Best CDR for video

[zeb7]

One thinks for sure which ever brand survives 100 or 5 years , its a long time of something else to go wrong ?
such as fire's, floods, children , that can attack the CD's far harder/faster than the material decay

Surely the best way to insure the contents of the cd's would be to
Create the Image of the disc so that you could burn multiple identical copy's of the disc
A CD recovery would be able to make use of identical Disc in the recover process.
Or you would simply to be able to store them in different places to prevent instant destruction

I buy the cheapest Spindles i can find (100cd's at E 0.30 per CD) all most 4x cheaper than Mitsui,
so i could burn each disc 4 times to1 Mitsui ,their checked after buning , then 3/4 years after burn
And before the end of this year i'll have a DVD-R so any decaying CD's many go to DVD ( 6 to 8 per DVD)
Then we have the 27GB bu-ray format disc's soon, so i think a 200 Year Archive life is unnecessary
Will there be CD-Rom drive's in 200 years time ?

I understand that you wish to buy the best , most of the time i think that way too.
In this case i preferrer the re-burn upgrade method, and to this day i have only needed the re-burn 3 cd's out of 400
the files where easily recovered when this happened .
The next 3 years will tell if i'm wong ( as i moved to very cheap CD 1year ago )

Have a look at this program
http://www.elpros.si/CDCheck/

[Brian Ellis]

Zeb7

Your CDCheck utility is very interesting. I downloaded it and used it to check several home-burnt CD-Rs:
1. SVCD 628 Mb on Memorex, ~3 months old, burnt on Memorex TwelveMAXX4 at 12 x
2. VCD 610 Mb on BASF, ~6 months old, burnt on Memorex TwelveMAXX4 at 12 x
3. Data 220 Mb on LaCie, ~3 years old, burnt on prehistoric HP SCSI unit at 1 x (max speed!!!)
4. Archive data 60 Mb on 3-M, 7 years old, burnt on prehistoric HP SCSI unit at 1 x (max speed!!!)
5. Archive AVI file, 480 Mb on LaCie, ~2 years old, burnt on prehistoric HP SCSI unit at 1 x (max speed!!!)

I checked these on:
a) a 48x CD-ROM drive of indeterminate make (WIN 98SE)
b) the 24x read of the Memorex burner (WIN 98SE)
c) a 16x Mitsumi drive of indeterminate age (WIN 2000)
d) the read function of a Pioneer DVR-AO3 (WIN 2000)

All CD-Rs gave 0 errors, 0 warnings with drives a) and b)
Disks 1 and 3 gave errors and warnings on drive c), others 0 errors
I tried only disks 1 and 4 on drive d) and this gave 0 errors/warnings.

My conclusion: obviously this series of tests is anecdotal rather than scientific as the samples are too small and the number of errors are too few to be statistically valid. However, it suggests that the various makes of disks I've used are all OK, even up to 7 years old. I would suggest that the drives may be more critical than the media. This tends to confirm my hypothesis of buying expensive disks for archiving is wasting money. I also suggest that to use CD-Rs for long-term archiving is invalid, not because of the stability of the media but because the drives do wear out. For medium term archiving, I suggest that buying a top-class drive and keeping it in its box until it is really required for reliable reading (ie, your current drives are creating problems) may be a solution but, by that time, it may not be possible to interface such a drive to a computer in 10 years, or even 5 years, time, such is evolution in this industry. It is possible, also, that the CD-ROM may even be obsolete, or at least obsolescent, in 5 or 10 years. It is obviously impossible to forecast the evolution of technology (I've been working on an interface between my crystal ball and the PCI bus for a couple of years now, but I don't seem to be having much success ), but I think it is probably stupid to rely on any of today's technologies being around for any length of time. I have the accounts of my old company dating back 10 years (as the law obliges) on QIC 1/4" tape cartridges but I no longer have a computer which reads them (I have kept one Iomega deck, just in case, but it's gathering dust in a cupboard). However, I'm darn sure that the tapes will not be readable, now, in any case, but I comply with the law!

