why didn't cd players ever re-buffer?

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • pvadbx
    Junior Member
    • May 2006
    • 24

    why didn't cd players ever re-buffer?

    this is mostly a historical question, since i think traditional cd players are a dying breed and I'm installing my transporter tonight, but i'm still curious. why didn't cd manufacturers stick 20-30 megabytes of memory into their machines and rebuffer and re-clock the bitstream to avoid all the sonic problems associated with a mechanical transport? it would be like an SB3, except the network would be replaced with the cd.

    it seems like this would have made some very good sounding cd transports for not a huge cost. am i missing something? why didn't manufacturures do this?

    .phil
  • Pat Farrell
    Senior Member
    • Apr 2005
    • 4294

    #2
    why didn't cd players ever re-buffer?

    pvadbx wrote:
    > why didn't cd manufacturers stick 20-30
    > megabytes of memory into their machines and rebuffer and re-clock the
    > bitstream to avoid all the sonic problems associated with a mechanical
    > transport?


    Because the RedBook CD was designed in 1975 or so, when memory cost
    thousands of dollars per kilobyte.

    I bought my first CD player in 84 or 85.
    When I bought a PC in 1990, I got 5MB of ram, and folks asked
    what kind of fool I was for getting all that unusable memory.

    > am i missing something? why didn't manufacturures do this?


    Because it would have cost too much.
    And 700 MB was an inconceivable amount of storage.

    Plus RedBook was invented to crush casettes, not LPs. At the time, the
    record companies were worried sick about lost revenue from people
    bootlegging albums onto cassettes, and recording directly from FM radio.
    In the olden days, a DJ would say "Coming up, the complete Led Zepplin
    IV" ablum, which meant everyone put a fresh cassette into their deck.

    There are audiophiles that claim that vinyl is still vastely superior to
    RedBook. It might even be true, but I don't care, records are such a
    hassle, and a SqueezeBox, let alone Transporter, is so much more convenient.

    --
    Pat



    Pat
    http://www.pfarrell.com/music/slimse...msoftware.html

    Comment

    • empty99
      Senior Member
      • Jan 2006
      • 153

      #3
      Am curious myself...

      Since they do that on portable CDs to fight the skipping problems. Too late now, computer music storage is the new killer apps for the home audio market.
      SB3>MF3.24DAC>Nak PA7s>B&W N802>Nice!
      SB3->Benchmark DAC1HDR->Pass X250->12g Romex->B&W N802

      Comment

      • pvadbx
        Junior Member
        • May 2006
        • 24

        #4
        Originally posted by Pat Farrell
        pvadbx wrote:
        > am i missing something? why didn't manufacturures do this?


        Because it would have cost too much.
        And 700 MB was an inconceivable amount of storage.
        sure, when CDs were invented, I understand. but say in 2000, why didn't the audiophile CD manufacturers take a different approach. when you are selling a machine for $1K, $2K plus, you can do a lot of things.

        speaking of cost for technology, in a past life, I bought 300 megabyte (not gigabyte) hard drives for the college I was working for. they were the size of washing machines, and cost (if memory serves) about $17,500 each.
        Last edited by pvadbx; 2006-10-13, 01:32.

        Comment

        • seanadams
          Founder, Slim Devices
          • Apr 2005
          • 2879

          #5
          Even better that just having more buffer, and I'm not sure why nobody thought of this for DVDs, would have been to distribute the data in a RAID-like way, with parity blocks on diametrically opposite areas of the disc. So you could completely obliterate any part of the disc and as long as enough bits could still be read elsewhere you would sill get 100% of the data. As it is, everything is still laid down in a continuous spiral (not even concentric circles like a hard disk) and assumes the data is accessed sequentially, with the additional error correction bits right next to the data they're protecting. I should file a patent...
          Last edited by seanadams; 2006-10-13, 03:16.

          Comment

          • pvadbx
            Junior Member
            • May 2006
            • 24

            #6
            Originally posted by seanadams
            Even better that just having more buffer, and I'm not sure why nobody thought of this for DVDs, would have been to distribute the data in a RAID-like way, with parity blocks on diametrically opposite areas of the disc. So you could completely obliterate any part of the disc and as long as enough bits could still be read elsewhere you would sill get 100% of the data. As it is, everything is still laid down in a continuous spiral (not even concentric circles like a CD) and assumes the data is accessed sequentially, with the additional error correction bits right next to the data they're protecting. I should file a patent...
            so true. optical drives are fast enough now, why not treat them more like data drives than real-time data delivery mechanisms. but i once worked for philips nv and got a good understanding of the standards process that went into the cd and dvd formats. you can't even imagine how long it took and the compromises that were made, as well as the interesting politics between the consumer electronic companies (even standardizing the details of MPEG 2 took forever). just look at the HD format wars now. we're lucky we got anything standardized.

            so thank the audio gods for a company like slim devices that lets us get away from all those compromises!

