Protection format names

[hooverphonique]

MFM encoding is not quite that simple, and doesn't allow any adjacent one-bits.. the disk controller uses the encoding as a clock when reading back the data in order to detect the bit rate. Also, the problem of adjacent ones is not an Amiga problem; MFM encoding was used on many floppy and harddrive systems back then...

If JW just got lucky that the controller would read back his strange encoding, I don't know.

[dlfrsilver]

the fact is that he used a correspondance table. that how he is reconstituting the 8 bits of data from the 6 bits of data saved and crammed on disk.

imagine, it's like if i said in a table that 00101111 must be saved on disk as 01111.

Of course saving 01111 instead of 00101111 takes less space on disk.

It's like he used the MFM bit encoding system as a compression system directly

[ReadOnlyCat]

Originally Posted by dlfrsilver

the fact is that he used a correspondance table. that how he is reconstituting the 8 bits of data from the 6 bits of data saved and crammed on disk.

I think that part is not in question. What NorthWay and hooverphonique probably meant is that in order to reconstitute the clock signal the stream of bits must respect some constraints such as making sure that the flux reversals happen within a certain frequency range and they assume that JW's encoding must somehow respect both the lower and upper limit of this range.

Originally Posted by dlfrsilver

Of course saving 01111 instead of 00101111 takes less space on disk.

It's like he used the MFM bit encoding system as a compression system directly

Actually he did not use MFM at all. He replaced it with his own encoding which is a bit more compact than MFM.
MFM(1,3) which is used on the Amiga is not the most compact encoding available: the wikipedia page on the topic lists RLL 1,7 which has the same information density as the one used by JW (three encoded bits for two signal bits).

Encodings used on floppy disk (and other streamed media) must satisfy two constraints:

• Avoid too frequent flux reversals which might exceed the magnetic media resolution.
  This is why it is "forbidden" to have consecutive 1s: they might be incorrectly written.
• Have frequently enough flux reversals that the clock drift can be detected and corrected "live".
  This is why too many zeros are bad, they prevent clock resynchronization.

In short: consecutive 1s are bad but consecutive 0s are bad too.

The many 0s puzzled NorthWay but I guess clock drift does not happen that fast so maybe it is fine to have quite a few zeros every now and then. I guess JW must have been experimenting with different combinations of bits and found out that the media density was a more constraining factor than clock drift.

[hooverphonique]

Exactly ROC,

it's still interesting that it works, though, since the disk controller must know the encoding in use to actually know how to retrieve the clock. When you program it (i.e. the hardware), you have the options of MFM or GCR, which tells me that the controller relies on the encoding, at least to some degree.

[ReadOnlyCat]

Originally Posted by hooverphonique

Exactly ROC,

it's still interesting that it works, though, since the disk controller must know the encoding in use to actually know how to retrieve the clock. When you program it (i.e. the hardware), you have the options of MFM or GCR, which tells me that the controller relies on the encoding, at least to some degree.

Oh, interesting, I did not know that.
I guess the hardware relies on derivative properties of MFM/GCR which JW's encoding probably exhibits too, likely accidentally though.

Update: actually, I think I understand why the encoding is needed by the controller. The section about densities of the Wikipedia RLL page seems to give a hint about it: since encoding can differ in compactness, that influences how the clock should be deduced. An encoding which guarantees that each 1/reversal is followed by two 0s/non-reversal can effectively store three encoded bits on the same surface as an encoding which guarantees only one 0/non-reversal in-between. Thus, the former encoding's clock will tick three times when the latter ticks two times. The controller must be informed of it otherwise it will misinterpret the stream of flux reversals.

Update2: the book "Amiga disk drives inside and out" by Abacus/Data-Becker actually explains that when GCR is set the reading/writing speed is halved. This makes sense since GCR allows consecutive 1s to be written which would exceed the resolution of the magnetic medium. MFM 1-sparseness on the other hand probably allows it use the maximum read/write speed.

The problem is that several one bits can follow each other in this format and the controller cannot process these. This has the result of changing the data recording density. To work with the GCR format without error, a switch is made to half (sic) the recording speed (from 2ms to 4ms).

So the controller probably does not need to know the encoding, it probably re-clocks using the same technique all the time.

Update3: I noticed that there is also a GCR dedicated bit in

ADKCON

in addition to the speed selection one so maybe the controller does need to know about the encoding after all. So mysterious.

Also, could you just refer to me as "Cat"? I am not too sure why but "ROC" feels horribly impersonal and cold.

[Keir]

Originally Posted by dlfrsilver

basically, when you have an alternate version, the whdload slave doesn't work with the game.

Yeah Gremlin name is a bad idea, because Gremlin, us gold, and even psygnosis used the protoscan longtrack protection.

One thing I'm missing in my Disk-Utilities is proper/accepted names for the various protections, where they have such a name. It took a while to work out what Protec was (some game had a fragment of the Protec source on disk, which gave that one away). I had no idea about Protoscan... And it's not like this info is easy to look up Sometimes WHDLoad notes give it away...

I will fix this one at least

[BarryB]

Sometimes the Softpres WIP pages give a particular protection/format a name, it's just trawling through all the WIP pages to find them!

[BarryB]

Two examples:

B.A.T. is identifed in Disk-Utilities as bat_longtrack, Softpres WIP: Protec variant on BAT, named Prolance by the game.

