Boombats wrote:Oh, and I was never here and you didn't see me. And no I'm not Frank Decent.

Huh?




*(i'm almost always out of the loop)

steve wrote:

cjc166 wrote:If DVDs go out of fashion, fuck it, you tried your best.

I'm going to say about three years.

I have some of these 100 year Gold-Archival DVDs at work. They are quite good- but as Steve mentioned, you better revisit them in three years or so. Using the method someone else recommended of exporting the individual tracks to .wav format is a pretty good idea for a short term solution.

I use .aiff format, in 24 bit quality, 96000khz because I was under the impression that file format was the standard (aren't red Book CD tracks in AIFF format?)

My concern even with this solution is the storage medium.

What's just as disconcerting is the myriad of proprietary file types that litter the digital recording landscape of the last 10-15 years. There's no guiding best practice or industry standard that is as safe as analog. Not because there couldn't be, but because the government doesn't care, the software companies actually encourage the siloed file formats and consumers are mostly too dumb or ignorant to get angry.

Although it is very unlibertarian to say, seems like there needs to a guiding hand with digital files and storage, including certifications and standards at the IEEE level. If your shitty audio software doesn't fit within the standards and best practices framework, it doesn't get the seal of approval and it shouldn't be used by professionals. I am sure this is more difficult in practice than I am envisioning here, but it seems the only way to go.

There are standards for file Encryption that happen to be government sanctioned. Why not file formats and storage media too?


I have a server at work that was "brought offline" about 4 years ago. It has the company's financial records on it, across a series of raid 5 SCSI drives. Probably ten years ago, this was the best way to store important data.

I have no doubt that I could fire the server up and extract some of the files to another medium, but I doubt I ever will. The system has also been backed up to DLT IV series tapes, another medium you need a specific type of tape drive to get to the data on them. Already, we've moved to LTO2, 3 and 4 media as they hold hundreds of gigs of data, not 40. I cannot put a DLT IV tape in an LTO drive and expect that it can understand the data on those tapes. In short, these back ups are fucked from a longevity standpoint.

In 20 years, any data on these systems and media is probably going to be extremely hard to work with, even if the file formats used to store some of the data (Fox Pro DB, some early version of SQL) are available via whatever technology is available in the future to extract, import and parse them. I am sure the FBI could extract and use the data off them, but for anyone else, it's going to be extremely difficult.

It seems like Analog/Tape/Electrical recording had the benefit in its inception of best practice, standards and backwards compatibility nearly always incorporated into its design. Digital does not have this at all.

Who is to blame? Government? The Music Industry? Software companies?

The government is certainly responsible on some level. I was at a presentation by the national standards agency. They were support to create a license level to make burners and cd at rated levels the best would be archival designated and you could only get that designation if your disks were made to certain tested standards.

They took about 3-4 years building the testing equipment and doing the testing. The final results were compelling. Some burners and certain cds can do a burn that would last for 90 - 150 years. The problem was two fold over 95% of the burners would not burn the disks properly or read a disk at the level required. The 4 burners that passed the test were no longer manufactured as they were more expensive than the average cd burner and the consumer could not tell any difference no burners of that quality were thereafter manufactured (from the specs they seemed to be the high end pioneer and Yamaha cd burners from around 93’-4’.

The disks again there were one or two brands of disks that could hold a image longer (they were again of the most expensive gold type) although several of these which looked like the expensive ones actually were rated worse than the cheap disks. But without the better burners although they were better but were not getting past 20 year kind of marks.

DVDs burners were just coming in due to the nature of the multi layer nature of the burn and the multi color laser readers they were considered in basic testing to have more to go wrong with them and that the dyes used in layers seemed to age with errors. You get more over time and of this may not render the disk immediately unreadable but will over time cause audio artifacts, corruption of files and eventual un-readability. The I can read a disk I burned 5 year ago argument is nice but you really don't know the incremedal increase in the number of binary errors (what the government testing is about)

I did a album in 94 that was mastered to gold cd. It is readable only on one of my external Yamaha drives at this point.

That aside those who have the money have a tape back up, and keep the files living on a raid where it lives in two separate identical hard drives in different arrays when one hard drive fails the backup copy immediately copies itself. This is how networks and large corporations store files that might be subpoenaed. This is also done for important video edits.

Running programs in the future on virtual machines means that they will have the plugins and virtual hardware all runtime extensions will see to run the programs including dongle cracks.

Don't get me started on ADAT recording or the stability of DAT.

Just so you know I am a Licensed Archvist who specializes in doing digital archiving.

Yeah, regarding the metal pit thing...thanks for giving away my big money-maker ideas on the internets, Scott! It would certainly be possible, but probably not a feasible solution given today's technology and the costs involved. The storage density is nothing all that crazy given some of the current high density media, but the longevity would be much, much greater.

Just an example:
Take a 1cm square piece of robust material (eg silicon). It would be quite possible to etch pits of 50nm diameter with a number of techniques (for one-off archiving, I'd guess electron beam lithography or focused ion beam lithography would make the most sense). If you divide that 1cm square up into 50nm square bits, you get 40 billion bits. Divide that by 8 (for bits to bytes) and you get 5 billion bytes, or 5GB. 5GB in a square centimeter! Of course, this doesn't include any reference points for tracking. Even with today's patterning technology you could probably push the bit size down to 10nm, so that would give you 125GB.

