Why were CDs recorded in 16-bit/44.1khz?

Discussion in 'Audio Hardware' started by MZ_RH1, Feb 5, 2017.

  1. Tim Lookingbill

    Tim Lookingbill Alfalfa Male

    Location:
    New Braunfels, TX
    What are you talking about? How does that have anything to do with this topic?
     
  2. harby

    harby Forum Resident

    Location:
    Portland, OR, USA
    @john morris has filled several pages of several threads with 75% his posts. Kind of bad forum etiquette.
     
  3. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario

    I agree. The studio burnt CD sounds the
    same as the 16/44.1 on our hard drive as well. I am sorry that I didn't make that clear. I was referring to the CDs we get back don't. I don't know why that is. It's a mystery to me. And I did say that they were bit for bit perfect. So nothing wrong with the CD.
     
  4. What do you think is in those extra 8 bits?
     
    john morris likes this.
  5. Tim Lookingbill

    Tim Lookingbill Alfalfa Male

    Location:
    New Braunfels, TX
    You didn't answer Black Elk's main question, again!.

    I'll ask it again...Why didn't you ask the manufacturer that made the CD's at the factory why they didn't sound the same as your studio burnt CD?
     
  6. enfield

    enfield Forum Resident

    Location:
    Essex UK
  7. Tim Lookingbill

    Tim Lookingbill Alfalfa Male

    Location:
    New Braunfels, TX
    Now I understand. Thanks for making that clear.

    What's funny about that article you posted is their explanation showing slight anomalies to single waveforms within playback equipment processing that implies it improves the sound or changes it in some way. But when I zoom in on a cluster of CD standard resolution waveforms in Audacity spread across a 27in. diagonal display screen I can see all the bumps, notches and wrinkles that look worse than what's shown in the article. But when I try to play back that section I get no sound and that's because I haven't selected enough peaks and valleys to make a sound.

    HILARIOUS!

    CD Standard resolution waveform. All those sample points aren't enough to make a sound...

    [​IMG]
     
    Last edited: Oct 3, 2019
    Stanley1970 likes this.
  8. Stanley1970

    Stanley1970 Well-Known Member

    Location:
    North America
    Thank you for sharing the article, interesting read this morning!
     
  9. harby

    harby Forum Resident

    Location:
    Portland, OR, USA
    The anomalies depicted in the third page are the preamp in the Yamaha - switching on and off +12dB gain to go between 16 bit and "18 bit" modes while using a 16 bit DAC, and how they mitigated analog impulses that require the gain boost to be accurate to .000001 or so.

    While CD only has 16 bits of audio stored on it, if then oversampled, the interpolated points (which are a 100% correct representation of the audio, BTW), theoretically have unlimited bits, or bits limited by the floating-point math accuracy of the processor.

    Modern DACs do CD audio at 24 bits; 5MHz+ internal multibit; without breaking a sweat, though.

    I said in another post it would take 31 4k TVs stacked on top of each other to render all possible levels of 16 bits in your waveform view; what you see on the monitor doesn't relate to digital audio in traditional ways.
     
    Last edited: Oct 3, 2019
  10. vwestlife

    vwestlife Forum Resident

    Location:
    New Jersey, USA
    There were some early experiments back in the 1950s doing that, and there were even some receivers back then made with separate AM and FM tuners that could be used at the same time to facilitate that. Here is a video of such a beast in operation:
    But that idea was quickly abandoned once multiplex FM stereo was approved by the FCC in 1961.

    AM Stereo is transmitted using the Motorola C-Quam (Compatible Quadrature AM) system, which uses conventional amplitude modulation for the L+R (mono) component and quadrature modulation for the L-R (stereo difference) component, on the same signal.
    [​IMG]
    Yes: Template:AM Stereo radio stations in the United States - Wikipedia
    Plus some in Canada and Mexico.
    The big problem is that there were four competing systems for transmitting AM Stereo when it was first introduced. The FCC declined to choose one as the single standard and said "let the marketplace decide", which turned out to be a disaster. Nobody wanted to invest in a system only to have it be the loser in this race. (Can you imagine if Sony and Philips had released competing and incompatible CD formats in 1982, instead of working together on a single format?)