Batman The Caped Crusader and The Untouchables are identified in Disk-Utilities as batman, Softpres WIP: SFX, SpecialFX: variants on Batman The Caped Crusader, Head Over Heels, Midnight Resistance, The Untouchables

[dlfrsilver]

Originally Posted by BarryB

Two examples:

B.A.T. is identifed in Disk-Utilities as bat_longtrack, Softpres WIP: Protec variant on BAT, named Prolance by the game.

Batman The Caped Crusader and The Untouchables are identified in Disk-Utilities as batman, Softpres WIP: SFX, SpecialFX: variants on Batman The Caped Crusader, Head Over Heels, Midnight Resistance, The Untouchables

indeed. Prolance format on B.A.T., and SpecialFX MFM longtrack format.

[BarryB]

Plus the multitude of early Rob Northern protections that have no 'official' name!

[Keir]

Originally Posted by BarryB

Two examples:

B.A.T. is identifed in Disk-Utilities as bat_longtrack, Softpres WIP: Protec variant on BAT, named Prolance by the game.

Batman The Caped Crusader and The Untouchables are identified in Disk-Utilities as batman, Softpres WIP: SFX, SpecialFX: variants on Batman The Caped Crusader, Head Over Heels, Midnight Resistance, The Untouchables

Thanks for those. Fixed.

[BarryB]

Originally Posted by kaffer

Thanks for those. Fixed.

Super

I'll see if I can find some more!

[BarryB]

Looking at formats supported in Disk-Utilities and identifying 'variants' so they can all be grouped together under one umbrella:

Seems DMA Design formats were dubbed DMA DOS: DMA DOS: various DMA Design titles, found on Beast 3 and Lemmings so far.
DMA DOS, a different one: Menace, Blood Money (encrypted). Alternate DMA DOS format found on Lemmings Demo

archipelgos format seems to be called LGO or Logotron: LGO, Logotron: variants on Archipelagos, Kid Gloves, Gauntlet 2, Golden Axe.

typhoon format on Typhoon Thompson, Prince of Persia is named Broderbund: Prince of Persia renamed to Broderbund as it is used on the US version of Typhoon Thompson.

My dumps of Puffy's Saga/Iron Lord (not supported) identified as Ubi Soft: Ubi Soft protection found on RanX, Iron Lord, Puffy’s Saga and possibly all of their own original releases.

barbarian_ultimate_warrior protection identified on my dumps of Buggy Boy, Football Manager 2 and Three Stooges (although Three Stooges takes ages to pass the protection checks!). Seems to be a Rob Northen 'flakey bit' protection: Flakey bit protection on the original European releases of Barbarian (Palace), Armageddon Man, Buggy Boy, Three Stooges, Wizball, and an alternate version on ECO (Ocean). Likely to be other games too.

A few games with the apprentice format can be renamed CHW (My dumps of Gazza II, Flip-It & Magnose, Star Goose and 9 Lives also use the same format): Apprentice: Another CHW variant. CHW (alternate), on Conqueror.

scooby_doo, hi_tec_b appears to be a Gary Antcliffe format: GA, Gary Antcliffe: Scooby-Doo & Scrappy-Doo, Universe. Yogi’s Great Escape: A Gary Antcliffe format without a checksum.

Also, some formats were either named after the author, game or company if no other name was identified!

Also, maybe these 'format/protection' discussions should be moved to kaffers original thread: Protection format names

[dlfrsilver]

Originally Posted by BarryB

Looking at 'variants' so they can be grouped together under one umbrella:

Seems DMA Design formats were dubbed DMA DOS: DMA DOS: various DMA Design titles, found on Beast 3 and Lemmings so far.
DMA DOS, a different one: Menace, Blood Money (encrypted). Alternate DMA DOS format found on Lemmings Demo

Archipelgos format seems to be called LGO for Logotron: LGO, Logotron: variants on Archipelagos, Kid Gloves, Gauntlet 2, Golden Axe.

'typhoon' format on Typhoon Thompson, Prince of Persia is named Broderbund: Prince of Persia renamed to Broderbund as it is used on the US version of Typhoon Thompson.

A few formats can be renamed CHW (Gazza II and Star Goose also use the same format): Apprentice: Another CHW variant. CHW (alternate), on Conqueror.

Also, some formats were either named after the author, game or company if no other name was identified!

Also, maybe these 'format/protection' discussions should be moved to kaffers original thread: Protection format names

CHW = Christian Weber from SCA

[BarryB]

Another case where no name is known for the format so it's named after the author!

[DamienD]

Originally Posted by BarryB

Also, maybe these 'format/protection' discussions should be moved to kaffers original thread: Protection format names

Done

[BarryB]

Originally Posted by DamienD

Done

You are soooooo helpful Damien

[jotd]

IIRC, the MFM format of Agony & Unreal disks is the same (disks start by "ur00"). I don't know if the format has a name though.

[BarryB]

Originally Posted by jotd

IIRC, the MFM format of Agony & Unreal disks is the same (disks start by "ur00"). I don't know if the format has a name though.

From SPS WIP 18 July 2002: Ordilogic: Found on Unreal, Agony.

[Hewitson]

Originally Posted by BarryB

Plus the multitude of early Rob Northern protections that have no 'official' name!

Rob Northen. Rob Northern was a member of Classic