Results 1 to 19 of 19

Thread: The dynamic range of audio - reproducing the sounds of real life at home

  1. #1
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default The dynamic range of audio - reproducing the sounds of real life at home

    This thread is concerned with dynamic range, by which we mean the loudness range between the quietest and loudest sounds that can be captured and reproduced by a recording/playback system - and related matters.

    Please bear in mind that the dynamic range quoted for a recording/playback system (example, "120dB dynamic range") is, in practice, meaningless. That's because we listen to audio in real rooms in which there is always noise. A wide replay dynamic range cannot be achieved in a city for two reasons: high ambient noise masks the quietest sounds (can you hear birds twittering in the city?) and the impossibility of playing very loud because of neighbour issues. So, city listening has a more restricted dynamic range than countryside listening for the same equipment and the same music. It could be argued that city listening would benefit from loudness compression - the technique to boost the quietest sounds above the ambient noise threshold and to squash the loudest sounds within a neighbour-friendly maximum loudness .... and that is in fact the trend in recording as we'll read ....
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

  2. #2
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default "What's happened to dynamic range" - a magazine article

    During my archive scanning to PDF, I've uncovered many interesting clippings. This wery accessible article (an excellent example of how to write for a non technical audience) appeared in a pro audio magazine last year.

    I've highlighted what I think are the really important points. One I would like you to consider is, 'don't immediately blame your equipment if you don't like what you hear'. You should check against an armory of recordings that you have confidence in, and they may well be recorduings from the 70s or before, recorded before the 'MP3 generation' started to mould recording balance. Do not for one second think that the commercial recording industry lovingly crafts recordings for the hifi enthusiast! The economic reality is that audiophiles are a commercial irrelevance to the recording industry!

    Article is attached.

    >
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

  3. #3
    Join Date
    Feb 2010
    Location
    germany
    Posts
    57

    Default Check your own dynamic range

    Here is a link to an association where you can download a tool for the measurement of the dynamic range of recordings. the only thing you have to do is donate US$ 30,- http://www.dynamicrange.de/en/our-aim

    i did it and use the tool; it is quite interesting to measure my CDs and compare the results to what i hear from the CD...

    best,
    D

    {Moderator's comment: That's an excellent link. Checking the recorded dynamic range is one thing. But that is not the point Alan was making. What matters is the dynamic range you hear at your ears not what is recorded. What you hear will always be a smaller dynamic range. As Alan explained, it is not possible to listen in absolute silence nor can you replay at real-world maximum levels without cracking the walls or windows. Please consider this point.}

  4. #4
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default The "Loudness war"

    Since the advent of affordable digital signal processing in the 1990s, I think we can say without fear of contradiction that dynamic range compression has become the norm, certainly in the pop music industry. Track by track, case by case, a little or a lot of compression (that's another world for squashing the dynamic range of a recording) is applied.

    How much should be applied? A little or a lot of compression? That depends upon many factors but the ultimate decision is not technical - it's not even artistic - it's made by the producer mindful of selling more records. There are very good arguments for compression; the speakers inside everyone's TV set are tiny and they can't take much power so to reduce the strain on their tiny diaphragms, the TV studio (equally this applies to all broadcast radio) will first limit how loud the signal can be, then bring-up the quiet part so they are audible. That's an entirely legitimate use of signal compression.

    Likewise, if we listen to streamed audio on the internet over plastic PC speakers, there would be a good argument for compression - and on car radio receivers too.

    Like it or not, almost all the music we buy is compressed. Here is a very good example of an audio waveform in its original and compressed versions. Listen and decide. For me, the compressed version sound uninvolving and dull - and sounds dirtier too with less sparkle.

    Video here.
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

  5. #5
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default Too loud, highly compressed = listening fatigue

    Here's another excellent example of why "modern CDs do not sound as good as original CDs".

    I've been saying this here for years (see my analysis of the remastered Beatles here somewhere).

