Today's Industry Standard For Audio Format?

[jeddypoo]

That seems like a confirmation to me!

[buzzaudioguy]

So, basically if I were a bat maybe I could hear the error corrections? Otherwise since I'm not (at least I don't think I am...) I'm probably pretty safe doing what I've been doing. On second thought, screw it all! I'm dumping all my mixes to reel.

[GooberNumber9]

You're correct! Bad science! Errors are corrected for in data discs, and verified bit for bit via a checksum after being burned. If you want the hard science, it can be found in Principles of Digital Audio by Ken Pohlmann.

Block error rate can be counted, and is corrected for. If you take a DVD, burn it off, read that in, burn from that, and repeat 1000 times, the block error rate will not go up, and you will have a bit for bit copy of the data.

Hmm.. I had been led to believe completely the opposite. I haven't confirmed it one way or the, other, but I'll apologize at this point and maybe post again if I find some documentation that contradicts Kevin.

Even if I'm right about errors on DVD-Rs, I'm sure if you burn a 24/96 file onto a Taiyo Yuden with a modern burner, you're going to be fine. Even if there are errors, I can't imagine they would be audible.

Right now I will say that a checksum can only help you correct some errors, and is only guaranteed to tell you whether or not you have an error, it's not guaranteed to let you fix it. There are correctable (at playback time) and uncorrectable errors.

Todd Wilcox

[acjetnut]

Actually, firewire drives (or any hard drive) is a WORSE storage medium! The mechanical parts they consist of wear down, break, go to crap quicker then a well burned, well kept optical disc. Hard drives can also be effected by magnetic fields, where optical discs can't.

When you burn, use high quality discs, the lowest burn speed possible and make a .md5 file for the tracks and include it on the disc. When it is all burned, check the .md5 file to see if everything was burned correctly.

[GooberNumber9]

After failing to find anything conclusive on the web, I decided there was no reason why I couldn't do a real-world test just to see what would happen.

I took a 320 MB file, copied it to an internal hard drive, then to a USB memory stick, then I burned it to a CD-R using Roxio at 48x (it actually burned at around 24x).

Then I compared them all using Windiff, which is a Microsoft utility that does a bitwise comparison of two files. I compared all three files and they all came out identical. I used a built-in OEM Dell-branded CD burner and I don't even know the brand of the CD-Rs, but they are whatever is lying around at my day job.

So, never mind what I said about not using CD-R or DVD-R. Call me paranoid.

I think the relative cost of a firewire drive versus a DVD-R is a bigger issue than potential mechanical failure. :)

Todd Wilcox

[Kevin Kitchel]

You're correct! Bad science! Errors are corrected for in data discs, and verified bit for bit via a checksum after being burned. If you want the hard science, it can be found in Principles of Digital Audio by Ken Pohlmann.

Block error rate can be counted, and is corrected for. If you take a DVD, burn it off, read that in, burn from that, and repeat 1000 times, the block error rate will not go up, and you will have a bit for bit copy of the data.

Hmm.. I had been led to believe completely the opposite. I haven't confirmed it one way or the, other, but I'll apologize at this point and maybe post again if I find some documentation that contradicts Kevin.

Even if I'm right about errors on DVD-Rs, I'm sure if you burn a 24/96 file onto a Taiyo Yuden with a modern burner, you're going to be fine. Even if there are errors, I can't imagine they would be audible.

Right now I will say that a checksum can only help you correct some errors, and is only guaranteed to tell you whether or not you have an error, it's not guaranteed to let you fix it. There are correctable (at playback time) and uncorrectable errors.

Todd Wilcox

Yeah, you burn a disc, you check it. If it fails, then you fix the problem. If it passes, then it passes. Your data is not changed.

How could 1s and 0s just be switched without problems being caused? A disc being read by a laser isn't any different than the same data being sent over a network. These are far from perfect for exact recreation of data, this is why error correction is built into each system. Highly optimized executable files are read from discs without issue when installing an operating system, and audio FILES aren't any different.

[Kevin Kitchel]

Actually, firewire drives (or any hard drive) is a WORSE storage medium! The mechanical parts they consist of wear down, break, go to crap quicker then a well burned, well kept optical disc. Hard drives can also be effected by magnetic fields, where optical discs can't.