The question then arises how to archive over the long term (say, > 10 or 15 years). I don't believe that any magnetic method is valid. For simple data/text, printing with a laser printer with certified toner onto archive quality paper (non-recycled, cotton or retted linen paper, untreated with alum or halogens) is, by far, the best bet, provided that it is adequately stored (preferably in multiple copies in different places). This will be good for 1 to 3 centuries. For monochrome photos, silver prints do cause problems but are OK for up to 50 years on condition that archive paper is used and that the prints are individually tested for the lack of processing chemicals (esp. fixers). For longer periods, prints made from the carbon-transfer or bromoil processes would be more stable, as the image is made up essentially of carbon pigment, provided that they are varnished. Colour photos and computer print-outs are more problematic because the dyes employed are never 100% stable. There are colour print archiving processes using pigments (poorer colour quality) but beware of synthetic pigments (e.g. dyed zeolites or suchlike, used in colour photocopiers etc.) but they are horrendously expensive.

How then to archive video? I think the best way is pro reel-to-reel tape (1") on condition that it is rewound at least yearly and recopied every 5 years. Obviously, this is out of the question for the amateur so I suggest that 2 or even 4 hard drives on a RAID 1 array with DV recording may be one possibility, provided that they are transferred at regular intervals, say every 2 or 3 years). This would allow technological evolution to be taken care of, with format changes, as and when. The CD (or even the DVD) is too small in storage capacity for raw data to be archived and highly compressed formats (MPEG) are unsuitable for real archiving.

Just some thoughts...

[Flying dutchman]

The test reports I've read about Emtec CDR's were rather poor (high number of block errors, also called "BLERs").

It may very well be that the average CDR has a shelf life of 100 years for data tracks, but video/audio tracks are a wholly different thing !!! One should always keep in mind that audio-, VCD- and SVCD tracks are written without any sophisticated error correction (unlike "normal" data tracks). Because of this, you can fit 10% more video/audio on the CD. Unfortunately, this also means that they're much more sensitive to scratches, dust, fingerprints and dropouts on the media themselves than tracks written in "data mode".

A 10% deterioration of the reflective surface would ruin the quality of a video CD but remain completely unnoticed on a data CD, so I would strongly advise to use the best media available and store them dark and at a constant temperature.

I agree about the quality of Mitsui Toatsu CDR's (also sold under the TEAC label) but they're pretty hard to get. I use Kodak ultima-80 myself. I pay some Eur 0,60 for them (through our company, ex VAT). They may cost Eur 1,- street price. What the heck, I can fit 40 minutes of camcorder video on it in really good quality (SVCD, CQ 75 in tmpeg).

By the way, this "CQ" mode in tMpeg is truly amazing. I can fit 70 minutes of letterboxed movie (DVD rip) on a CD in stunning SVCD quality- never thought it would be possible. No more CBR for me!

BTW, has anyone ever tried the freeware program "CDR identifier"? It tells you who is the real manufacturer of a CDR rather than what's printed on the label!

[kwmarriott]

Anyone tried gold discs? I have done a bit or research on the net and most reviews say gold discs will last longer than any other disc-but will they play in a DVD?

The other shocking bit was-it seems data discs (when you save data to a cdr) have error correction built in so the data can be retrieved even if the disc has moderate scratching or data errors: audio and video disks don't have error correction so the discs will fail with relitively few errors.

It may be worth creating svcd on good quality discs while saying the original mpeg2 file as a data on a cheaper cdr: put the data cdr away and you could rebuild the svcd from the data disc if the svcd became damaged.

[Brian Ellis]

KW

Don't worry about audio, CDs have redundancy and checksum correction, so that a half-decent player will restore audio from even quite badly damaged discs.

This partially explains why you have only about 70 mins on a CD, yet can have up to nearly 60 mins of compressed video and audio. Theoretically, you could have 2-3 hours of audio from a CD if the redundancy were eliminated.

Take an old, but working, CD and draw 2 or 3 radial lines with a fine permanent marker on the non-label side. You'll find it will probably still work in a quality stand-alone CD player (I find CD-ROM drives are less tolerant, probably because they don't have the necessary correction firmware).

Your idea of saving the mpeg-2 file as data won't work, sorry. If you look at it on both disks, you'll find they are exactly the same length. It is the video format itself that removes any security, to give you the max video/Mb, not burning the CD/DVD. Same for AVI. Not sure for DV, though. Probably the best bet would be to have a second archive on a DV tape, provided it is rewound regularly and then recaptured and re-taped every 5 years or so. That way, you can sleep on both ears!

