Burning data: any reason to use Mode2 instead of Mode1?

[Awatef]

I'm talking about normal 700MB data that is to be burned on a normal 700MB-CD, I'm not interested in 800MB tricks.

Is there any advantadge of using Mode2 instead of Mode1 with normal data CDs?

[avih]

no. there's no advantage. if you select mode 2 with nero, it uses mode 2 form 1, which is VERY similar to mode 1.

both have disc capacity and error correction exactly the same way. but mode 1 is considered a bit more 'compatible' format, although afaik all windows versions can read mode 2 form 1 with no problems.

if you're not interested in '800M tricks' stick to mode 1.

cheers
avi

[DeXT]

Mode1 is mostly dead, Multisession requires Mode2 as well as Joliet, as seen in the specs (in Nero you can use Mode1 for Joliet but it's out of the specs). Of course VCD/SVCD requires Mode2 for all tracks, too (in fact you cannot build a VCD with Mode1 for the ISO track, since *.RTF files are burned in Mode2/Form2, and a CD-ROM/XA session can only contain Mode2 tracks).

So all "recent" standards are built in top of Mode2. Any player willing to support Multisession or Joliet must support Mode2, too. Mode2 has been around for quite some years. Probably half the CDs you already have, either originals or burned ones, are Mode2. A player that can only play half your CDs cannot success in the market, so I would not worry about potential compatibility issues.

When compatred to Mode1 there are no "advantages", other than the ability to mix Form1 and Form2 files in a single CD, and using the XA ISO extensions (rarely used in Windows, in favor of Joliet).

[Awatef]

@ Dext
you mean, Joliet is normally only possible in Mode2, but Nero cheats and make it possible with Mode1?

[DeXT]

Well seems I was wrong about Joliet and Mode2 relation, from the Joliet specs:

The data area for a Joliet volume on a CD-ROM shall be comprised of either Mode 1 or Mode 2 Form 1 sectors. CD-ROM media utilizing the multisession recording techniques outlined above may not contain any Mode 1 sectors anywhere on the media. Mode 1 sectors are allowed only on single-session media. So you cannot use Mode1 on a multisession disc, but you can use it on single session discs with or without Joliet.

In any case Nero "cheats" many things regarding the ISO specs, as an example you should not put any character other than the allowed ones in the ISO-9660 part of the filesystem but Nero allow you to do so, even spaces. Also, every file name must contain a dot (.) with or without an extension and Nero does not put one when no extension is found. These are some examples.

But as I said since Mode2 has been in the market for many years I doubt there is a single player out there not supporting Mode2Form1 tracks.

[Awatef]

Ah ok, seems like I'm going to use Mode2 more often :D

[SirDavidGuy]

Mode 1 is certainly not dead!

More than 90% of Data CD's are mode 1. It has the best there is in terms of error recovery.

Another question is: Is there any use for mode 0? (Yes, there is a mode 0)

[Awatef]

@ SirDavidGuy
What do you mean by "it has the best there is in terms of error recovery"?

avih stated that Mode1 & Mode2Form1 have the same error protection, so why should be Mode1 be better in that matter?

[SirDavidGuy]

Originally posted by Awatef
@ SirDavidGuy
What do you mean by "it has the best there is in terms of error recovery"?

I mean exactly what I said. 2 things can be of equal qulaity and still be the best.

True mode 2 (ie. Following ECMA-130 / ISO/IEC 10149), actually defines no form 1 / form 2. It is defined in CD-ROM/XA.

[amni]

bytes, can really correct error in the 1024 "main data";
unless you consider error correction done only on-the-fly
(while burning).

Sems to me that there is false myth about that
"error correction scheme"

For long lasting error correction you might need
at least additional 1024 bytes per 1024 bytes in the
"main data" (in other words you need a second backup
for the 1024 bytes in the "main data".



Originally posted by SirDavidGuy


I mean exactly what I said. 2 things can be of equal qulaity and still be the best.

True mode 2 (ie. Following ECMA-130 / ISO/IEC 10149), actually defines no mode 1 / mode 2. It is defined in CD-ROM/XA.

[mpucoder]

The extra bytes are not redundant bytes, they are reed-solomon codes. And they can correct short burst errors very well (without Reed-Solomon your HD would be a miserable device, and DVD would suck)

[amni]

on-the-fly CHECK SUMS (in my terminology) and I said that
on-the-fly correction makes sense.