    All summarised in this video here.

    And then this one here.

    The CD (music) industry has an agenda that is completely alien to and disconnected to the quality audio industry. That's a sad fact.
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

  6. #6
    Join Date
    Jun 2009
    Location
    Canada
    Posts
    905

    Default Dynamics and masking of a noisy listening environment

    Quote Originally Posted by delgesu View Post

    {Moderator's comment: That's an excellent link. Checking the recorded dynamic range is one thing. But that is not the point Alan was making. What matters is the dynamic range you hear at your ears not what is recorded. What you hear will always be a smaller dynamic range. As Alan explained, it is not possible to listen in absolute silence nor can you replay at real-world maximum levels without cracking the walls or windows. Please consider this point.}
    Moderator: I think Delgesu has a point, though. Of course, what you hear will always be less than what's recorded, because of the masking effect of the ambient sound in your environment. But environments vary: ultimately, the recording is the fixed limiting factor. If the true dynamic range of the recording is only, say, 10dB, then that recording can never really sound natural, even in an ideal listening environment. Conversely, you may not fully appreciate a high-dynamic range recording if you listen in a noisy environment - but at least the potential is there. It might be useful to know this.

    I imagine that there are other factors that have a bearing on subjectively-perceived sound quality, but surely something approaching a natural (rather than highly-compressed) dynamic range must be one of them.

  7. #7
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default Dynamic range - summary so far

    OK, so where have we got to with the subject of dynamic range? Let's make another tentative step forward. To summarise so far:

    1. By dynamic range (of an audio signal) we refer to the potential loudness scale between the loudest sound that can be captured and/or reproduced by the system and the quietest.
    2. An example would be the piano. We know that a piano can play loud but there is a minimum amount of pressure needed to make the hammer hit the string so there is a minimum loudness below which the piano is, in actuality, silent.
    3. All pop music and I'd suspect most/all classical music has some form of dynamic range compression whereby the dynamic range is deliberately compressed to suit modest replay equipment that would be damaged by a wide (natural) dynamic range.
    4. There has been a trend towards making recordings seem louder as louder is initially more attractive - and they say, sells more v. a quieter recording. This extra loudness is critical to commercial 'pop chart' success. This extra loudness is achieved by increasing the overall level during mastering (of the recording).
      See this video.
    5. There is a maximum loudness that any recording process can tolerate so when 4) is applied, the peaks in the recording have to be crushed to keep the whole recording within the maximum allowable signal. Again see video in 4).
    6. Our replay listening at home cannot take advantage of a super-wide dynamic range because the quieter sounds are masked by everyday domestic and environmental noises, nor can we replay at the full loudness without destroying our speakers and the structure of our buildings.
    7. All and every analogue recording/delivery/replay format has a much narrower dynamic range than digital formats. This is because all analogue systems are bedevilled by hiss at the lower end of the dynamic window and by saturation limitations at the top of the dynamic range. This is not a real limitation domestically, because even the limited dynamic range of analogue is normally sufficient for high quality domestic listening.

    Are we in agreement with this? Next I'd like to look at analogue in more detail.
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

  8. #8
    Join Date
    Feb 2006
    Location
    SW France
    Posts
    84

    Default High dynamics recording ....

    With all of this digital technology available, would it not now be possible for a separate signal to be recorded at the mastering stage, which would allow the re-expansion of the dynamic range of a recording in suitable playback systems. This signal would carry all of the compression and limiting data in a real-time stream which could drive an expanding amplifier, much like dBx used to do in the latter days of analogue? Come on Mr Dolby, you know it could be done.
    Paul

    "If all else fails, read the instructions"

  9. #9
    Join Date
    Sep 2009
    Location
    Malaysia
    Posts
    508

    Default Wide dynamic range sample

    I am not entirely sure if I understood No.1, at the first glance I thought I understood what's dynamic range, but now I may have misunderstood dynamics range that is often talked about in Hi-fi Magazines.