[zeb7]

Thanks Brian , you have covered a wider range of media and hardware than i have,
I also have had CD's that are not readable on a CD-ROM drive.
Maybe these issue will be muted by small cottage industry ,like the Film to video advertisements you see in magazines ?

Your long term storage Solutions , lead me to the conclusion that any archive format would need maintenance
As i would be extremely unwise for people to buy 200year CD's and not check them in 5 years time anyway .
although i use the cheapest available to me , I think mid priced branded CD the wiser choice ( with the view that
hardware will not be available or compatible in 5 / 7 years ?

One other factor is burn speed my CDRW is 16x speed recorder but i have found ( by using the CD check program)
that my Cheap spindle CD's get the following results ;

LG (came with the recorder) = Burned at 16x no errors
Memorex 10pack with jewel case's = Burned at 16x no errors
Cheap spindle CD's = Burned at 16x lot's of errors (some files non recoverable )
Cheap spindle CD's = Burned at 12x small amount of errors
Cheap spindle CD's = Burned at 8x with no errors (120 cds tested )

So i ask my self what is happening at slower speeds ?
I believe at the slower speed the Laser has more time to burn the dye-layer ?
This would mean the cheap CD 's use dye chemicals that are harder to burn ?
and thus better at withstanding UV or sunlight damage ?
If the Cheap chemicals are stable in all other ways (not sure what's needed) that would make
the Cheap spindle CD's best for archiving ?( that would be ironic )
When using branded CD's ,their would be an advantage to burning CD'
at lower than the Drive's maximum speed also ?

@KuroNeko ; if you can afford to keep buying the Mitsui
then do so , but i would still check them in 5 years.

> BTW, has anyone ever tried the freeware program "CDR identifier"
No not yet! here's a link ;

http://www.cdmediaworld.com/hardware/cdrom/cd_utils_2.shtml

[Flying dutchman]

@brianellis: I hate to disappoint you, but I must really correct you here.

Computer data tracks on a cd are written in sectors of 2 KB plus some 300 (!!!) bytes error correction per sector. There is a lot of redundancy here. This data format was introduced because in computer data a single defective byte can ruin the whole file (unlike audio data).

Audio/video tracks, however, utilize the full CD sector size (more than 2300 bytes) with almost ZERO error correction. There's almost no redundancy here. If I am not mistaken, these two formats are called "mode-1" and "mode-2".

This is the reason why VCD's and SVCD's are multiplexed into those funny-sized sectors of 2300-odd bytes ; one sector on the CD corresponds to one muxing unit thus making "seeking" possible for the VCD player.

The fact that an audio CD still works fine even if it has scratches is simply because an audio CD player interpolates between audio samples if a few are missing. Not because of redundancy. It's hardly audible anyway. It is the filtering/interpolation that does the trick. The scratch becomes very noticeable as soon as you try to make a backup copy of such an audio CD : The Cdrom drive in your PC will skip a part of the faulty sector and it becomes VERY audible on the copy!

In data mode, a 74-min CDR fits only 650 megs. In audio/video mode, it fits over 700 megabytes. Don't believe me? Try to rip a 74-minute audio CD and save the WAV files on hard disk. Then try to copy the WAV files back onto a 74-minute CD in data mode, not audio mode. They won't fit! Do you have a really full VCD? Check the size of the files and compare it to the capacity that's printed on the box. You'll be amazed...

BTW, a VCD or SVCD always contains at least two tracks. The directory and the accompanying CD-I application files are in one track (with error correction), the video clip(s) are in separate tracks (without error correction). So technically they are "mixed mode" CD's, just like some audio CD's that also contain a computer game.

My conclusion: Don't trust error correction on (S)VCD's. Backing up Mpeg files on data CD's is a very good idea but keep in mind that a CDR has 10% less space in data mode.


A good tip for Yamakawa/Raite DVD owners: These DVD drives not only play MP3 files from data discs, they also play video mpeg's if you simply rename them to MP3. That way you have the benefit of error correction but can still play them in your DVD player...

[Brian Ellis]

Zeb7

Some CD-Rs state on the package the maximum speed that they should be used at. My Memorex burner adjusts itself automatically if you put in a lower-than-max (12x for CD-R) CD-R to the disc's max speed, so the manual tells me, the same as it drops automatically to 4x or 2x for CD-RW, according to the medium, presumably read from the ATIP.

