Recover Data From a Bad Disc, Restore Damaged CD/DVD [GUIDE]

[admin]

By the time most people learn that optical media are not bullet-proof, it's too late.

The internet is full of horror stories about "lost data" and "dead discs", and most often the blame is put in the wrong place (on the format itself, instead of the lack of testing or use of poor quality media). Understand that much of the time, it's the DVD hardware at fault, your disc isn't really bad. Consider trying a better quality DVD drive, known for its reading quality, such as a Pioneer ROM/burner or a BTC burner. Many burners, such as Samsung/Sony, LiteOn and LG, are not the greatest at reading discs.

What To Do When You Find a Bad Disc

When you run across bad discs, there are two things the must be done:

1. Learn about good and bad media. Buying the cheapest no-name brand, or budget brands like Memorex, Ridata or Magnavox are often why you're in this predicament. While this may seem like an "I told you so" slap in the face, take it as a learning experience.
   .
2. Recover the data and/or make a new disc, as best as you can:
   • In some cases, this means re-burning the disc from the archival master disc or master location (such as a hard drive). It's not a big loss of time to make a new disc.
   • In most cases, however, this was the only copy of the material, and you must rely on quality hardware and software to extract your data. That was a mistake, of course, to only have the one copy.

How to Recover / What Software to Use

Use one of several programs to try and extract data.

1. The first course of action, if this is a DVD-Video disc, would be to simply try and create an ISO with normal DVD software, such as DVD Decrypter. You can download it here. The ISO read mode available in ImgBurn may work similarly.
   .
2. The second course of action is to try the software ISO Puzzle. It requires that Nero be installed, or the Nero WnASPI, or a compatible WnASPI layer. These files have also been attached to this post for your convenience. Let it run for at least one hour, using the "recycle tray" option.
   .
3. ISO Buster is another popular option, but in all honesty it is inferior to ISO Puzzle and DVD Decypter, for the recovery of DVD-Video content. This may work well, as a second runner-up behind ISO Puzzle, for non-video data recovery. In many cases, the free version of ISO Buster will not be adequate, while the $30 pay version (with its UDF reading mode) will work nicely..

In most cases, the DVD Decrypter or ISO Puzzle read will return a partial or complete image to create a new ISO. This ISO can be re-burned to a new disc, or loaded as a drive letter using Daemon Tools Lite or Gizmo Drive (use the freeware versions, paid versions not needed), allowing you to once again access the data.

What Next?

In the future, use only good media, and test the discs if the data is important.

[jmac698]

I think there's more you could add to this FAQ. I have used two disc scratch removal products. One is a handheld repairer. You crank the handle, and a felt band rubs across the disc as it turns. The disc must be wetted first to reduce friction.
Another type is a Philips motorized repairer. It has 3 discs, from rough, to polishing, to cleaning. You use them in order.
I practice though, I just use the rough disc and it works with one 2minute session.
Another tidbit I could add. Why do they always say to clean discs from the center outwards? Well for CD's I can say for sure, it's how the error correction is layed out on the disc.
I can explain it very simply as follows: if you store data, add an extra digit which is 1 if there's an even number of 1's in the data. So 1111 adds 1, 1110 adds 0. Then if there's ever the wrong number of 1's, you know there's a problem. Great, this is called parity and detects an error. Now arrange the data in a two dimensional table. Add the parity for each row and column. Now, not only can you detect the error, but you can pinpoint the exact point at which it ocurred. To fix the error, just flip the bit.
Good data:
11 1
01 0
--
0 1

Bad bit:
01 1
01 0
--
0 1

You can see that row 1 and column 1 have the wrong parity, thus bit (1,1) should be flipped.
Yep, that's basically how error correction works. You can read more about Hamming codes. There are other codes and it's pretty cool how they work. The one on CD's is called Reed Solomon code. There are codes to detect insertions and deletions, bits flipping in just one direction, and other variations. Satellite transmissions and cell phones use other types. You can read about Turbo Codes which are close to theoretically ideal (best protection per extra data), but also fast to decode.

Anyhow, the error correcting codes are stored circumfrentially, so scratches from the hub radially only affect several bits, which is easily correctable. I believe up to 1mm scratch is correctable, I'd have to look it up.

[admin]

Quote:

One is a handheld repairer. You crank the handle, and a felt band rubs across the disc as it turns. The disc must be wetted first to reduce friction.

These are infamous products, because they're known to more often further ruin the disc, as opposed to removing scratches or fixing your issue. The main problem with these units is that the mix of pressure, torque and rotation is very uneven. Been there, done that, ruined several discs myself.

It's basically a really shoddy hand-crank version ($50) of the professional RTI DiscChek devices ($3k+).
Want to buy a DiscChek? They're on Amazon: http://www.amazon.com/gp/product/B00...SIN=B001J5TH1S
There's just really no comparison.

I know that's quite costly for casual home users, but $50 for the cheap crank is a bit of a kick to the guts, too. You'd generally be better off locating a local video service or DVD rental store, which will charge a few dollars to clean your discs. A local rental place charge $3 per cleaning, and even a dozen discs is cheaper than than the crank -- and with better results.

