, a desirable state of affairs.

These rules prevailed for all of the time that CD was mainly a format used in large (and expensive) hi-fi systems owned and used by experienced listeners. However, as cheaper players, portables and in-car models appeared the CD went from being exclusively high-end to become a mass market product. The needs of the users of these new types of player are different to those of the serious audiophile, so the way in which CDs were made also started to change. It has long been known that people respond more favourably to music that is played loudly than to music which is soft. The non-linearity of the ear in the frequency domain plays a part in this, the listener's perception to both the high and the low frequency extremes improves greatly with increasing sound level and a lot of what makes music interesting and satisfying can be found in these ranges. Of course any CD can be made to sound louder by simply turning the volume of the amplifier up, but there is a commercial (as opposed to a technical) advantage to be had if the discs themselves simply sound "louder" without having to do this. CDs with a compressed dynamic range can also sound clearer in car players where background noise would otherwise drown out the quieter parts of the music. The last thing a driver wants is to be startled by brief, shatteringly loud passages that suddenly rise above this level.

Remembering that the average level on a CD was set to record transient sounds accurately and that the maximum level that can be recorded on a CD is fixed at an absolute point that cannot be changed, obviously something had to give. The answer was to compress the transients and raise the average level of the rest of the music. This certainly gave a sound that was loud but the dynamic range of the signal is actually less. Unlike raising the volume control to make a quiet CD loud, there was no way that the listener at home could recover the lost information. Music without vivid transients is bland and dull, it's difficult to put one's finger on exactly what is wrong at first but it is instantly obvious that something is missing. Even this wasn't enough in the quest for more volume; some producers pushed the level right up to 0dBfs and beyond. Of course with CD there is no beyond, the peaks in the music crashed into the 0dBfs barrier and could go no further so they became flat topped, loosing most of the information that had been contained within them and giving a harsh, distorted sound that could be really quite unpleasant to the serious listener.

Such practices were at their peak at around 1999 and although the majority of producers have pulled back from the brink there is still a great deal of processing that takes place when a CD is mastered in order to improve how the discs sound on cheap portables and car players. To those with high quality equipment this is a nuisance as it spoils the enjoyment of music that could be obtained if only the original standards were observed. Pop records aimed at teenagers tend to be the worst, although "digitally re-mastered" albums that appeal mostly to older listeners can be just as bad, often the earlier CD releases are preferable if you can find them. Amazingly, searching out old CDs is becoming increasingly popular among some of the hard core parts of the audiophile community. Classical music has been largely left alone by studio tricks and really benefits by CD's huge dynamic range, a well produced disc played on one of the better Beosystems can still be an absolute joy.

The CD system is under pressure from those who profit from persuading you to regularly replace your equipment and everything you play on it with "the next big thing". Even though one of the reasons given is the limited 16 bit resolution of what is now almost a 30 year old format, don't believe a word of it. Try to hear CD at its best before you go the trouble and expense of buying all that music yet again.
Ach, kikker toch op!

gslikker

Leuk artikel. Heb er nooit bij stilgestaan.

Al valt het natuurlijk wel eens op dat muziek van diverse cd's ronduit slecht klinkt, natuurlijk.

Maar als het mij al opvalt, is het dan vast zo erg slecht, dat dit niet per se aan de range van het medium kan liggen, want techneuten horen doof te zijn  ;)

minplus

Goed artikel. Het geeft precies aan met uitleg waarom ik de vroege CD's slecht vond klinken (metalig) vergeleken met LP's en de latere jaren 90 en lang daarna vaak opnieuw (loudnesswar).
If nothing goes right, go left; If nothing's left, go right

Vincent Kars

CitaatInitially the advice was that the bulk of the music should be at a level no higher than -18dBfs, that is, using only 13 of the 16 bits of digital capacity. The reason for this was to allow transients, like the crash of a cymbal or the snap of a snare drum, to be reproduced accurately and completely without distortion. The trade off was of course that the overall level of distortion present in the rest of the signal was worse than necessary, giving some early CDs their characteristic steely, edgy sound. The average level was soon revised upwards to -12dBfs, doubling the available digital resolution for the bulk of the programme content but still leaving a possible fourfold increase in signal level for the loud bits.

Ik vraag mij af of dit de oorzaak is van de metalige sound.
In mijn oren klonk de eerste generatie CD spelers beroerd, een scherpe klank.
Ik denk dat het vooral de jitter geweest is die dat veroorzaakte.

Ik heb veel opnamen uit het eerste uur, ADD. De bulk is kamermuziek, ze klinken absoluut niet scherp maar hebben een forse headroom.

In de strijd tegen de Loudness war wordt normalisatie toegepast.
Dan praat je over -18 LUFS
Dat zou dan ook zeer metalig moeten klinken.

Tom!

Een goed opgenomen en gemasterde CD klinkt fantastisch. In de popmuziek moet je daar tegenwoordig wel goed naar zoeken inderdaad.
Geluid is een van de meest overschatte eigenschappen van muziek.
--- E. Schönberger

minplus

Citaat van: Vincent Kars op  6 mei 2022, 12:15:59
CitaatInitially the advice was that the bulk of the music should be at a level no higher than -18dBfs, that is, using only 13 of the 16 bits of digital capacity. The reason for this was to allow transients, like the crash of a cymbal or the snap of a snare drum, to be reproduced accurately and completely without distortion. The trade off was of course that the overall level of distortion present in the rest of the signal was worse than necessary, giving some early CDs their characteristic steely, edgy sound. The average level was soon revised upwards to -12dBfs, doubling the available digital resolution for the bulk of the programme content but still leaving a possible fourfold increase in signal level for the loud bits.