    My understanding of dynamic range is the difference between the loudest and softest signal in the recording. If the softest passage is 20dB and the loudest passage is 100dB I consider them to be wider dynamic range as compared to the same recording with the lowest at 50dB and the loudest at 100dB. Is this correct?

    I am attaching two files. One is Nero Waveeditors image of Telarc Erich Kunzel_ Cincinnati Pops Orchestra - Strauss Jr. (J)_ Explosions-Polka, Op. 43. And short clip of the recording towards the end. I am not even sure if this is on topic. I have never dared to play this CD at the normal loudest level.
    ST

    p.s I am unable to upload about 15seconds of 1.84mb wav file. I am getting file invalid message.

    {Moderator's comment: Please - again - deci-BELL, named after Alexander Graham Bell so out of respect we *always* write capital B for Bell. So we must write dB never db every time. Also your image edited to comply with posting rules (it was not annotated).}
    Last edited by STHLS5; 16-10-2010 at 04:33 PM. Reason: changing db to dB.

  10. #10
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default

    Quote Originally Posted by STHLS5 View Post
    My understanding of dynamic range is the difference between the loudest and softest signal in the recording. If the softest passage is 20dB and the loudest passage is 100dB I consider them to be wider dynamic range as compared to the same recording with the lowest at 50dB and the loudest at 100dB. Is this correct?
    That seems correct to me. But I can see how confusing this can be so I've made a very quick little sketch (attached) which may clarify this a little.

    To explain it:

    (1) This is the theoretical (marketing) position. The recording process is advertised as having "120dB dynamic range". Sounds very impressive, and it is. The starting position is that we must establish a reference point when we are working in decibels because decibels are not absolutes; they are a handy way of expressing ratios. Since we can't define in advance how loud something is or could be, we have to set our reference point at the opposite end of the scale - the quiet end. We can define silence; silence means absolutely no sound at all. For example, in a vacuum (on the moon) there is total silence as there is no air to carry sound. Quoting silence relative to a vacuum is not much use for air breathing humans on earth so several generations ago, engineers decided that there was a minimum amount of motion of air particles battering the ear drums which was just - barely, under optimum conditions - detectable by a human with normal healthy ears. You can think of this sound minima as only as loud as a whisper of a whisper. So that was set as the 0dB on our scale. An air motion below that is, to a human, undetectable so is by definition, silence. At the other end of the loudness scale we have exceedingly loud sounds (over 100dB) which can and will destroy your hearing if you have much exposure to them. Luckily, they don't feature much in normal human activity or mankind would be deaf.

    So (1) represents a really wide usable, theoretical dynamic range between the loudest sound cleanly reproduced and the quietest under laboratory test-equipment conditions. It's not relevant to the real world and simply cannot be achieved at home.

    OK, so that's the lab situation. Now we want to take that wonderful technical system and apply its potential to recording some extremely energetic wide-dynamic music music in a first class hall or studio - eg. the 1812 Overture. We make the recording and then analyse the results giving us (2). Two things we detect. First, the microphones have strained to reproduce those loud canons and simply couldn't move their tiny diaphragms under the bombarding of high pressure air molecules and they have distorted somewhat, so we have lost that end of our dynamic range as it is not linear. Second, at the bottom end of the scale, those microphones and mic amplifiers generated random electronic hiss, plus the air con system in the hall wafted randomly moving air particles over the microphones, neither of which bear any relation at all to the music. Also, there is some passing traffic rumble, some distant aircraft noise, an office toilet flushing and many other quiet noises which have obliterated the bottom 25dB or more of our dynamic range. If there are any micro details in the performance, they are well and truly masked by the hiss. Gone forever. So by this stage we are working with a dynamic range of perhaps 80-85dB from the real-world "silence" to the maximum signal our mics will handle.

