h a l f b a k e r yWe are investigating the problem and will update you shortly.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
CD-Text
Software-inscribe human-readable text on unused portion of CD-Rs | |
Find a CD-Recordable disk with only a small amount of data
recorded on it (say, 30Mb.) A brief visual inspection will reveal
a boundary where the "grayed" (recorded) portion of the disk
(when searched from the center outwards, as the recording
goes) gives way to the empty (wasted, unless its a MultiSession)
portion
of the disk. Most CD-Rs have a fairly large unrecorded
area.
This strip (it is usually a strip of 1-2 cm width) ought to hold
text. This may be inscribed in software, by plotting the
ultra-high DPI text/graphics (why not?) using large aggregates
of 1s or 0s (inversion may enhance readability.)
This should not affect the disk's data stored elsewhere, as
the format allows even "small" CDs (has anyone seen them
sold?) where the diameter is reduced.
However, this practice, if the software is written, may
give rise to a new artform: CD-engraving. Imagine, the DPI
here is higher than that allowed by the most expensive
laser printers. Vast paintings (greyscale), to be viewed in their
infinite detail under magnifying-glasses, would come into
existance... I suspect that a large portion of blank platters
bought would be put to this use.
If the software for this does not exist, I may even attempt
to create it. Does anyone know of past "baked" instances of
this?
Artwork on silicon chips
http://micro.magnet...reatures/index.html
Pictures are already etched into your hardware, why not CDs? [PotatoStew, Sep 17 2000, last modified Oct 17 2004]
(?) DiscT@2
http://www.yamaha.c...ts/crwf1/crwf1.html
Allows graphics and text to be burnt onto a CD-R's unused area [wiml, Jun 19 2002, last modified Oct 17 2004]
(?) DiscT@2 example
http://www.giles.co...c/discTa2_0602.html
What it looks like [sadie, Jun 26 2002, last modified Oct 17 2004]
(??) Patent 6,264,295
http://patft.uspto....=6264295&RS=6264295
Sadly baked. [sadie, Jun 26 2002, last modified Oct 17 2004]
(?) Another example of Disc T@2
http://www.kunarion...o/gallery/hbcd1.jpg
as a breakfast toast stand. [Amos Kito, Oct 17 2004, last modified Jul 16 2005]
[link]
|
| |
Interesting. I imagine the recording format uses an encoding (like the RLL schemes hard drives use) that prevents long runs of '1's or '0's. Depending on the format, it might still be possible to create something visible... |
|
| |
I've never seen anything like this. Bake away! |
|
| |
"Help, I'm Being Held Captive by RIAA." |
|
| |
I've thought of this same notion. I expect that some hardware changes might be needed on many drives to make this technique usable; among other things, when writing a CD precise alignment of data on adjacent "tracks" isn't normally necessary. There are three things going on: the disk is going around, the head is moving slowly inward, and data is being clocked out to the head. Provided these events are happening at about the right relative speeds it doesn't matter if they're 100% precise. By contrast, for the application described here even a 0.1% error (or medium-term jitter) between the data rate and the CD rotation would cause the "text" produced to look quite ragged, and a 1% would render any but the largest text illegible. |
|
| |
The only way I can see this working given my understanding of how the drives function would be for a program which was running with all other interrupts disabled (to avoid timing variations) to start out by making one sync on the disk in about the center of the desired text field, then repeatedly (with CD running at constant speed): |
|
| |
- Go to sync track
- Look for sync mark and start timer when found
- Go to another track
- Wait for timer to expire and lay down a sync mark |
|
| |
Once this was done, the system could "measure" the sync marks' accuracy and use them to write the rest of the required visual data. |
|
| |
BTW, paper tapes used to be labeled in a fashion much like you're describing: since the reader would ignore everything until a certain "start" pattern was observed, the header of a tape could be punched with any design of holes not containing the start pattern. Cool, eh? |
|
| |
I've seen this on CDs already. Of all things (gasp!) an AOL cd. The outer .5" of the disc were these huge AOL logos in a big circle. |
|
| |
I've also seen stuff like serial numbers and the name of the mastering houses burned graphically on the inside ring of CDs. Pick up a few and look at them. |
|
| |
Based on what I can find out about the CD recording format, this is unfortunately not possible to do simply by carefully shaping the bits supplied to the burner. If you can drive the burning hardware directly (bypassing even the very low level data encoding algorithms) you might be able to make pictures. For all I know, the AOL logos were impressed with a completely different mechanism. |
|
| |
You could reverse-drive almost everything in the encoding chain (and there's a lot of stuff in there!); you wouldn't be able to control every bit, but you could at least arrange for some areas to have many more '1' bits than '0'. Even the so-called 'scrambling' system could be defeated, since it's a simple XOR against a determinstic algorithm. |
|
| |
