I agree that the literature is poorly written and doesn't explain it well. I actually spent a good 30 minutes in the PS Audio room at the RMAF last fall chatting with Paul and others about the technology and I think I understand it pretty well.
A typical CD player does have a buffer, but it basically amounts to a fraction of a second in most cases (good players will have a second or more). This is to allow for the variations in spin speed and movement of the laser changing the timing data being pulled off the disk. So long as the data pull stays within the margin of the buffer you are fine in this regard. But pull too much data and the buffer doesn't have room for it (buffer over run). Pull not enough data and the buffer empties out and the digital stream stops (buffer under run). This is basically the second paragraph Al quotes above.
The Perfect Wave has a significantly larger buffer which provides much more room for error. Let's say the PW holds a minute worth of music and begins to generate the digital stream when it is half full. This gives the laser 30 seconds either way to keep the buffer filled with enough data to keep music going.
The second part of the equation, however, is the important one. A CD laser is a single pass reader and uses error correction (ECC) to try to clean up dropped sectors - it has one chance to get it right and one chance to clean it up (with imperfect correction data no less). The PW laser operates like a computer drive, which will read a sector multiple times if necessary to assure it has the data right. If you read up on the technology in the ripping software Exact Audio Copy (EAC), you will get a better understanding of the principle at work here. In fact, if I heard Paul correctly, the PW actually uses EAC to read the disk and generate the buffer.
The CD player has to spin at a constant rate that basically reflects the stream rate (this is "1x" speed in CD-Rom speak). The Perfect Wave can spin up significantly faster because it's building a significanlty bigger buffer data file (think a CD-Rom that runs at "16x" or "24x" speed). This is how, when they demo the player, they will eject the disk after about 30 seconds and the music continues. The PW is capable of pulling the data much faster because it is acting like a computer drive to create what amounts to a file, which is then turned into the digital stream.
Finally, the PW generates the digital stream like any other CD player, but if you take the stream out to the PW DAC, it keeps it in I2S format all the way to the DAC chip. A traditional transport device must convert the stream to a different format to accomodate Toslink, SPDiff, or USB, and then the DAC converts it back to I2S to send to the DAC chip. By doing this, all sorts of timing errors and re-clocking problems are introduced which creates jitter and hash. This is where we spend money on external re-clockers (like the Pace Car) to better manage this process. Using the PW combo eliminates this problem because the data is always in the optimal I2S format.
Does that make sense?
A typical CD player does have a buffer, but it basically amounts to a fraction of a second in most cases (good players will have a second or more). This is to allow for the variations in spin speed and movement of the laser changing the timing data being pulled off the disk. So long as the data pull stays within the margin of the buffer you are fine in this regard. But pull too much data and the buffer doesn't have room for it (buffer over run). Pull not enough data and the buffer empties out and the digital stream stops (buffer under run). This is basically the second paragraph Al quotes above.
The Perfect Wave has a significantly larger buffer which provides much more room for error. Let's say the PW holds a minute worth of music and begins to generate the digital stream when it is half full. This gives the laser 30 seconds either way to keep the buffer filled with enough data to keep music going.
The second part of the equation, however, is the important one. A CD laser is a single pass reader and uses error correction (ECC) to try to clean up dropped sectors - it has one chance to get it right and one chance to clean it up (with imperfect correction data no less). The PW laser operates like a computer drive, which will read a sector multiple times if necessary to assure it has the data right. If you read up on the technology in the ripping software Exact Audio Copy (EAC), you will get a better understanding of the principle at work here. In fact, if I heard Paul correctly, the PW actually uses EAC to read the disk and generate the buffer.
The CD player has to spin at a constant rate that basically reflects the stream rate (this is "1x" speed in CD-Rom speak). The Perfect Wave can spin up significantly faster because it's building a significanlty bigger buffer data file (think a CD-Rom that runs at "16x" or "24x" speed). This is how, when they demo the player, they will eject the disk after about 30 seconds and the music continues. The PW is capable of pulling the data much faster because it is acting like a computer drive to create what amounts to a file, which is then turned into the digital stream.
Finally, the PW generates the digital stream like any other CD player, but if you take the stream out to the PW DAC, it keeps it in I2S format all the way to the DAC chip. A traditional transport device must convert the stream to a different format to accomodate Toslink, SPDiff, or USB, and then the DAC converts it back to I2S to send to the DAC chip. By doing this, all sorts of timing errors and re-clocking problems are introduced which creates jitter and hash. This is where we spend money on external re-clockers (like the Pace Car) to better manage this process. Using the PW combo eliminates this problem because the data is always in the optimal I2S format.
Does that make sense?