    Canada adopted C-Quam as its single standard for AM Stereo in 1988, but the FCC didn't until 1993, by which time many AM stations had switched to talk formats and nobody really cared about AM Stereo anymore.
    But most of what we perceive as bass comes from the upper harmonics, not the fundamental. A pure 70 or 80 Hz tone would just sound like rumble, with no musical timbre to it.
     
  11. BruceS

    BruceS El Sirviente del Gato

    Location:
    Reading, MA US
    Not for me. Kind of the opposite. While I do have a ~12k cutoff, that doesn't mean everything below that is patent. Some days are better than others, for sure, but I have spent enough time by now trying to figure out what settings/configs to use to improve sound, when the issue was me. I'm sure years spent on the telephone (headset) providing support didn't help.
     
  12. BruceS

    BruceS El Sirviente del Gato

    Location:
    Reading, MA US
    Shannon made #10 on the Science News "Top 10 revolutionary scientific theories."#1 was Copernicus' "Heliocentrism," c. 1543. Albert E.'s Relativity was only #4. FAIK, it too works only under ideal lab conditions. Anyway, Mr. Shannon's theory is in elite company.
     
  13. Doctor Fine

    Doctor Fine "So Hip It Would Blister Your Brain"

    I haven't compared 24/192 to 32/384 on my own system so I can't say with first hand listening experience.
    If I had to guess ---the greater bit DEPTH would probably present as a thicker more fully fleshed out sound of instruments and voices. The higher 384 khz sample rate would probably improve the tone of the upper frequencies, increase stereo separation in general and help clear up the soundstage.
    These are the same areas of improvement I can hear happening as I move UP from CD at 16/44.1 to on line Qobuz streaming at 24/192.
    I am so happy that digital is finally sounding great.
    CD always worked OK in a mid-fi sort of harsh fuzzy way...
    I use a Benchmark DAC and a Bryston DAC for digital and they both struggle to make CD sound very musical.
    CD just hasn't got the beans to put together an adequate signal replication, I'm afraid.
    What do you want from a 35 year old computer format.
    35 years is too long a run for ANYTHING digital, isn't it?
    But now that digital is finally finding a mass market "upgrade" in sight I hope the trend continues.
    All us music lovers are all way overdue for a FORMAT upgrade and now that folks are paying attention to the mastering AND the format it is mastered to---well, let's just say Steve Hoffman has some WORK to do!
    Let's hear MORE high res!
    I love how it makes my classical recordings sound clearer and more life like with much better separation, clearer treble and more "life" in the sound.
    And I hope everybody trying to shoot this progress down can be the minority.
    We lost SACD already.
    The MINIMUM I will accept at this point is 24/192.
    As should EVERYBODY that calls themselves a music loving audiophile.
    I really thought my Denon DL103r phono cartridge with ruby cantilever and fine line stylus was the shizzle.
    UNTIL I heard the same music at 24/192 and it sounded even MORE analog than my phonograph record version.
    Better separation.
    Same identical great tones (unlike CD which hardens everything).
    It was an easy 15% upgrade from the best phono sound I have heard.
    Maybe I now NEED a five thousand dollar Lyra cartridge or something to even up the playing field?
    I can't tell you what a shock it is to finally have TWO high resolution formats.
    LP AND 24/192!

    My two cents.
     
    Kiko1974 likes this.
  14. Black Elk

    Black Elk Music Lover

    Location:
    Bay Area, U.S.A.
    I'll try to keep this brief:

    1. I have spent the last 25 years involved in the development and introduction of DSD/SACD, high-resolution digital signal processing, a DSD-based DAW and other audio-related projects. My work has taken me to many of the world's most famous recording/mastering studios.

    2. I never said a master clock could mess things up, I responded to your claim (the use of a MASTER CLOCK is essential to avoid clicks, jitter, pops and all sorts of digital nasties) to point out that the addition of an external master clock does not necessarily improve jitter performance.

    3. my test was carried out either in the studio of a famous recording/mastering engineer or in a quiet room at several audio shows. Listeners included studio professionals, audio journalists and audiophiles. Many hundreds of people in total.

