2. tech-talk
  3. 149591-112572

Theoretical question about how CD's work

Theoretically, can the contents of a CD be printed out onto sheets of paper in 1’s & 0’s, re-entered digit by digit (say, by a generous helper monkey with an infinite lifespan) into some sort of program, and the same sound will be replicated? Just trying to understand how CD’s work (though I’ve been trying for 25 years and it still seems like magic to me).
sealrock
nonoise's avatar
nonoise
10,647 posts
 
So it's settled. The monkey on the player piano is the correct outcome.
🐒 + 🎹 = 🎼

iasi
115 posts
 
here is a somewhat related comment...i'm sure that many of us played a CD in the car while driving...i bet it sounded fabulous...regardless of the laser scattering, the bumps in the road, and all the motion the car experiences during a normal drive...athat cheap CD player sounded pretty good...
geoffkait's avatar
geoffkait
23,365 posts
 
Well, the CD player in the car buffers the data, otherwise the CD would not play at all. However, data buffering cannot fix the damage already done by scattered light and the vibration CD, which occurs in the first picosecond as the laser reads the disc. So, you would have thought the sound was even more fabulous had the scattered light and CD vibration been eliminated. Everything is relative.

A lot of people think their system sounds fabulous. I get it. 😬

No matter how much you have in the end you would have had even more if you had started out with more. 😛
Post removed 
kijanki's avatar
kijanki
4,413 posts
 
Most of CDPs play in real time - there is no time to read particular sector again.  Ripping program can access each sector unlimited number of times to obtain right data (proper checksum).  Car or portable CDPs have to make up time to be able to return to troubled (bump) moment/sector and reread it.  In order to do this CD is rotating a little faster than necessary and the data is buffered.  I had once Panasonic portable CDP with a switch to enable/disable this mode.  I could clearly see thru the window CD rotating faster in bump-proof mode.  I'm only not sure what is considered a "bad data".  Reed Solomon code allows not only error correction up to certain number of missing data point (2.5mm scratch along the track per Al's post), but also data interpolation for a little longer scratches along the track.  CDs in really bad shape might still play fine, but a lot of data will be interpolated.  Ripping program can obtain right data to create new CD-R with better sound than original scratched CD.  I was even able to save that way few completely unreadable CDs (but it took long processing time).
More to discover