The unfortunate kicker is the low-level encoding technique, which is a process known as "EFM", which stands for "Eight-to-Fourteen Modulation". It bears that name because, for every eight bits of input data, fourteen bits of output are generated. This modulation is used for several reasons, but unfortunately for us, one of them is to keep the "digital sum" low. The "digital sum" is the number of one bits minus the number of zero bits -- that is, the degree to which the data isn't an even mixture of 1's and 0's. |
|
| |
Unfortunately, to draw pictures, we specifically want to alternate large regions with many '1's (one "color") with large regions with many '0's (the other "color"). EFM (and the three "merging bits" that go between each 14-bit word) are designed specifically to prevent this. (I'm not sure exactly why, but apparently it's easier to design electronics if the DC signal is kept small.) |
|
| |
Incidentally, what with EFM, merge bits, synchronization patterns, two levels of error correction, and two levels of "packet headers", the amount of data stored on a CD-ROM (approximately 650MB) is well under one third of the total number of bits physically encoded on the disc. |
|
| |
(And did you know that the track on a CD-ROM is three miles long?) |
|
| |
Finally, this doesn't even bring up the sychronization issues which supercat touched on. It should be very clear by now that if you want to put rainbow doodles on your CD-R, you either need very low-level access to the burning laser, or you need to break out your permanent markers. The standard SCSI interface to a CD-R won't cut it... |
|
| |
Thank you egnor. I agree 99.9% with what you said (and might have touched on it myself except that I didn't know the exact details). There are two ways I can see of getting around the issues you describe; can you tell me if either of these would have any useful effect? |
|
| |
[1] While bit sequences must have a relatively constant proportion of holes to gaps, it may be that an area of the disk filled with low-level "101000001010000010100000" would appear slightly visually different from an area filled with "100010001000100010001000" due to the different-sized gaps. Since the gap sizes in question are relatively small multiples of the wavelengths of light, it would seem that they might exhibit slightly-different irridescence. |
|
| |
[2] At some level, there have got to be commands to format a new part of the disk. While it's possible that the only such commands either have to start with the head at the center of the disk or else "append" to existing data, that wouldn't have to be the case. On the other hand, I just realized something [darn] |
|
| |
On many cheap CD players, there is no absolute positioning available other than the very center of the disk; all other positions are found by reading "markers" on the disk which say where the head is currently, and going inward or outward from that. I would not be surprised if many CD-RW drives behave similarly, but with their read/positioning head located a "track" away from their write head. If this is so, the only way the drive would be able to move the head to the outer parts of the disk would be to have first written the inner parts. |
|
| |
Perhaps it would be possible to "trick" the drive into formatting a block near the outside of the disk, then formatting a block one track outside of that (using the just-formatted block as a placement reference), and continue outward from there. This would create a rather coarse outward 'spiral' (which might go around the disk once every 1/20" or so) but it might then be possible to use that spiral as a "foundation" from which to build more detailed designs. |
|
| |
Just a side-issue - Why do people (like [egnor]) talk about the 'track' on a CD? I had thought that CDs were made up of multiple concentric tracks, rather than a single spiral track (like records). |
|
| |
Hippo: Nope, CDs are made with a single spiral track, like a vinyl record (and unlike a hard disk). |
|
| |
Supercat: That's true, I hadn't thought about diffraction effects. (Ooo, could you make a hologram, a la Beaty's handmade holograms? No, better keep it simple...) |
|
| |
hippo: A CD-rom is written with one big long spiral track. There are several reasons for this. For one thing, it allows music (or data) to be read out continuously without any brief gaps while advancing the data head. Additionally, CD's are written with a constant linear density; the use of a spiral track means that it's possible two read/store a fractional number of sectors per revolution. |
|
| |
The Windows 2000 distribution CDs have anti-counterfeit holograms on them. Are they burned into the subtrate (on the reverse side), or just applied after the CD is mastered? |
|
| |
Mass-produced plastic holograms and mass-produced CD's are both produced by the same method; press some heated plastic against a master which contains lots of teensy bumps, thus producing lots of teensy dimples. While the techniques used to make the master for a hologram are probably different from those normally used to make the master for a CD, there'd be nothing preventing someone from taking part of a hologram master and putting it in the middle of a CD master. |
|
| |
00000101110101110101110100001101110100000101110
11011101110101110100100101110101110101110110101
11011100000101110101010100011101010100000111011
11011101110101110101110101110100100101110110101