    4. some of the professional listeners were responsible for the titles many on this Forum rave about. They are all familiar with high-resolution sound quality!

    5. you stated that manufactured CDs sound different to your CD-Rs and reference files! I asked whether you had raised that issue with the pressing plant, and then asked how you were doing your comparisons. You avoided answering both.

    6. you used the term 'reference' to describe your converters. As someone who knows many of the best converter designers, and has been hands on with lots of product, I was merely pointing out that the term 'reference' is subjective, but that any converter which could be considered 'reference' grade should be source agnostic thanks to its jitter immunity.
     
    nosliw, GreenDrazi, jbmcb and 2 others like this.
  15. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario

    And yet a lot of musicians are mixing down to 32 / 384. 24/384 is now used as an archive standard for reasons which are not mine. And the so called 32 bit converters are impossible and yet some converters are called 32 bit. It is just adding 8 more bits. There is a belief that everytime you: EQ , edit, or whatever that somehow bits are lost and that affects sound quality. Again not my opinion. And I don't even work at the studio anymore.
     
  16. missan

    missan Forum Resident

    Location:
    Stockholm
    I don´t think that is two cents.
     
  17. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario
    First thank you for your civilised reply. I wish others were so capable of mature post behavior.

    I have been very tired over the last few days and I probably mixed everyone's posts up. I must have seen bizzare.

    Did not realize you were in the business of designing. Or maybe you did but I missed it.

    Well said . No argument there. I will let my Uncle speak as my views here (by some) are not considered valued. So reply is not possible. He is too busy and isn't even a member.

    "....Hello people. My name is Stephen Jack Morris I have been working in one way or the other in the mixing /mastering field since 1968. I used to cut records for RCA back in the 70's. It is good to know that DAC development has not stood still. Question Mr. Elk: We need at least 62 channels of A/D conversion. How much would such a converter cost that you are talking about in your post? I think it is time for an upgrade at some point. I am very impressed by your story. Since you are a DAC designer I trust what you say.
    My nephew although well meaning does not speak anymore for Seven Nations Studio. And I am not him. Please excuse his vigor and vim. But as you know they are a lot of as you say "armchair audiophiles" (not you of course) but some have some funny beliefs.

    As regards to the factory CD not sounding exactly like the master 16/44.1 file it is strange yes. I have made calls. I have gone down to the plants myself. And all they can say is, "They don't know." This is not something you would hear on monitors .
    It is when you put on mastering grade headphones (I won't say reference Mr. Elk ) that you notice the soundstage is reduced. Most people would not hear it. This is a known phenomenon in the industry. Less so then say 25 years ago but it does happen. But many of us have come to expect it. Keep in mind many engineers will deny that this happens. The math does not add up as you say. Again, not always but sometimes. Again my nephew over states the problem. But we are working with the CD plants to fix whatever the problem is

    Any information you can send me. Or perhaps a link would appreciated. Thank you again for your time. ..
     
    Last edited: Oct 3, 2019
    enfield likes this.
  18. Black Elk

    Black Elk Music Lover

    Location:
    Bay Area, U.S.A.
    Firstly, I'm not a DAC designer, but I know a bunch of them.

    For multi-track ADCs one has to consider budget, size, power consumption, I/O, supported formats, etc. Do you want the same quality on all channels as a stereo mastering converter, or not (that can increase the cost significantly)? Companies like dCS and EMM Labs seem to have dropped out of the Professional arena, which is a shame, as they made excellent converters, though none bigger than 8-channels in a unit. At the moment, I would be inclined to look at:

    Prism - Prism Sound - Recording and Production

    Merging Horus - Merging Technologies | Horus Mic Pre & AD/DA Converter with RAVENNA & AES67

    Mytek - https://mytekdigital.com/professional/

    Merging has the most options, especially for large numbers of channels. I don't think you could go wrong with any of the above. The number of options for ADCs is significantly reduced compared to DACs. Legacy companies like Apogee are still in the business, but I don't know of any high-end studio still using their converters.

    Of course, AVID/Digidesign have their own 24-channel converters which a lot of ProTools owners use.
     
    john morris likes this.
  19. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario
    Thank you. Your post said for the past 25 years you were involved in the development of DACs. I assume that meant designer. We both assumed. Mistake.