What I object is claims like "if you use MODE2 FORM2 data"
you will not be able to read that datam, if you have
"errors due to stains and scratches on the CD" because their corrections reside in the extra 228 bytes.

I guess that SOLOMON REED are sort of CHECK SUMS which
make it easy to find locations in the 1024 bytes
which are "hard to read", but these locations
should be re-read from the main 1024 bytes and not
from any other place.


Originally posted by mpucoder
The extra bytes are not redundant bytes, they are reed-solomon codes. And they can correct short burst errors very well (without Reed-Solomon your HD would be a miserable device, and DVD would suck)

[mpucoder]

All I can say is, you should do some reading on the science of error detection and correction. The Reed-Solomon codes provide for the detection and correction of bursts of errors typical of those found in any stream of data. They also protect themselves. They can, if used properly, decrease the hard error rate by three orders of magnitude (.001). The system is much more sophisticated than a simple checksum or CRC.

The claims that mode 2 is more likely to be rendered useless by coffee stains and scatches is a valid one. CD's were designed for music, and in order to get the greatest capacity some compromises were made in error rates. To music it doesn't matter much, the error will be short and masked.
Also in those extra bytes are positioning data. You may not know this, but CDs, again made for streaming audio, do not fully encode the sector address into each sector. Insteead the positioning data is contained in 2 bitstreams (2 bits/per sector) so that it takes awhile to find out what sector is passing over the pickup. Old CD-ROMs had difficulty ripping audio because of this, newer CD-ROMs track the position better and buffer more data so that the disk can be read easier. In mode 2 you are fully dependant on the drive firmware to supply the correct sector, there is no confirmation in the "extra" bytes.

[DeXT]

Well AFAIK (and as avih said, too) Mode2 Form1 has EXACTLY the same error correction, detection, and positioning (subchannel) information that the one present in Mode1 tracks. The only difference is that the sector contents are arranged in a different way and it takes use of the unused (and wasted) 8 bytes of "intermediate data" to store the XA sub-header.

Seems some people tends to think that Mode2 sectors are always read as raw, and because of this neither the drive nor the OS itsef makes use of the error detection/correction information. This is totally wrong.

I made my own tests. Burned a Mode2 sector (no matter if it's Form1 or Form2) with an invalid EDC information. The OS reported a read error when trying to read such sector (so it makes use of the error detection). Then burned a Mode2/Form1 sector with valid ECC and EDC information but intentionally changed 1 single bit inside the user data area before burning it. When reading through the file system the sector was corrected on-the-fly (i.e. the bit was back to its original state), and after reading the track in low level my modification was still here (so it was not corrected while burning, as expected). So it makes use of the ECC too.

Of course this behaviour should have been exactly the same with a Mode1 track. So I think you can sleep quietly.

[amni]

I wouldn't like to read on Error Correction Schemes, but as a proffssional
mathematician I have a fairly good idea about what can be done
and what cannot be done in such schemes.

True correction can be
done only on-the-fly while the burning CD process continues;
after the CD burn terminates no true error correction can be done
(because there is no redundancy in writing).

After CD burned retrieving erronous-suspected data can be done
only by re-reading more carefully the suspected areas
on CD; error correction scheme may accelerate the process of
re-reading by better "focusing" on the suspected areas on the
CD, using what you call "positional" data.

So the difference in reading a stained CD in MODE2FORM2
and MODE2FORM1 (or MODE1 BINARY DATA) is only in the time
needed to re-read (and the software which enable carefull re-reading);
I guess that using software like BLIND READ (which does very
careful re-reading) I can read from stained MODE2FORM2 not less
than what I can read from a stained MODE2FORM1.




Originally posted by mpucoder
All I can say is, you should do some reading on the science of error detection and correction. The Reed-Solomon codes provide for the detection and correction of bursts of errors typical of those found in any stream of data. They also protect themselves. They can, if used properly, decrease the hard error rate by three orders of magnitude (.001). The system is much more sophisticated than a simple checksum or CRC.

The claims that mode 2 is more likely to be rendered useless by coffee stains and scatches is a valid one. CD's were designed for music, and in order to get the greatest capacity some compromises were made in error rates. To music it doesn't matter much, the error will be short and masked.
Also in those extra bytes are positioning data. You may not know this, but CDs, again made for streaming audio, do not fully encode the sector address into each sector. Insteead the positioning data is contained in 2 bitstreams (2 bits/per sector) so that it takes awhile to find out what sector is passing over the pickup. Old CD-ROMs had difficulty ripping audio because of this, newer CD-ROMs track the position better and buffer more data so that the disk can be read easier. In mode 2 you are fully dependant on the drive firmware to supply the correct sector, there is no confirmation in the "extra" bytes.