11011101110100000101110100001101110101110101110
Look closely, and you will see that I wrote my name. |
|
| |
This is an idea a few of us tossed around when I was working for the government (your tax dollars at work). Technically, as I learned from my Fourier Optics class, CD/DVDs/etc *are* holograms - that is - they are encodings of amplitude and phase on a medium. Your (our) problem is that you want, as someone pointed out, non-white holograms. Because of multiplication by a whitening matrix, you get a psudeorandom bitstream from what could be a non-random stream. However, there is no reason you could not work out backwards what the input has to be to recieve a non-whitened output b/c, and here's the catch, the drive is "expecting" a near-random stream of bits. It is certainly not expecting a stream designed specifically to fool it. If I can work out the reverse mapping, I'll post it. Incidentally, is it bad that I flunked out of Random Processes 101? |
|
| |
Ok, upon investigation, it would appear that EFM, b/c of the 8-14 modulation, defeats us somewhat. It was pointed out that EFM is done to keep the "digital sum" low. This is somewhat accurate, however, more importantly, it keeps the RF signal bounced back from a CD from constructively or destructively amplifying itself by having two "ones" next to each other. Put simply, 14 bits is the fewest number of bits you can use to describe 8bits of binary data without ever having a "one" next to a "one". So, this kills the idea of printing a picture. Any pictures you have seen were certainly printed on (as all mass-produced CDs are printed, as was pointed out). |
|
| |
But nevermind pictures! The real idea that should be expounded on was already proposed and sumarily dismissed! Holograms do *not* need large areas of 1's and 0's. Holograms, typically, are quite white. Additionally, there is more than one possible method of encoding said hologram and, as such, I am confident that a method could be developed. For example, while 1's are prohibited from being neighbors in a linear sense, they are certainly not prohibited from being "radial neighbors". This feature could certainly be exploited to produce holograms. However, my ideas end there. If anyone wants to code something to do this, let me know byu posting and I'll work out the math nessesary to convert stereo images to bit-shifted holograms. |
|
| |
Baked; see link. (Since Yamaha controls the drive firmware, of course, they don't need to work around the various encoding & interleaving schemes.) |
|
| |
Regarding "DiscT@2": I would love to see the actual date of their patent filing. |
|
| |
Me too. It looks like US Patent #6264295 covers this idea, though, and it was filed in 1998. |
|
| |
(It's kind of convoluted though. Maybe because you can only patent the process for doing something --- in this case they're patenting the application of cartesian->polar transforms to rotating media --- not the idea of something that would be nifty to do.) |
|
| |
//Me too. It looks like US Patent #6264295 covers this idea, though, and it was filed in 1998.// |
|
| |
That hardly makes this a 'me-too', since neither the author nor anyone else on this forum was aware of any such idea existing elsewhere until quite recently (DiscT@2). Seems more like a case of 'great minds think alike'. |
|
| |
// neither the author nor anyone else on this forum was aware of any such idea // |
|
| |
I think they were suggesting that Yamaha got the idea from us. Imagine a class action from the members of half-bakery.com (mostly anonymous) against a major corporation. Is there a lawyer in the house? |
|
| |
But yes, that patent does look real enough. |
|
| |
And things come full circle. bakery-->fleshed out idea-->corporate idea-->product-->Amos-->bakery |
|
| |
Amos - I am quite jealous of your toy. |
|
| |
The Yamaha system looks absolutely fabulous, but is probably pricey as heck. I have a regular Samsung combo drive. Does anyone know where I can get third party software to do this? |
|
| |
//An ideal this wonderful is baked, you have the baked product, and yet you complain that it's not good enough? Only a true halfbaker would have the nerve. Shirley you should have mentioned aliasing issues back in 2000 when yamaha was taking this idea as their own.// |
|
| |
Well, I would have expected that input in a polar-coordinate-based format would be a natural logical method, since most disks have little enough "free" space at the perimeter to make it hard to do anything useful without wrapping text around. |
|
| |
That being said, if I could find an API for the thing I'd try my hand at writing my own application. |
|
| |
FloridaManatee: The degree of mechanical tolerance required to make the drawing work is such that no drive that wasn't explicitly designed for such a purpose would have any chance at doing it effectively. For burning "normal" CD's a track-to-track jitter of even 1/4" would not be a problem, but even 1/40" of jitter would render any 'picture drawing' abilities useless. |
|
| |
"I've also seen stuff like serial numbers and the name of the mastering houses burned graphically on the inside ring of CDs. Pick up a few and look at them." |
|
| |
Wow...cool! I just picked one up right off my desk and there it was. That's cool. I've never noticed that before. So is that done in the same way this idea is describing? |
|
| |