    Lets assume that fine audiophile classical recording is not further compressed (that's for later discussion) in the post-production but issued to CD as is. What about when we play it at home? That takes us to (3). Our homes are not silent, especially not in the city. Again, we have random noise which obliterates the finest details, and our speaker/amp etc. may not have enough power to properly reproduce the really loudest parts of the signal. So again, the dynamic range constricts, perhaps down to 60dB.

    Now, interestingly, I'd argue that 60dB is probably enough for normal hi-fi listening, although if offered 70dB or 100dB I'd go for that. But my figure of 60dB is - coincidentally? - about the dynamic range of the LP record which many people greatly enjoy as an audio carrier. So should we chase a super-wide dynamic range for real world high quality audio at home? Would a 192kHz 24bit recording system necessarily produce better sounding audio for home use?
    Attached Images Attached Images
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

  11. #11
    Join Date
    Sep 2009
    Location
    Malaysia
    Posts
    508

    Default Telarc CD - too much dynamic range for home listening?

    The sketch was very useful as I now I understand the subject better. Thank you.

    Quote Originally Posted by A.S. View Post
    ...Now, interestingly, I'd argue that 60dB is probably enough for normal hi-fi listening, although if offered 70dB or 100dB I'd go for that. But my figure of 60dB is - coincidentally? - about the dynamic range of the LP record which many people greatly enjoy as an audio carrier. So should we chase a super-wide dynamic range for real world high quality audio at home? Would a 192kHz 24bit recording system necessarily produce better sounding audio for home use?
    IMHO, I find the so called wide dynamic range recording of Telarc in my previous post wasn't practical to listen at home. Most of the musical content in the cd is much lower than my preferred loudness level at the usual level of my volume control. If I were to increase the volume then the highest peak (such as the explosion sound of pistol firing, ariel bombs) is too scaringly loud. But then thankfully 99% of music need not be in such widest dynamic range.

    Curiously, the dynamic range is more important for cinema where you need to hear a soft whisper and at the same time you need to feel the impact of a jet taking off.

    ST

  12. #12
    Join Date
    Feb 2010
    Location
    England and Cyprus
    Posts
    370

    Default Analogue dynamic range

    Quote Originally Posted by A.S. View Post
    All and every analogue recording/delivery/replay format has a much narrower dynamic range than digital formats. This is because all analogue systems are bedevilled by hiss at the lower end of the dynamic window and by saturation limitations at the top of the dynamic range. This is not a real limitation domestically, because even the limited dynamic range of analogue is normally sufficient for high quality domestic listening.
    Now this comment raises issues for me.

    Not that a digital system has lower noise and higher dynamic range, but that the poorer spec. of an analogue system is good enough for domestic listening.

    I have an early AAD recording of Brendal on the piano. It is lovely, and one of my favourite CDs; but I can hear the hiss of the master tape; and I would prefer not to. Later DDD CDs lack the hiss. I can hear that an analogue recording is inferior in my domestic surroundings.

  13. #13
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default Hiss, hiss removal and dynamic range

    Quote Originally Posted by Labarum View Post
    ...Later DDD CDs lack the hiss. I can hear that an analogue recording is inferior in my domestic surroundings.
    But did you turn up the volume to hear the hiss? That's cheating! That means you slid the narrow dynamic range window of the recording upwards and yes, no surprise that your attention is then drawn to all the nasty stuff.

    I don't doubt that if there was irritating analogue tape hiss the CD production studio digitally processed the hissy analogue master tape to remove it to make it more palatable (marketable) to the CD buyer. The dynamic range of the very best analogue recording tape without Dolby noise reduction is about 65dB, and about 75dB with Dolby both for about 3% harmonic distortion. Digitally signal processing the hiss in post production does not expand the dynamic range; it just conceals it. Once the microtones of the piano are buried under the layer of analogue tape (or LP) hiss they are scrambled by the hiss (which is just another word for random noise) and generally irretrievable as music. In short, the hiss wipes out the smallest details, and once obliterated, they are gone forever. But my point was that many audiophiles rave about analogue, when in the lab, analogue is completely hopeless as a recording/replay medium primarily because of its very narrow usable dynamic range. Those who rate analogue can still hear enough detail to satisfy them, so 60dB or so seems to be good enough considering all the other intrusive noises that restrict the replay dynamic range at home (see my earlier postings here).

