ECM using - is it necessary?

[Smiling Spectre]

Today I test ECM, that, as it seems for me, very often used on this site.
The results seems very discouraged.
I tested NRG of Resident Evil, CUE/BIN with soundtracks of Gateway 2, ISO of Heart of Darkness and another CUE/BIN without soundtracks for Space Empire 5.
I ECMed file, then RARed it. And also I RARed it directly. I used WinRAR 3.70.
On every file, there was gain after ECM. Result file was smaller.
And every time result file was smaller... just on number, that ECM was able to extract from image!
In every time it was about 500-700 bytes. Except Gateway 2, where ECM was whole 5MB from 31.

So I can make beforehand conclusions:
1. ECM is absolutely uneffective on any file, except those with soundtracks.
2. ECM not helps to compression at all.

Can anyone prove or refute this?

[MasteromaN]

Uploaders aren't idiots. If they notice a gain, some of them will use ECM. I don't know what example I could give you. I think the ISO of Project IGI rared and ecmed is 30MB smaller than if it was only rared. If the gain was only 10MB I'd have just rared the ISO.

[hfric]

you are one funny guy Smiling Spectre
read here and then spit your hate
http://www.neillcorlett.com/ecm/how.html
knowhow is the key

ECM (Error Code Modeler) lets you prepare CD image files (such as BIN, CDI, NRG, CCD, or similar)

in english : it is effective for images with EDC and ECC (alcohol images and others that dump the hole CD/DVD count too but not those that dump only data like ISO), don`t you compare ISo to Bin - it has none or little error correction code

[Smiling Spectre]

MasteromaN, I never take other peoples as idiots, at least while they not prove it. If my English is so bad, and suggest this conclusion, I am sorry. I don't mean anything like this.

What I mean is
a) I am totally wrong, and there is something bad with my testing. But I don't see, what exactly.
or
b) ECM could be effective before, but now WinRar overcome it - and no one notice this yet.

That's all.

hfric I could be funny very often, yes. I know, that I cannot know everything, and I don't fear to ask stupid question, if I don't see clear answer on it. Even if it prove my stupidity.

As you can see, I tested four different images. Totally different - from nrg to bin/cue. It's all types of files, that I have - only mds and ccd was not tested, because I haven't it at hand. No one file shows significant reducing in size - except those with soundtracks. And no one shows any progress in packing with WinRar.
I read ECM pages before testing it - it seems very convinceable for me. But facts says, there is no advantage in using ECM for packing (but there is some advantage in raw size with multi-tracking images). So it seems for me, there is no point to use ECM for any file, except such multi-tracking, and also that modern packer (such as WinRar) already can pack effectively files without ECM.
Of course, it all can be horrible incident, and this was only four images, that not benefit from ECM, yes. But at present moment 7zip seems much more effective for me (it gives me at least 2-3Mb, not 100-200Kb in advance, and not required double packing/unpacking

What can you say about it?

[hfric]

no one forces you to download our thing , don`t wanna then don`t get it simple as that

[Virgil]

b) ECM could be effective before, but now WinRar overcome it - and no one notice this yet.

Now? Packing algorithm, which is used in rar is indeed much older, than that utility. Difference in size with ecm'ed images is quite noticeable - around 100 megabytes from one full(650-700 Mb) CD-image. I think, since you wrote somewhere, that you on a lousy internet connection you would appreciate anything, what will reduce size of download.

[I_AM]

The effectiveness of ECM largely depends on the format of the image you wish to compress.
If you have a MODE1 image of 2048 bytes per sector, usually an .iso file, then ECM will have no effect on it.
For a MODE1 image of 2352 bytes per sector, such as most .bin files, the compression ECM+RAR gives should be better than just RAR. That extra 304 bytes per sector is the data that ECM works with, and should do a more efficient job on it that RAR alone.

7-zip seems hit and miss to me. Sometimes it is wildly better than RAR, and others falls short

[Smiling Spectre]

hfric, don't be offended, please! If you'll understand Russian, you'll see, that I could try to express my doubt in all, that I say. In English it seems harder. :/ I'll try this again: I don't want to offend anyone. I never think, that you all on this site did something wrong (because I think, this site is great . I only want to clear my understanding - in that case with ECM - and understand, when it have sense to use it and when it's not. It seems, my provocative tone in English seems more sure, then it's in Russian, so I repeat it clearly: all my conclusions (especially those, I myself marked as "beforehand") is not final in any means, but only points, that I want to clear.

Virgil, you are right about WinRar. Then I don't understand, why I cannot achieve any progress with compression.

I_AM, is this exist simple method to detect images, that could potentially be packed better after ECM? Because now I tested already 14 images, blind-packed six of them, and no one make anything!