[SirDavidGuy]

@amni:

No offense, but you're full of crap (I don't know how that couldn't be offensive, actually).

After CD burned retrieving erronous-suspected data can be done
only by re-reading more carefully the suspected areas
on CD; error correction scheme may accelerate the process of
re-reading by better "focusing" on the suspected areas on the
CD, using what you call "positional" data.

Simply not true.

True correction can be
done only on-the-fly while the burning CD process continues;
This is only true of CIRC. ECC/EDC can be written (and read) RAW by any MMC compliant burner.

You cannot make wild assumptions about the way that the error correction on CD's works, while not examining it yourself. Try reading Annex's A and C of ECMA-130.

[amni]

unless I am be asked to design low-level
burning software (which handles EEC).


Originally posted by SirDavidGuy
@amni:

No offense, but you're full of crap (I don't know how that couldn't be offensive, actually).



Simply not true.


This is only true of CIRC. ECC/EDC can be written (and read) RAW by any MMC compliant burner.

You cannot make wild assumptions about the way that the error correction on CD's works, while not examining it yourself. Try reading Annex's A and C of ECMA-130.

[mpucoder]

That's funny - talking about mathematical principles that you are not familiar with. As a mathematician you should be dying to know how modular math works, and what it can accomplish. Have you never heard of Gauss or Fermat? And as far as computer error detection and correction goes, try looking up Hamming, and Reed-Solomon. It is mathemetically possible to determine the location and content of a burst of errors in a data stream using very little overhead. Total redundancy, btw, is useless - which copy is correct?

[theReal]

So you cannot use Mode1 on a multisession disc, but you can use it on single session discs with or without Joliet.

That's not true. You can definitely write multisession Joliet CD's in mode1. The only thing you cannot do is mix mode1 and mode2 on multisession discs (obviously).
Not only the Nero help file says so, but I have burned many many multisession discs in mode1, I definitely know it is possible.

[amni]

I don't think that you say something which contradicts
what I have said. I use my terminology of "on-the-fly"
correction because I'm not familiar with
the terminology you use. I think that "correction of
burst problems" is included in what I call
"on-the-fly" correction, and I never said this is not
feasible to be done in the "extra" bytes of 2352 sectors
(beyond the 1024 bytes alloted to main data in that sectors).
Also I agreed that it is possible to write additional
"location data" in these extra bytes, so that re-reading
can be made more efficient.

My claims were againts the myth about the bad durability
of binary data, written in MODE2FORM2.
As a matter of fact, I announced in this forum
about 4 months ago that
I filled 6 CDs of 80 minutes with about 800 MB (ZIP files)
in MODE2FORM2 and there were ZERO (!!!) errors
(I tested them using the CRC of ZIP files).
I checked some of these CDs 3 months after writing them
and the errors of writings are still ZERRO (!!!).




Originally posted by mpucoder
That's funny - talking about mathematical principles that you are not familiar with. As a mathematician you should be dying to know how modular math works, and what it can accomplish. Have you never heard of Gauss or Fermat? And as far as computer error detection and correction goes, try looking up Hamming, and Reed-Solomon. It is mathemetically possible to determine the location and content of a burst of errors in a data stream using very little overhead. Total redundancy, btw, is useless - which copy is correct?

[SirDavidGuy]

I think that "correction of
burst problems" is included in what I call
"on-the-fly" correction.

What exactly do you define "on-the-fly" correction as?

I never said this is not
feasible to be done in the "extra" bytes of 2352 sectors
(beyond the 1024 bytes alloted to main data in that sectors).

There is 2048 bytes of user data (which you call "main").

Also I agreed that it is possible to write additional
"location data" in these extra bytes, so that re-reading
can be made more efficient.

It is possible, but not done.

I checked some of these CDs 3 months after writing them
and the errors of writings are still ZERRO (!!!).

Proving nothing. All that you showed was the CIRC was sufecient to compensate for any errors created in the burning process and received during storage.

[mpucoder]

OK, seems to be some confusion on what mode 2 means. I'm not exactly sure myself what the correct name for it is, but I think this thread started out talking about the disks made by "mode2cdmaker". Possibly mode 2 isn't the correct name, but I think that's what started the thread.