Ik vraag mij af of dit de oorzaak is van de metalige sound.
In mijn oren klonk de eerste generatie CD spelers beroerd, een scherpe klank.
Ik denk dat het vooral de jitter geweest is die dat veroorzaakte.

Ik heb veel opnamen uit het eerste uur, ADD. De bulk is kamermuziek, ze klinken absoluut niet scherp maar hebben een forse headroom.

In de strijd tegen de Loudness war wordt normalisatie toegepast.
Dan praat je over -18 LUFS
Dat zou dan ook zeer metalig moeten klinken.


Ik denk dat de spelers een (groot) deel van het probleem waren, want pas de Cambridge CD2 klonk in mijn oren prima zoals een goeie platenspeler, zonder die metalen rand (en zonder overdreven laag, zoals de Marantz).
Als ik nu de CD van Grace Jones uit 1986 beluister (Inside Story) dan klinkt die voortreffelijk op de gemodde 104 en 204, zoals ook Sinead O'Connor (The lion and the cobra) uit 1987 waarop ik CD-spelers testte toen de CD2 net uit was. Hij was helaas duur maar liet alle andere die ik beluisterd had ver achter zich.
If nothing goes right, go left; If nothing's left, go right

Luistervriend

En vergeet de invloed van de dacs van betreffende spelers niet, die bepaalde eigenschappen van de cd wel of niet boven halen.

Maar eigenlijk de hele keten, van opname tot en met weergave, incl. akoestiek.
Kortom, alles is zo goed als de zwakste schakel.

Voorganger en collectant van onze Overtuiging Belijdenis  Gemeenschap (OBG)  "Het Laatste Woord"
Zoals het al zegt, wij creëren dagelijks een nieuwe realiteit, zoals AI, onze Schepper, ons inspireert.

Aanmelden kan via PB. 🤪

minplus

Geen idee of het de Dacs waren, maar tot die CD2 regelmatig CD-spelers beluisterd en vond ze niks. Moet bekennen dat ik de Flipsen uit die tijd toen ze uitkwamen niet beluisterd heb. Maar in '87 een hele serie beluisterd en de Cambridge was de enige die ik goed vond.
Was trouwens iets anders dan de huidige firma; meer een paar ontwerpers in een garage of zo.
If nothing goes right, go left; If nothing's left, go right

Tom!

Ik heb mijn relatief goedkope Sharp CD-speler in de jaren '80 een tijd aan de set van mijn vader gehangen (ik was de eerste in huis met CDs en mijn vader had een goede set) en we waren allemaal diep onder de indruk van de geluidskwaliteit. Het verschil tussen de diverse opnamen was wel heel duidelijk en sommige van de eerste CDs waren niet om aan te horen. Met name de modernere jazz blonk toen uit in kwaliteit.
Geluid is een van de meest overschatte eigenschappen van muziek.
--- E. Schönberger

Heintje

Ik denk dat van alle klank bepalende eigenschappen in een CD-speler de jitter echt een gehyped modewoord is van de laatste tien jaar.

Motivatie: als je kijkt wat er in een CD-speler allemaal meespeelt in de klank aan digitale, analoge, voedings en mechanische technieken, is het uiterst kleine beetje tijdsvariatie in de sampling / reconstructie zo ongeveer de kleinste factor die je kunt bedenken, véél kleiner van invloed dan ingebouwde oversampling, eenbit vs. multibit en wijze van implementatie van brickwall filters zijn.
Het is wel één van de zeldzame van invloed zijnde factoren die je gewoon goed en betrouwbaar / herhaalbaar / vergelijkbaar ("kwantificeerbaar") kunt meten en ik denk dat juist dáárom zowel journalisten als fabrikanten "er helemaal weg van zijn".

metalfreak


minplus

If nothing goes right, go left; If nothing's left, go right

Regenpak

Probeer eens zo'n transient in een groef te snijden van een LP. Gaat je niet lukken. Dan maar de beperking van 0dBFS. Kom ik nog altijd mijn bed voor uit. En dat zal ook wel zo blijven. De kans dat ik de verbetering ga horen van betere DACs dan die van mijn geliefde CD100/30 is nul. Zeker met het klimmen der jaren...

A propos jitter: ik kreeg hiermee te maken toen ik met een PICkie een precisiesinus wilde maken. Ging me ook niet lukken:



Die pieken zijn het gevolg van jitter. Bij een SPDIF signaal is die irrelevant omdat de data ingeklokt wordt.

Tjerk
It's been Monday all week long!

mediator

Ergens mid jaren tachtig kwamen bij mij via broers en vrienden de eerste geluiden over de CD speler en dat dit superieur was aan vinyl en veel meer dynamiek had. Eerste klanken via dit medium waren uiterst teleurstellend. De dynamiek vond ik geen aanwinst. In die tijd noemde ik het maar een 'constante dynamiek'. Noem het schreeuwerig. Waar dat aan ligt laat ik maar aan de experts over. Wel is er zoals eerder aangegeven ook genoeg verandert in de DA conversie en inderdaad is de manier van het maken van een master voor digitaal en analoog verschillend. Dat zal dan wellicht ook gelden voor wat er gebeurt in de opname studio. Wat ik verder ook wil benoemen is dat onze sets veelal de aanleiding zijn voor wat we horen. En in mijn optiek meer dan die opname.