            .phil

            Comment

            • Pat Farrell
              Senior Member
              • Apr 2005
              • 4294

              #7
              Re: why didn't cd players everre-buffer?

              pvadbx wrote:
              > sure, when CDs were invented, I understand. but say in 2000, why did't
              > the audiophile CD manufacturers take a different approach. when you are
              > selling a machine for $1K, $2K plus, you can do a lot of things.


              In the 2000s, they conspired to sell you DVD-A and SACD. Of course,
              since the format wars prevented either from reaching mass market, both
              are dead.

              > speaking of cost for technology, in a past life, I bought 300 megabyte
              > (not gigabyte) hard drives for the college I was working for. they
              > were the size of washing machines, and cost (if memory serves) about
              > $17,500 each.


              Must of been 81 or 82. I remember in 1980 paying $31,000 each for 200 MB
              drives, took three phase power, were the size of said washing machines,
              and had a cable the size of your wrist. They were essentially just
              rebadged Memorex IBM drives, but since they had a Digital logo, they
              where special. They might have been the most expensive disks, as soon
              thereafter, prices droped to the $20K range you mention. I think I paid
              about $4000 for a 5MB 5-1/2" disk in 83 or 84.


              --
              Pat Farrell pfarrell (AT) bioinformatx (DOT) com



              Pat
              http://www.pfarrell.com/music/slimse...msoftware.html

              Comment

              • mbonsack
                Senior Member
                • Aug 2005
                • 324

                #8
                Originally posted by pvadbx
                speaking of cost for technology, in a past life, I bought 300 megabyte (not gigabyte) hard drives for the college I was working for. they were the size of washing machines, and cost (if memory serves) about $17,500 each.
                When I was at HP in 83 and 84, we affectionately called our 404MB drives "dishwashers", as they looked almost *exactly* like one, replete with front-panel controls and door-locking mechanism for the top-load disk pack. They too required 220V power, weighed a ton, and used a removable pack that looked like a stack of dinner plates in keeping with the kitchen motif. It even sounded like a dishwasher and practically shook itself off the floor when doing heavy I/O.

                These drives were the ones used in the data center that got shot up in one of the "Die Hard" movies. It was good to see these things still had a useful life well after the last bit was transferred from them.

                Comment

                • cliveb
                  Senior Member
                  • Apr 2005
                  • 2071

                  #9
                  Originally posted by pvadbx
                  why didn't cd manufacturers stick 20-30 megabytes of memory into their machines and rebuffer and re-clock the bitstream to avoid all the sonic problems associated with a mechanical transport?
                  I can't believe what I'm reading in the responses in this thread.

                  Every CD player, since the year dot (ok, 1983, when CD players were introduced) has done precisely this. The data coming off the spinning disc is decoded and placed into a small buffer, from where it is clocked out by the master clock at the CD player's DAC. You don't need megabytes of buffering to do this, because the DAC's clock is also used to modulate the rate at which the disc is read.

                  And I'm a little surprised by Sean's statement about how data should be distributed around the disc so that if a small area of the disc is obliterated the data can still be recovered, because this is exactly what is done on CDs, to a certain extent. Perhaps Sean was suggesting that the way it's done doesn't go far enough, but the way I read his post it seemed to imply it wasn't done at all.
                  Until recently: Transporter -> ATC SCM100A, now sold :-(
                  House move forced change to: piCorePlayer(RPi2/HiFiBerry DIGI2 Pro) -> Meridian 218 -> Meridian M6

                  Comment

                  • ncpl
                    Member
                    • Jul 2006
                    • 82

                    #10
                    ROM drives....

                    Pre-Squeeze I have used Meridian players and for quite a few years they have been using DVD ROM drives in their CD players. They spin faster, read faster and can therefore re-read when required to reduce read errors. These have fed reasonable sized buffers prior to the output stages.

                    That said, the CD's don't spin at 7200rpm and the players don't have the same level of buffereing as your average HDD.

                    Comment

                    • pvadbx
                      Junior Member
                      • May 2006
                      • 24

                      #11
                      Originally posted by cliveb
                      Every CD player, since the year dot (ok, 1983, when CD players were introduced) has done precisely this. The data coming off the spinning disc is decoded and placed into a small buffer, from where it is clocked out by the master clock at the CD player's DAC. You don't need megabytes of buffering to do this, because the DAC's clock is also used to modulate the rate at which the disc is read.
                      So the speed of the transport motor is constantly changing to keep the small buffer at an appropriate level? Seems like there is a very small margin for error in this approach. If your statement is true, why do CD players of similar componentry to the SB3 sound so much worse?