Quote:

Another type is a Philips motorized repairer. It has 3 discs, from rough, to polishing, to cleaning. You use them in order. I practice though, I just use the rough disc and it works with one 2minute session.

Although better than a fully manual hand crank toy cleaner, these are still rather inferior cleaning methods.
And there is still a risk that you'll just make the disc worse -- something that almost never happens from an RTI DiscChek.

Quote:

There are other codes and it's pretty cool how they work.

Yes, very interesting, albeit a bit complex. Most of the DVD formats have their own unique ECC, as did the CD formats. In fact, when it comes to testing media, a large portion of the test is dedicated to monitoring how often error correction is needed, and if it worked successfully. (Of course, that opens a whole new conversation, best fit for another thread by itself. Many users online fail to understand what a test is actually reading, or is able to differentiate burner vs disc qualities.)


Thanks for posting.

[jmac698]

I have to agree with you in a way, it's obvious that my hand crank device leaves new scratches in a specific pattern. My motorized device is better in this respect. However, it has saved some discs for me. If a disc still isn't made better, I just take it to a video rental place who will fix it. Some people use their power tools with a buffing attachment. These generally make a much smoother looking surface.

As for DVD error correcting codes, you are probably referring to the C1 and C2 error rates. You can read highly technical tests of these types on www.cdr-info.com. C1 errors are ok and are normal. C2 errors on data disks are bad. If beyond a certain point, you've lost data. I can't find any good references right now, but I had a book on this. I believe they are based on Reed-Solomon codes. These are specified by (n,k), which means the number of bits it takes in total and the number of bits it protects. A (255,223) RS code protects 223 bits with the addition of 32 bits, totalling 255 bits. A cd typically has 2352 byte blocks which protect 2048 bytes of data. The RS code has the least amount of overhead for what it protects, for that type of code. You can theoretically protect data with almost no overhead, this is what Shannon's Theorem states. We have come quite close to this limit and it's used practically today in satellite transmissions and cell phones. Quite amazing really.

[jmac698]

ps could you comment on my entry to the disc labelling FAQ? I put a lot of research into that. Short answer, disc labels are actually OK and the weight doesn't make it wobble. The problem lies in the tension added by the label; if too stiff it pulls the disc back towards the center which makes it wobble.

[admin]

Quote:

Some people use their power tools with a buffing attachment.

I've seen these kinds of methods discussed on various forums, along with references to all kinds of random polishes and cleaning products -- from toothpaste to peanut butter to Brasso. But if a damaged disc is really worth the effort and/or expense of a repair, then it's worth doing with a proper method. And that doesn't include home power tools, items from under the kitchen sink, or something you'd more often spread on toast. Honestly, those methods are rather idiotic, and not at all suggested if you really want the disc to be recovered. I'm somewhat reminded of sad incidents where somebody has ingested rubbing alcohol (and often died) because he/she wanted to "get drunk" and not understood that not all alcohol is the same. The same wisdom applies here. Not all buffing tools and cleaning products are equal, and meant to be used in just any scenario. Discs are fragile and touchy, and this sort of unorthodox cleaning method is considered to be mishandling and misuse -- i.e., user error. By no coincidence, it's generally user error that caused this entire damaged disc scenario.

Quote:

You can read highly technical tests of these types

When it comes to authoritative information on disc testing and error correction, I'd actually suggest reading some of the white papers from Mitsubishi, Pioneer and DaTARIUS. Of course, the main issue with these documents is they can be hard to find (especially the older MCC and PVC docs), and are really not all that accessible to the layman due to excessive jargon and optical engineering concepts.

Quote:

I believe they are based on Reed-Solomon codes

Correct. This is addressed in quite a bit of detail in Jim Taylor's DVD Demystified (the book, not the free website). If you're truly interested in advanced information on DVD, and are interested in something a little more authoritative than random websites, then it's worth the $33 from Amazon. (Of course, I have found some errors on his site, so even he's not immune to errors. But it's still better than a lot of the usual places, including wikipedia and certain popular-but-amateur forums.)

Quote:

ps could you comment on my entry to the disc labelling FAQ? I put a lot of research into that. Short answer, disc labels are actually OK and the weight doesn't make it wobble. The problem lies in the tension added by the label; if too stiff it pulls the disc back towards the center which makes it wobble.

It was just replied to by staff.

[jmac698]

Thanks, you have significant knowledge! I wouldn't necessarily dismiss 'peanut butter' as a type of abrasive. In the food industry, one term oft-discussed is 'mouth-feel' which actually relates to texture. It's quite possible that even peanut butter has a well-specified granular size. I can't quote you the exact micron-dimensions, but it's quite likely to be within a typical 'spec' for ground products. After all, all grinding is done by the same processes. This is only an opinion however. I've seen specs for example for solder paste which is very highly controlled. This is an example of 'happenstance' which gives you a product of specific granularity which is useful for polishing. Such coincidences are the boom of DIY processes. Very clever in fact, so don't dismiss out of hand!

Look up 'Jerri Elsworth' on YouTube, she is one my my favorite DIY'ers (featured in 'MAKE Magazine', a source of DIY projects, for those who like to discover 'how things work'.)