Give me a cleanroom and a few million bucks and I could start doing this today. The real problem is how to read it, error check, etc. I could imagine a customized automated scanning electron microscope or atomic force microscope, but read times would be pretty slow.

How much would you be willing to pay to have your songs last an eternity?

Will it work better than the Voyager record?



How about instead of pits you make a tiny nano turntable keep it analog but small.

In my opinion, assuming that people won't be able to figure out a software codec 55 years into the future is the same as assuming somebody won't understand how to hook up a transistor.

If a transistor's datasheet disappears, it is more or less useless. However, if enough samples of electronics are around, than you can figure out how the transistor was used. Some transistors are so popular that it would take a catastrophic disaster for everyone to all forget how they were used fifty years in the future.

Magnetic hard drives are a fairly simple technology, I dare say as simple as magnetic tape. The british government thinks magnetic hard drives are so good at storing data that top secret drives are wiped over multiple times, incinerated, and the ashes are stored in a basement.

I have programmed in an "ancient" file format called DST for work. It was originally written in fortran, then ported to C with fortran compatibility, and now we've abandoned fotran support in order to reprogram it once again and implement indexable data. This format is based off digital tape, and before it was purely sequential read access, which made fast file access a headache. The fact that I have learned enough fortran and this entire data format in a matter of a couple of weeks is actually somewhat amazing to me. I've also had to read through BZ2 code in order to understand how to implement it with our file structure in the best ways possible, and it's not that complicated. This format, DST, has survived well over 20 years without a hitch. The C programs can read all the old DST formats.

So, in my opinion, if we assume NOBODY in the future will have the ability to decode some compression algorithm or a data format is ridiculous. Magnetic hard drives, especially with ECC, are a good way to store information. At the sizes we have today, a session could be rewritten 20+ times on the same hard drive for redundancy at a cost of about $40. Gold CD-Rs/DVD-Rs are also a good way. Fundamentally, assuming there are no DVD drives or CD drives is nearly assuming that we will no longer understand interferometry or the fundamental forces of nature. I don't believe the technology behind magnetic tape and other digital formats are completely decoupled.

Also, the PCM format HAS been around for a very long time, and there are 100MB drives you can still access with a PC that are over 20 years old.

The first transmission of speech by digital techniques was the SIGSALY vocoder encryption equipment used for high-level Allied communications during World War II from 1943. In 1943, the Bell Labs researchers who designed the SIGSALY system, became aware of the use of PCM binary coding as already proposed by Alec Reeves. In 1949 for the Canadian Navy's DATAR system, Ferranti Canada built a working PCM radio system that was able to transmit digitized radar data over long distances[1].

juice wrote:So, in my opinion, if we assume NOBODY in the future will have the ability to decode some compression algorithm or a data format is ridiculous. Magnetic hard drives, especially with ECC, are a good way to store information.

Of course someone will be able to recover the data, assuming it is not corrupted, but that person likely works for the FBI or similar. It's not "easy" or convenient for professionals who work in the music industry to recover a early proTools 1 session that used some mac OS 7 TDM plugins of a company that went under in 1996.

However, if that session were saved in Aiff or WAV format as individual tracks starting at 0, then yes.

Right now the process is incorrigable, because of the forward leaning, always look ahead, never look back new file format race madness going on. No software company has much incentive to make something backwards compatible because there are no defined best practices or standards.

ebeam wrote:Yeah, regarding the metal pit thing...thanks for giving away my big money-maker ideas on the internets, Scott! It would certainly be possible, but probably not a feasible solution given today's technology and the costs involved. The storage density is nothing all that crazy given some of the current high density media, but the longevity would be much, much greater.

Just an example:
Take a 1cm square piece of robust material (eg silicon). It would be quite possible to etch pits of 50nm diameter with a number of techniques (for one-off archiving, I'd guess electron beam lithography or focused ion beam lithography would make the most sense). If you divide that 1cm square up into 50nm square bits, you get 40 billion bits. Divide that by 8 (for bits to bytes) and you get 5 billion bytes, or 5GB. 5GB in a square centimeter! Of course, this doesn't include any reference points for tracking. Even with today's patterning technology you could probably push the bit size down to 10nm, so that would give you 125GB.

Give me a cleanroom and a few million bucks and I could start doing this today. The real problem is how to read it, error check, etc. I could imagine a customized automated scanning electron microscope or atomic force microscope, but read times would be pretty slow.

How much would you be willing to pay to have your songs last an eternity?

Hey, this wouldn't cost a few million bucks. It exists; its called ROM. Its routinely used in chip design every day.

Being on tape didn't help preserve all three first Rifle Sport album masters. Somebody lost them all.
There's always the human element in these problems.

megathor wrote:Hey, this wouldn't cost a few million bucks. It exists; its called ROM. Its routinely used in chip design every day.

Sure, but that requires a mask for the photolith and can't achieve the same density. This could be directly written from your digital audio file.