I couldn't get CDR Identifier to work on my system. Unfortunately, it would seem that each drive has to be individually added to the list within the software and additions to this have been discontinued (see http://www.gum.de/it/download/english.htm ) since January 2001. Incidentally, the link you gave and also the ones on the above page don't work. I found it on a Czech site at http://www.cdr.cz/software/cdr_identifier/ (the software and readme file are in English, so don't be afraid of downloading it!).

As for stability of photosensitive layers, light and especially IR light will play a role. However (CD-Rs and not CD-RWs which use a different principle), I understand the mechanism is that the layer contains a photoinitiator that releases a free radical that decomposes the chemical, causing it to become less transparent by the release of carbon atoms. This is easy at UV wavelengths but we have to have something that works with IR lasers and there are relatively few photolytically sensitive molecules that work in the 600-650 nm region. Once the carbon atoms are released, they are stable and theoretically, nothing under normal storage conditions can cause them to revert into transparent substances by the formation of new covalent bonds. I can therefore understand that the carbon "pits" will last a very long time. However, it is also essential that the unconverted layer surface should not spontaneously decompose into carbon. Here, I'm on shakier ground, but I think that the photoinitiators may be unstable (hence the restricted lifetime of unburnt discs) and may become inactive through a hydrolytic decomposition. If this is so, then nothing can happen after a given length of time. So what would provoke the hydrolysis? If the layer contains anything that is remotely hygroscopic, then it will attract moisture from the atmospheric humidity through the protective layer and metallisation, or even through the polycarbonate (all polymers are like sieves to the tiny individual water molecules and the metallisation is so thin as not even to be a barrier). This would take time and could be dosed by the judicious selection of materials. If my hypothesis is correct, then it is evident that any changes after the photoinitiators have decomposed would be extremely unlikely, so the data would remain intact for as long as the polycarbonate support, the metallisation and the protective coating remained unchanged. The big danger is therefore restricted to c. 3 - 5 years and, as you suggest, the shorter the shelf lifetime so may the storage time be better.

CD-RWs work on a totally different principle and depend on the phase changes of a metal alloy layer. This makes the shelf life much longer but the data less stable.

As for gold and "silver" discs, this makes no significant difference. At 600-odd nm, the spectral reflectivity of gold and aluminium is pretty much the same. And don't think that the gold discs are worth more: the thickness of metal is measured in angstroms: in fact, the layer is transparent to short wavelength light, it is so thin. The intrinsic value of gold on a single disc is probably less than a hundredth of a cent, even at over $300/oz. It costs umpteen times more to evaporate the Al or Au layer than the value of the metal itself! Oh! and the "silver" discs are aluminium and not silver!

[Flying dutchman]

@Brian: I know aluminum is used on CD's, but is it used on CDR's as well? Aluminum is terribly reactive to oxigen, so the shelf life of the reflector would depend very much on the porosity of the protective laquer layer. O2 molecules diffusing through this thin laquer layer would "eat up" the reflector.

Silver is much less reactive to O2 (but more to SO2, which turns it black) so the same applies here but in a lesser degree. Gold is chemically very stable, so it would really make sense to use it.

[Brian Ellis]

FD

As far as I know, aluminium is used. There are very strong economic incentives for this, not because it is cheaper but because it can be evaporated at c. 750°C, as against over 1,000°C for silver, resulting in much lower energy levels required for achieving a given vapour pressure. OTOH, gold requires about 1,250°C. This may be the reason why the first CD-Rs (sold at several bucks apiece) were gold types, as energy costs were a relatively small part of the total, then.

It is relatively easy to keep oxygen out with a polymeric coating because the O2 molecule is big, compared with, say, water. What would be dangerous for silver would not be SO2, which also has a largish molecule, but H2S, considerably smaller and very reactive to silver. Gas diffusion through polymers is a science in itself and much of modern gas chromatography depends on it but the general rule of thumb is that, provided there is no gas phase adsorption by the polymer, the smaller the gas molecule, the faster is the diffusion through any given polymer composition. I would suggest, then, that aluminium is stable under the conditions of any kind of CD. Spectographic istruments, including astronomic telescopes, that handle UV, have evaporated Al, rather than Ag on their mirrors, because silver absorbs UV and even some blue light (hence its use in photo emulsions: the light energy absorbed causes an outer electron in the halide ion to change its energy state, rendering it easier to reduce the molecule to elemental silver during development).

