Specialized Practice in File Recovery
Damaged or corrupted compressed files are difficult to repair
If your compressed files have been damaged or corrupted, standard decompression programs will extract gibberish. The typical answers regarding file recovery are:
- “There are free or cheap programs that fix the checksum so you won't see an error message any more. (But they don't actually fix the corrupted data, they just make the error message go away.)”
- “You did your file transfer with the wrong transfer mode. Don't do that any more.”
- “You'd better have good backups.”
The reason for this unhelpful litany is that every bit counts in a compressed file. In normal compression, all of the preceding input contributes to the encoding of subsequent bytes of data. Errors in the data corrupt not only a few bits, but also corrupt the decompression process for everything that follows. Corrupted bits are usually not detected until well into the decompression process, at which point it's too late: you've decompressed bad data, your output is mostly gibberish, and the process finishes with a checksum or CRC error. Early detection of errors is only possible if there is redundancy in the data; compression deliberately removes as much redundancy as possible, to make the data smaller.
There is hope for data recovery
Damaged compressed files can still be recovered, however, if:
- There is some knowledge about how the file was corrupted.
- There is some knowledge about the format and content of the uncompressed data.
- Sufficient computational resources are available to try many possible solutions until the correct one is found.
To recovery the original uncorrupted data, one can search through many possible reconstructions, trying each, until plausible solutions are found. Naive implementations of this idea require enormous computation for all but the smallest files. Without enough knowledge, finding solutions is not feasible.
Knowing more about your data means that possible solutions can be found with less computation. If you know enough, the search can become feasible.
We have developed tools that make it possible to recover data from corrupt compressed files. Brute-force searching only works on the tiniest of files, so our tools apply a number of advanced techniques to guide the search for solutions: breadth-first search guided by custom heuristics, and machine learning to let the system train itself to recognize good data. This is not a turn-key program; the process must be customized to your data. If you have lost critical compressed data, this may be the only way to recover some or all of it.
For an example of how these tools have been applied successfully, see the case study.
How to decide
To discover whether this service may be applicable to your situation, fill out our questionnaire, and we will advise you on whether your cases resembles other we've seen, whether our tools will be able to help, and how much effort and cost will be involved.