    I know longer work at Seven Nations Studio but I will pass it along. Our 98 channel DAC was custom built by Bur-Brown 8 years ago. I am told there is no other like it. Linking was an option but for reasons too long to go into we didn't.

    Thank you for your information. I will email to my Uncle.
     
  20. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario
    Yes very true. You are referring to what we mastering engineers (ex) call the Cloud.(90 - 250hz). It is where all the detail of the bass is. And yet most modern mastering engineers severely reduce it. If you guys had time I could tell you some stories. But remove those last two octaves 20 - 80 hz and the music is incomplete and unsatisfactory. Low bass is felt in the body. It is part of the music. Go to rock a show and filter out the bass below 80hz. You won't feel the kick anymore. And the low B on a 5 string bass is 31 hz. It's upper harmonics are not just in the Cloud.

    Four different standards for AM STEREO RADIO. AHHH!
     
  21. You import a 24/96 file into pro standard editing software and the software automatically sees it as a 32 bit file.
    The additional 8 bits are an automatic bolt on.
    There's mastering recovery information, but no additional musical information.
    I wouldn't be surprised to find out that DACs capable of handling 384kHz sample rates just add 8 to the 24 in the display, to give the owner of the DAC the satisfaction that it's a step beyond 24/192 :yikes:
     
    Last edited: Oct 4, 2019
    domfjbrown, john morris and enfield like this.
  22. vwestlife

    vwestlife Forum Resident

    Location:
    New Jersey, USA
    Not exactly. It is converting the audio from 24-bit integer to 32-bit floating point to make it easier to edit. One of the main advantages of floating point is that it gives you loads of headroom above what would normally be 0 dBFS, with no chance of clipping. So if the recording or mixing engineer gets sloppy with their levels, no damage is done.

    32-Bit Float Files Explained » Sound Devices
     
  23. Carl Swanson

    Carl Swanson Senior Member

    Sorry to hear about your layoff. Here's to better days!
     
    john morris likes this.
  24. Tim Lookingbill

    Tim Lookingbill Alfalfa Male

    Location:
    New Braunfels, TX
    Very informative article. Thanks for posting.

    Questions... do sound cards' DAC need to be 32bit floating point to hear this non-clipping amplitude advantage? And if this sound card doesn't have a high bit FPU will it result in the same kind of clipping distortion when editing? When I work on brickwalled 16bit/41Khz CD files that have been upconverted to 32 bit FP/41Khz in Audacity I can only go so far with the 32 band EQ before it distorts with crackly noise and loss of detail.

    I'm wondering what I should set my MacMini's MIDI headphone out bit depth and sampling rate at since it's adjustable in Device Settings. Thanks again for your input on all this.
     
  25. Black Elk

    Black Elk Music Lover

    Location:
    Bay Area, U.S.A.
    I wouldn't pay attention to that article.

    Firstly, PCM uses signed integers, so the numbers are in 2's-complment form (Two's complement - Wikipedia ) and range from -32,768 to +32,767 in a 16-bit word.

    Secondly, audio signals are bipolar, so the loudest signal is -32,767 or +32,767 since they are equal loudness (we will avoid the 1 value offset for negative numbers here).

    Thirdly, the lowest value (all zeroes) represents no signal, not the noise floor.

    Fourthly, when switching between -1 and 0 (decimal), ALL the bits toggle as you switch from 1111111111111111 to 0000000000000000, which can be a source of switching distortion.

    Fifthly, the signal-to-noise ratio for an n-bit PCM sample is given by:

    6.02n + 1.76 dB (https://www.analog.com/media/en/training-seminars/tutorials/MT-229.pdf)

    or 98.1 dB for a 16-bit system.

    Sixthly, floating-point is required where you have a large number of channels with widely different signal levels which need to be processed and combined (mixed), and you are limited by DSP processing power. For audio signal processing, fixed-point arithmetic is best, but is more costly. This is why on a recent project we used 64-bit fixed-point arithmetic to do EQ processing!
     

Share This Page

molar-endocrine