    A more pertinent observation would be to replay your Brendle off LP made from the same analogue master tape (the vinyl medium has a similar dynamic range to good analogue master tape) and then comment if you can hear the tape hiss or not. I'd suspect that the record hiss would just about mask the tape hiss. All recordings, digital or analogue have hiss - but at a normal fixed replay level that hiss is rarely intrusive. But if you turn up the volume, you magnify the hiss and it does indeed draw attention to itself. But that doesn't mean that the dynamic range is changed at all.
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

  14. #14
    Join Date
    Feb 2006
    Location
    SW France
    Posts
    84

    Default Producer's toys and dynamic range

    Quote Originally Posted by A.S. View Post
    But did you turn up the volume to hear the hiss? That's cheating! That means you slid the narrow dynamic range window of the recording upwards and yes, no surprise that your attention is then drawn to all the nasty stuff. (snip) All recordings, digital or analogue have hiss - but at a normal fixed replay level that hiss is rarely intrusive. But if you turn up the volume, you magnify the hiss and it does indeed draw attention to itself. But that doesn't mean that the dynamic range is changed at all.
    It's a fact of life though that as recording technology has improved, the temptation by producers and engineers to use more technological toys instead of less, has unfortunately led to the loss of many of the advantages a wider dynamic range reproducing system. I think that only amongst small classical labels will you find uncompressed recordings with as natural a dynamic range as it's possible to reproduce. Pop vocals are invariably compressed whilst being recorded, and even orchestral recordings have the master gain "ridden" to keep the quiet bits audible and the loud bits within the confines of the recording medium.
    Paul

    "If all else fails, read the instructions"

  15. #15
    Join Date
    Mar 2010
    Location
    uk
    Posts
    141

    Default All is not lost? "HD radio sound"

    HD Sound uses improved encoding and higher bit rates to boost quality. The BBC said improvements in sound would depend on the size or quality of speakers or headphones. It said HD Sound also offered a wider dynamic range, accentuating the difference in volume between quiet and loud sounds.
    from: BBC launches HD Sound for online radio

  16. #16
    Join Date
    Feb 2006
    Location
    SW France
    Posts
    84

    Default BBC HD audio - what about the dynamics?

    Quote Originally Posted by weaver View Post
    Wwith all due respect to my former employers, I cannot see how HD sound, which is derived from the same mix as "normal" sound can by itself have greater dynamics. The dynamics are decided mainly at source unless compression is applied AFTER the soundmix. By soundmix I mean the source audio, whether live outside broadcast or disc, or studio performance.

    Enhancing dynamics must start at the mixing desk, all subsequent processing can surely only diminish dynamics. Noise floor is still noise floor and peak is still peak, no matter how many extra samples of it are taken.
    If this announcement means that the HD sound is going to be less compressed in the audio domain, i.e. before the online distribution system's digital coding process, then it is to be welcomed. I listen via the satellite feed which is already at a higher bitrate than the online feed so any reduction in audio compression is a bonus.

    Now where are you Rupert?
    Paul

    "If all else fails, read the instructions"

  17. #17
    Join Date
    Aug 2009
    Location
    United States
    Posts
    73

    Default

    Quote Originally Posted by weaver View Post
    Thank you Weaver. In addition to that link being a citation to read how BBC will address dynamic range in their HD, it is a significant announcement.

    The BBC has announced plans to launch HD Sound - "an extra high-quality audio stream" - for online radio.It will be available for special live events, starting with Electric Proms on the Radio 2 website this month, and for all Radio 3 output from December.