In practice, my oldest commercial CDs (Al) must be well over a decade old, probably getting on for two (my memory fails as to when we bought our first ones ) and they still play perfectly and there is no apparent sign of deterioration of the "silver".

I've managed to get the CDR Identifier working on a different machine. I've had great fun identifying the source of CD-Rs that I can lay my hands on, including those sent to me by 3rd parties. I found only one Mitsui disc, on a TEAC label. I also found one Ritek one, with only 30 Mb of data, on an unknown label (it was covered by a paper label). This is 3 years old and I checked it for integrity: no faults or warnings. My old 3M gold disks were apparently made by TDK. I could find no correlation between the so-called "good" and "bad" makes and the type of photosensitive layer (so-called long cyanine and short phthalocyanine strategy, whatever that means). In fact, both Ritek and Matsui-TEAC use the short strategy. Only the CD-RWs show the max recording speed, so I don't know how the burner knows when to slow a CD-R down.

[Flying dutchman]

Ah, that makes sense. But as far as I can see, the use of aluminium also makes it necessary to use an (expensive) oxigen-free environment during the evaporation process. Doesn't that increase the cost enormously compared to gold?

I have some golden Fujifilm CD's here where the protective paint starts to come off. I could salvage the contents, though.

[Brian Ellis]

FD

There's no oxygen in either case. The evaporation is done under high vacuum!

[zeb7]

@Flying dutchman : don't be to hard on the first layer of error correction , it may be simple with little overhead
but it dose its job very well , as you may know (see above) i use very cheap CD-R's , last week i burned a SVCD,
only after watching a perfect 43.20mins SVCD on the other PC did i noticed a big 1milimere hole in the disc ! , then i put the disc in the CDRW and run the CDcheck program with no error ! , i still redone the SVCD from the MPEG2 file to be safe .

>The Cdrom drive in your PC will skip a part of the faulty sector and it becomes VERY audible on the copy
That because the copy program did a fast blind read from the Disc , Audio CD players read the Disc in the same way
, but they only read the CD at 1x speed

>In audio/video mode, it fits over 700 megabytes. Don't believe me?
Well its good to know its not just me that makes 820mb Mpeg2 file's when burning to SVCD

BTW thanks for the patch , i don't use it myself , but at least it made M release new drivers
And let us know were they stand ( Matrox removed support from the drivers as you know )

What are they trying to do ? run a business or something

Brian
You nearly lost me when you moved to the chemical level , do i understand you're post correctly ?
>"I'm on shakier ground, but I think that the photoinitiators may be unstable (hence the restricted lifetime of
>unburnt discs) and may become inactive through a hydrolytic decomposition. If this is so, then nothing can
>happen after a given length of time"
So the transparent (unburned) pits would become stable after a given length of time ( still in a transparent state)

> "hence the restricted lifetime of unburnt discs "
I though that the restricted lifetime was because the dye-layer was very slowly changing to carbon , but that would
mean a longer shelf life for unused Cd's, So i must be wrong you're hypothesis is much more plausible , how else would
the CD manufacture's clam 200 year life but must be burned in the frist 3/5 years .

>"the shorter the shelf lifetime so may the storage time be better."
So possibly my cheap CD's have a long shelf lifetime and have a longer
risk period before the Dye-layer become's stable ?

>"I don't know how the burner knows when to slow a CD-R down"
When starting the burn the drive checks the CD with the "Power calibration check" to adjust the laser power
needed for each disc and also to check the burn speed on CDRW & DVD-R's , never see a check for CD-R's
My L.G. drive is rated at 16x write but i don't think the laser is stronger enough to burn all dye-layers types
at 16x ( not all of the dye is changed to carbon at 16x )

[davpen]

I had never given much thought to CD-R quality before. They have always generally worked even the cheap ones. I may have to rethink things a bit for long term backups.

But I am shortly going to be getting a DVD-R/RW drive. Whats the longevity like for DVD-R? Does DVD include more extensive error correction for video than does CD-R?


Cheers,

David

[Flying dutchman]

As far as I know, DVD uses a UDF/ISO format which indeed has error correction.