When you burn, use high quality discs, the lowest burn speed possible and make a .md5 file for the tracks and include it on the disc. When it is all burned, check the .md5 file to see if everything was burned correctly.

Totally agree about the moving parts issue of drives. They are NEVER to be used as an archival medium. I wouldn't have an issue transferring stuff via hard drive if I didn't have to mail it though. Just make sure the data exists in more than one place.

And I would also recommend keeping the md5 file seperate from the disc, and making a checksum for the whole thing as a volume. A "control" for the "experiment" that is the physical medium. E-mail the checksum, or put it on something else. This way you are checking the integrity of the metadata on the disc as well.

[The Scum]

True checksums are really pretty worthless...if I subtract a bit from one word, and add the same bit to another word, I've corrupted 2 words in the data block, but the checksum is still correct.

A CRC will tell you if a single bit in the block was corrputed so it can be fixed, or if multiple bits are corrupted, the CDC will tell you that the block is beyond repair.

Most modern storage uses CRCs rather than checksums.

As far as the formats, red book CDs use less error detection than data CD/DVD. This checking is reasonably robust.

Consider all of the software that's distributed on CDRom or DVD. If there are bit errors when the software is installed, then the program is corrupt, and it will probably misbehave. There's a pretty low tolerance for that sort of corruption. I've had discs that the player or operating system would refuse to read, long before they installed bad software. I don't see how audio files would be any different.

There's some easy-to-digest information about this here:
http://www.pcguide.com/ref/cd/format.htm
and here:
http://www.cdrfaq.org/

[Kevin Kitchel]

Goober: I will hit the books this weekend to try to verify (or not) this myself. This is going to bug me all night!

[Derrick]

optical storage (CD-R, DVD-R) should be your dead last resort because what you burn is not necessarily exactly what is on your hard drive. If you can send a firewire hard drive, that is best. Todd Wilcox

So have we determined that this is shiser-von-crapovich?

[GooberNumber9]

optical storage (CD-R, DVD-R) should be your dead last resort because what you burn is not necessarily exactly what is on your hard drive. If you can send a firewire hard drive, that is best. Todd Wilcox

So have we determined that this is shiser-von-crapovich?

Well.. I'm not sure if I'd say "confirmed". I officially retract that statement, based on my little experiment of making the crappiest burn I could possibly make and then finding out it was completely identical to the source file.

I was basing my recommendation of hard drive/memory device over burned optical disk on the understanding that once the laser has "burned" a pit on the disk, it's there and you can't go back and fix it. Non-burn file copies involve error checking and behind-the-scenes re-copying if things don't go right.

It looks like despite the underlying theory, the reality is that even crappy OEM burners are good enough to write and read without errors.

*I* say, burn away!

And really, in terms of the original question, anything your mastering engineer or duplication plant or label or WHOEVER is comfortable with is something you can deliver on. MY opinion doesn't matter nearly as much as whomever you're sending your audio to. If someone says "Send me a DVD-R", send 'em a DVD-R! If they say "Send me a cuniform tablet", hire someone else.

Todd Wilcox

[Derrick]

And really, in terms of the original question, anything your mastering engineer or duplication plant or label or WHOEVER is comfortable with is something you can deliver on. ...If someone says "Send me a DVD-R", send 'em a DVD-R!
Todd Wilcox

Well shure, I'm just asking what people are asking for these days typically because I haven't done anything like that in 6 years. Do people tend to ask for DVDs, DATs, CDs, hard drives or something else?

[The Scum]

Last time I had something mastered, the options for bringing in the files were fairly diverse. They wanted standard files (WAV/AIFF in the highest resolution possible), and would accept any reasonable media (FW hard drive, USB drive, CD-r, DVD+/-r), and also had FTP.

They had a handy little booklet that listed the formats they worked with and what they expected from us.

We took 24-bit 44.1 kHz WAV files on both CD-r and DVD+r. They wound up using the DVD, because the DVD reader was faster than the CD reader.

We left there with a redbook CD-r and a sheet with the track listing/PQ codes, both headed straight to the dup plant.

[Derrick]

Good to know. Looks like I can support many of these now so looks like I haven't fallen behind the times... Thanks!