2All
I tested another ten images at home. Selected from nrg, cue/bin and mds. I make some progress with conclusions.
1. I was totally wrong about this "tracks" thing. I ECMed images with tracks and without it - it shows no dependance. Incubation: Wilderness Missions in NRG have whole 400Mb of soundtracks, but cannot be packed, and Total Annihilation: Battle Tactics in mds haven't tracks at all, but lose 50Mb after ECM. So it seems for me, there is no simple method to decide, if file could be ECMed or not.
1.2. Another beforehand conclusion: I notice, that all (two ) tested mds was ECMed almost to it's "data size". So if data was 600Mb, but image was 700, after ECM I received 605Mb. It could be criteria, IF it works always (or almost always - I remember about protection).
2. I blind-tested another five random images after ECM to packing ability with WinRar. It's still nothing. If file was 550Mb before ECM and become 500Mb after ECM, then when it was 300 after WinRar, it becames 250 after ECM+WinRar. In no case I manage to gain additional packing. In one case (it was Incubation: Time is Running Out - nrg) ECM manage to reduce file on 300 bytes, but ECM+WinRar makes resulting file on 1Mb bigger then simply WinRar.

I still not making final conclusions. I ask you: is it worth to continue testing? Is there exist cases, when WinRar in maximum compression makes file significally smaller after ECM, then difference between ECMed and nonECMed image? If answer is yes, I will try to decypher, when it worth to ECM all this files. If it is no...

[Egon68]

I blind-tested another five random images after ECM to packing ability with WinRar. It's still nothing. If file was 550Mb before ECM and become 500Mb after ECM, then when it was 300 after WinRar, it becames 250 after ECM+WinRar. In no case I manage to gain additional packing.

But still you get extra 50 Mb off the package. That's 16,6 % which is the
average gain with ecm. I made a test with DroneZ:
bin/cue = 365 Mb
with WinRAR = 336 Mb
with ecm = 318 Mb
with ecm+WinRAR = 292 Mb, 13 % smaller than without ecm.

Generally ecm works well with bin, mdf and img. Don't use it with nrg & iso.
Or don't use it at all! It's just that people usually have hd space & time
enough, but surprisingly many have very poor connections and if you save
10-20 % bandwidth, in my opinion that justifies the use of ecm.

[Smiling Spectre]

But still you get extra 50 Mb off the package. That's 16,6 % which is the
average gain with ecm. I made a test with DroneZ:
bin/cue = 365 Mb
with WinRAR = 336 Mb
with ecm = 318 Mb
with ecm+WinRAR = 292 Mb, 13 % smaller than without ecm.

It seems, my persistent non-ability to express my thinking clear and plain is even bigger on English. %)
1. I already understand this, I try to say this several times and I think this all this time: If ECM cut off significant hunk of image - it must be done. Yes!
2. Question is in another part: is this any sense in trying to RAR any file, if ECM not manage to cut off much? This is suggested as main purpose of ECM on author's site, but my testings not prove it. Never. Your too - file was ECMed by 47Mb, but RARed only to 44Mb difference.

[Virgil]

Товарищ, какую истину вы тут пытаетесь найти? Сказали же, что программа работает с bin/img/mdf, но не ISO и NRG. ECM давно используют на всяких консольных сайтах ради экономии места при больших коллекциях образов.

[Smiling Spectre]

Да можно и на "ты", я не обижусь. Я ж вроде сказал, ни? Я хочу понять, отражается ECM на уровне сжатия потом архиватором, как уверяет автор, или вот эти вот 50-60 метров, что выжмет ECM без архиватора - это и вся ее польза. Если первое - тогда узнать, когда именно оно отражается (это я уж как-нить сам ), ну, а если второе - просто иметь это в виду и пытаться выжать один раз - и все.
Кстати, NRG вполне себе выжимается - выжал же я 30 метров из Тотал Аннигиляции.

[hfric]

I blind-tested another five random images after ECM to packing ability with WinRar. It's still nothing. If file was 550Mb before ECM and become 500Mb after ECM, then when it was 300 after WinRar, it becames 250 after ECM+WinRar. In no case I manage to gain additional packing.

But still you get extra 50 Mb off the package. That's 16,6 % which is the
average gain with ecm. I made a test with DroneZ:
bin/cue = 365 Mb
with WinRAR = 336 Mb
with ecm = 318 Mb
with ecm+WinRAR = 292 Mb, 13 % smaller than without ecm.

and the Full Rip compressed in 7zip is 98MB

[Egon68]

hfric, we are discussing ISOs here...

[hfric]

yeah i know but its kinda funny that the full game (files from the ISO) compressed with UHA and then with 7zip take only 98mb and the ISO with the same files but compressed only WinRAR takes 336 Mb

makes you go dump 7zip in favor of UHA