                      Originally posted by cliveb
                      And I'm a little surprised by Sean's statement about how data should be distributed around the disc so that if a small area of the disc is obliterated the data can still be recovered, because this is exactly what is done on CDs, to a certain extent. Perhaps Sean was suggesting that the way it's done doesn't go far enough, but the way I read his post it seemed to imply it wasn't done at all.
                      As I understand it, the error correction for Red Book audio is distributed, but only nearby, and is nowhere near the quality of ECC data level error correction in Yellow Book CD-ROM data discs (let alone Sean's suggestion of RAID style, where data is actually duplicated...sort of). I think Sean is suggesting that the disc be treated as a data disc rather than a real-time delivery mechanism. Then you can go to another part of the disc to pick up error correction information, fix the errors, and go back filling the buffer with data. But this only works if you have a big buffer that can handle the buffer draining while the error correction takes place, and can then handle re-filling the buffer at high speed by running the disc at multiple times the data consumption rate.
                      Last edited by pvadbx; 2006-10-13, 18:35.

                      Comment

                      • pvadbx
                        Junior Member
                        • May 2006
                        • 24

                        #12
                        Originally posted by mbonsack
                        When I was at HP in 83 and 84, we affectionately called our 404MB drives "dishwashers", as they looked almost *exactly* like one, replete with front-panel controls and door-locking mechanism for the top-load disk pack. They too required 220V power, weighed a ton, and used a removable pack that looked like a stack of dinner plates in keeping with the kitchen motif. It even sounded like a dishwasher and practically shook itself off the floor when doing heavy I/O.
                        yep, those were the ones. ours used to wander around the computer room when a lot of students were accessing them. for the uninitiated, these were "super-mini" computers that used a time-sharing approach--I managed a lab with 35 terminals where they all accessed the same computer. does anyone remember Pr1me Computer?

                        Comment

                        • tyler_durden
                          Senior Member
                          • Dec 2005
                          • 799

                          #13
                          Originally posted by pvadbx
                          So the speed of the transport motor is constantly changing to keep the small buffer at an appropriate level?
                          This is how CD and DVD player's work. The data is recorded at "constant linear velocity"- i.e. each bit on the disc is physically the same size, so at the start of the disc when the laser is reading near the spindle, the disc spins very fast. As the laser moves away from the spindle, the disc slows down. The speed is constantly decreasing as the disc is played/recorded.

                          This is why those high end belt drive CD players with 5 kg platters that rotate with the disc seem very strange. The motor that drives that assembly is going to have to do a lot of work to get the right speed, mainly acting as a controlled brake once the
                          disc has been spun up to starting speed. All that mechanical stuff can be done away with by using a cheap CDROM drive dumping the data into a buffer.

                          The mechanism that moves the laser and focuses it is very slow compared to the head positioners in a HDD. That is probably why DVDs don't use a HDD storage/EC type scheme. Also, there is limited capacity in the discs to store all that extra ED/EC info.

                          TD

                          Comment

                          • Nikhil
                            Senior Member
                            • Jul 2005
                            • 509

                            #14
                            So basically if we could get all cd player manufacturers to read some uncompressed or losslessly compressed audio files off a data cd (much the way many cd players do mp3 and wma files), we'd be golden right? well not golden, but in a much better place than we currently are?

                            Comment

                            • Robin Bowes

                              #15
                              why didn't cd players ever re-buffer?

                              pvadbx wrote:
                              > mbonsack;145860 Wrote:
                              >> When I was at HP in 83 and 84, we affectionately called our 404MB drives
                              >> "dishwashers", as they looked almost *exactly* like one, replete with
                              >> front-panel controls and door-locking mechanism for the top-load disk
                              >> pack. They too required 220V power, weighed a ton, and used a
                              >> removable pack that looked like a stack of dinner plates in keeping
                              >> with the kitchen motif. It even sounded like a dishwasher and
                              >> practically shook itself off the floor when doing heavy I/O.

                              >
                              > yep, those were the ones. ours used to wander around the computer room
                              > when a lot of students were accessing them. for the uninitiated, these
                              > were "super-mini" computers that used a time-sharing approach--I
                              > managed a lab with 35 terminals where they all accessed the same
                              > computer. does anyone remember Pr1me Computer?


                              Hey, yes!

                              I wrote my final year undergraduate paper in Fortran77 on a Pr1me at the
                              University of Salford in 1991. Something to do with Digital Filter design...

                              R.

                              Comment

                              Working...