  18. #18
    Join Date
    Feb 2010
    Location
    England and Cyprus
    Posts
    370

    Default Dynamic range of radio

    Quote Originally Posted by Paul G Smith View Post
    With all due respect to my former employers, I cannot see how HD sound, which is derived from the same mix as "normal" sound
    Is this true HD ie 24/96 or better, or is it just an extension of the 320kb/s stream experiment that began with the Proms?

    The 320kb/s stream was very nice and very welcome, but I would not call it HD!

    I listen to a lot of classical music on internet radio, and often a station at or over 128kb/s that is less "loud" will sound more musical than one that uses a higher bitrate but more dynamic compression.

  19. #19
    Join Date
    Jan 2006
    Location
    South of England, UK
    Posts
    4,276

    Default

    Since I first read about the 2010 Proms being available at an experimental higher bit rate, I've been more than a little suspicious of the game plan. First, even though I didn't myself detect any improvement in audio quality (over my PC speakers) I intended (but forgot to) log on to the BBC's site and register my appreciation of the quality on the basis that if the BBC were offering it free, why not take it. The objective design of any market surveys is no casual matter - consultants are very highly paid to build in traps to detect false responses because market surveys do steer business decisions - and as I said, I never got around to looking at or completing their survey. But I was a little surprised to hear the Head of Technology* using the accumulated response as justification that the HD stream would be a 'good thing' and in the same breath admitting that it was only likely to appeal to audiophiles (his word) with sophisticated hifi equipment.

    Now to the rub: at an operational level BBC (and indeed all broadcasters) have one overarching remit. They must control the level - specifically the dynamic range - of broadcasts to ensure that the programme is clearly intelligible by the vast majority of listeners, 99% of whom are listening at home or in car using very low-fi speakers - the majority of which are not in separate systems but in plastic portable radios. The reality is that even on Radio 3, the classical service, the studio engineers must take a second by second interest in the loudness of the programme they are transmitting and either manually adjust the levels (so called "gain riding") or allow some equipment (a dynamic range compressor) to do that. Just look at how many knobs are needed to control the levels and dynamics of a concert hall recording here. Broadcasters absolutely must manage the dynamic range of their output, and by that I mean must restrict the loudness range. That's currently done at source, at the studio mixing desk, so that what leaves the desk on its way to the transmitter/distribution network is fit for transmission, i.e. processed, limited, squashed and of adequate frequency response and distortion.

    If a sub-set of the current audience (the audiophiles, the 1% or less) are to benefit from a greater dynamic range, then the work practice at the mixing desk is going to have to be rethought. There is going to have to be the one conventional stream of audio feeding the 'low-fi' network and another, wider dynamic (etc.) stream feeding certain specialist networks. Just how is this going to work in practice? The first question is, can the mixing desk generate these two streams in parallel, one conventionally processed for the normal audience and another (lightly) processed when conventional mixing desks have thus far been designed to generate the one, master (stereo) processed stream. Will there need to be more sound engineers sitting at the desk looking after the two streams individually twiddling the knobs? Or can it be automated? Or is this a trivial application for a modern all-digital desk? Or, as I fear may be the case simply for ease of implementation, the one conventional compressed stream may be used for HD audio too, so the only difference is that one is delivered at, say, 320kb the other at the normal 128kb rate.

    You'd not seriously expect to hear any significant difference at home on anything other than super hi-fi equipment would you? And pandering to the 1% at a time when the BBC is under enormous political pressure. Best use of public money? I'd say that largely depends on what extra dynamic range we will be allowed to hear at home on the HD service and that has nothing whatever to do with 128, 196 or 320kb delivery bit rate.


    * You can hear the interview here. Please scroll forward to 0:53.30
    Alan A. Shaw
    Designer, owner
    Harbeth Audio UK

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •