GB vs. GiB -- an explanation of storage capacity

[Aurora]

This is not smackdown material, but if it isn't put here, the people who need to read it won't read it.

DVD Forum approves 'bigger than Blu-ray' HD DVD

After reading this thread, I discovered that there is a ton of misconception about why the actual capacities of optical storage discs are lower than the actual capacities the OS reports. It's important to realize how this stuff actually works. So far, people have proposed
-There's software on the disc
-They can't produce a full BD50 because of poor yield, so they only make it 45 GB
-Historically, optical storage capacities are reported as actual, but BD is an exception

These are all completely and totally wrong.

First, there no "software" on the disc. An optical disc, like any other storage medium, must contain a file system to be readable by any operating system. If there are logs and file system dedicated space on everything, they don't take up GB of space. They take a few MB, max. I'm not intimately familar with the file systems of optical discs (ISO9660), so I can't say for sure if there are file system specific files taking up some space.

Secondly, a BD45 would have different pit widths than a pair of BD25's glued together, and that's not what they do. If you wanted to make a normal BD50 into a BD45, you'd have to not use some of the pits available. That is, you'd be taking a glass, and filling it only part of the way, declaring the top part of the glass unsuitable for containing liquid. It doesn't work that way. You wouldn't produce a disc with 50,000,000,000 bytes and only decide that 45,000,000,000 are good. That would be called a bad disc, and tossed out. If your goal was to make a disc with 50,000,000,000 bytes, you wouldn't accept less. If you could only actually make one with 45,000,000,000 bytes, you wouldn't call it a "50 GB" disc.

Historically, the available capacity on optical storage has NOT been the actual usable space. Ever try to put a full 4.7 GB on a DVD? Couldn't do it, could you?

Here's a lesson in how storage formats report their data. Storage formats--ALL OF THEM--HDDs, DVDs, HD-DVDs, BDs, flash drives, etc. use the SI definiton of KB, MB, GB. That is, 1 KB = 1000 bytes, 1 MB = 1000 KB, 1 GB = 1000 MB. Thus, when an HDD reports that it has 80 GB, they mean it has 80,000,000,000 bytes available on it. Computer OSes don't use that. They use the binary definition for KB et al. 1 KB = 1024 bytes, 1 MB = 1024 KB, 1 GB = 1024 KB. TO prevent this confusion, these are sometimes written as KiB, MiB, etc. ("binary kilobyte"). So, let's do the math

For DVD,
4.7 GB ==> 4.337 GiB
8.5 GB ==> 7.91 GiB

For Blu-ray,
50 GB = 50,000,000,000 bytes. To convert that into the GiB that the computer sees, we divide by 1024^3, and get...46.56 GiB.

For HDDs,
80 GB ==> 74.5 GiB
120 GB ==> 111.75 GiB

For HD-DVD,
30 GB ==> 27.93 GiB
51 GB ==> 47.49 GiB

They've reported pretty much every storage format using the SI meanings for over 10 years now, so there's no reason to expect a change now. The only exceptions to this reporting scheme I know of are the first Western Digital Raptor drives, which came in 36 GB and 74 GB (note that the new 150GB Raptor uses the 1 GB = 1000 MB), and enterprise SCSI drives. I don't know why they put actual capacities on those products. I'm guessing it has something to do with marketing, feeling 80 GB sounds better than 74 GB, for mass market products. For server products, the capacity marketing is less important, since users of those products are concerned with available storage space and things like reliablity, etc. You sell the masses on numbers, technical experts on real features.

Generally, if you see a capacity reported as a "nice number," like a multiple of 5 (25, 50, 100, 500, etc.), it's going to be the "not actual" capacity. If you see something strange, like 74 GB, 36 GB etc, then it's probably actual.

EDIT: Spelling, couple clarity issues

[Nighthawk]

Aurora,

Thank you very much for this post. You have addressed many misconceptions in this forum. You rule girl !!!

[WallyWest]

Perhaps this should be made into a sticky thread? Make sure its always up here for reference?

[TheGreenDude]

I really thought this sort of thing was more common knowledge than that... Or are there really that many people that buy a device with XX storage and not question why none of them ever show as much space?

[WallyWest]

Originally Posted by TheGreenDude

I really thought this sort of thing was more common knowledge than that... Or are there really that many people that buy a device with XX storage and not question why none of them ever show as much space?

I think peeps know it, they just don't think about it.

[bigshynepo]

Thanks for the informative post. Using computers most of my short life so far, I was always very aware of this fundamental rule of storage. However, I only thought it applied to hardware based media and not optical. I now know that to be wrong. My ancient external HP 4X CD-R burner barely got used due to its incredible write times so I never joined the optical media burning movement.

[james]

Thanks for the info, very good explanations.

[hi_i_am_wade]

You have to remember that computers work in 1's and 0's and storages manufacturers work on $'s and $$'s. The REAL capacity is in base-2. Storage manufacturers use base 10 to make you think you are getting more than you really are. That is why they do it that way. And it is too confusing.

It works like this:
2^10 (2 to the 10th power) is 1024, that is a kilobyte.
2^20 is a megabyte
2^30 is a gigabyte
2^32 is 4 gigabytes, the memory limit on most computers
2^40 is a terabyte

Nevertheless, since they are uniform in lying to us, it does make for a good comparison.

[AV_Integrated]

Thread moved - left redirect. Sorry

[firelizard]

They've reported pretty much every storage format using the SI meanings for over 10 years now, so there's no reason to expect a change now. The only exceptions to this reporting scheme I know of are the first Western Digital Raptor drives, which came in 36 GB and 74 GB (note that the new 150GB Raptor uses the 1 GB = 1000 MB), and enterprise SCSI drives. I don't know why they put actual capacities on those products. I'm guessing it has something to do with marketing, feeling 80 GB sounds better than 74 GB, for mass market products. For server products, the capacity marketing is less important, since users of those products are concerned with available storage space and things like reliability, etc. You sell the masses on numbers, technical experts on real features

Auroa, outstanding post, just to shed a bit of light on why SCSI or enterprise system drives are reported as actual size is simple. Kind of like you stated above "..masses on numbers, technical experts on real features", well storage engineers need to design systems based on empirical measurements. Larger systems require real finite calculations as typical a storage array that is > 1TB is mostly like using a highly propriety low level embedded OS ( no real wiggle room for padded numbers ) has to be precise ). So, thanking the enterprise industry wastes no time and provides factual unpolished numbers. Also, on a side not, the raptor is considered an enterprise class drive platform, so respectfully adheres to the same reporting standard as SCSI.