Just so no one misunderstood me or anyone else, CIRC is present at all times, and has the ability to correct a fair sized error burst. But the 2048 byte sector that we've been using for years used the additional space in the audio 2352 byte block to enhance the error processing.
At about 35 cents a CD I'll gladly give up the 12% gained by "mode 2" for the huge increase in error recovery. Like most people in the real world, I'd like to think I'll never scratch a CD, or expose it to too much sunlight, but stuff happens.
(Besides, I own a DVD writer, what's a mere 150M?)

[DeXT]

@TheReal: Well I'm just quoting the specs, which I usually trust, and in the specs it's said that you cannot write a Multisession disc in Mode 1.

If any software allows you to do it, it's out of the specs. When a standard is set, it usually means that anything outside these specs is just not guaranteed to work on every case.

In any case I tried it with Nero, and found the following: choose "Start multisession disc", go to ISO tab, select Mode 1 , go to any other tab, then back to ISO tab, you'll find Mode 2/XA checked again.

[theReal]

I just tried it with Nero 5.5.8.2:
I choose "start multisession", go to the "ISO" tab and "mode 1" is already checked by default. It stays checked no matter what else I click.

I found a list of possible CD formats here http://www.copytec.net/cdr_formats.htm , and it even seems to say multisession mode 1 was described in the orange book standard.
The same list is available here (http://www.tec.informatik.uni-rostock.de/(nobg)/RA/MKA/massstor/optical/cdnormen.htm#tabelle), only it's in German.


Another interesting statement I found:

"If multisession recording is essential, then use the CD-ROM Mode 1 format, and accept the fact that the discs may not be readable in a few older CD-ROM drives. Do not jeopardize quality by using CD-ROM/XA for other than multimedia. Never use CD-ROM/XA Mode 2/Form 2 for data. Instead, focus on achieving interchange and longevity in the enormous number of CD-ROM drives manufactured since 1991. Do not compromise quality by attempting to achieve interchange in a few legacy drives" (from http://www.mscience.com/faq62.html )

[DeXT]

Seems every source says a different thing according to this subject. It's hard to find a good source without having the original (orange book) specs. However as it's stated in that article (which I found a bit radical at some subjects) seems Multisession has been highly tighted to Mode2/XA at the beggining at least. This probably was corrected in subsequent revisions of the orange book specs, but seems this is why some sources state that Mode2/XA is a must for MS, like the Joliet specs.

This also applies to CD recording software and hardware manufacturers. Some of them simply does not allow you creating MS discs with Mode1 tracks. My CD-RW drive (Acer 24x) is an example.

Do not compromise quality by attempting to achieve interchange in a few legacy drivesI don't agree here of course. Using Mode2/Form1 (which is the format of choice for data) never compromises quality. It's only in Form2 when it does, as it's already known (due to the lack of ECC). This is why it's targeted for multimedia content only.

[theReal]

I don't agree here of course. Using Mode2/Form1 (which is the format of choice for data) never compromises quality.I think that's what the author of the article meant, too. A sentence later he saysNever use CD-ROM/XA Mode 2/Form 2 for data.

I understood that the first multisession was introduced as mode2/form2. So, writing multisession mode2/form1 doesn't make any more or less sense than writing multisession mode 1, does it?

[amni]

"on-the-fly correction": means "correction done by burner while burning",
this is different from "correction after burning"
(eg, by using a second backup).
Again, I'm not taking about tecnology in use,
only taking about what is feasible.

"1024 ...": my typo error I meant 2048.

My experiments with MODE2FORM2: they prove that
MODE2FORM2 handles well errors (at least for
data which is not of cricial importance, eg: movie files).


Originally posted by SirDavidGuy


What exactly do you define "on-the-fly" correction as?



There is 2048 bytes of user data (which you call "main").



It is possible, but not done.



Proving nothing. All that you showed was the CIRC was sufecient to compensate for any errors created in the burning process and received during storage.

[mpucoder]

The term "on-the-fly" has, for over twenty years, meant the detection and correction of an error without a re-read or retransmission, i.e. the application of FEC.
So, maybe to help us old farts understand your point, could you coin another term for whatever process your discussing?

[midiguy]

I use mode 10 to burn all of my CDs. It has full error correction and detection. not only that, I can snap my CD into 2 peices and poke holes through each half, and it will STILL read the CDs PERFECTLY!

*midiguy laughs like a computer nerd*