Understanding the files on the DVD of Arirang Karaoke (hk2 & midi0)

[bigboss97]

I've read the first 2000 song names from TLIST and marked with the MIDI file names as you provided
vn_mid.txt

HTH
Phuoc

[henry123]

vn_mid.txt just the song list
i din't see any ressult of marking to midi files

[bekon]

how is this relevant to what have been discussed?

henry123 wrote

if you can't see the lyric : just because of missing font

[bekon]

how is this going, would you mind update us and perhaps walk us to the right direction on extracting all the midi's and mp3's from a Arirang Midi DVD?

henry123 wrote

i'm having a vol36 disk
but no time to be with it.
I'll give all of you the result later.
Thanh Phuoc so much

[bekon]

i'm curious, what is the algorithm used to successfully decoded and extracted all the midi's and mp3's into a sequence and not to the designated song numbers? thanks.

henry123 wrote

i was spending about 3 days to get all of Arriang DVD KARAOK' original midi files,  if anyone can help me to arrange the midi source of vol 34 (all of arriange  Karaoke DVD midi and mp3 files : order number from 00001   ) . I'll give you alll the midi files as well as uploading for dowload to all for you.
i got a trouble in extracting the resource of LISTSONG.DAT FILE (I THOUHG IT A DATABASE FILE NOT A DVD FILE)
midi.move.JPG
mp3.JPG

picofkaraoke.JPG
ok.JPG
ok2.JPG
ok.JPG

[bigboss97]

There's no golden bullet. It's simply a lot of patience and luck
For MP3 see:
http://old.nabble.com/Arirang-MIDI-Karaoke-DVD-storage-file-struct-and-MP3-Extraction-p29129944.html

bekon wrote

i'm curious, what is the algorithm used to successfully decoded and extracted all the midi's and mp3's into a sequence and not to the designated song numbers? thanks.

[bekon]

i understand luck, but what we are talking about here is purely brilliant achievement. The whole purpose if anyone decided to engage in such a project is to gain experience and most important the love in doing.

Why not share the knowledge so others be-able to accomplish the same goal and benefit from it. Isn't that the whole purpose of public forum? I'm just disappointed that someone would just ignored and decided to abandon it. I do have reasons to doubt but that is not even the point.

bigboss97 wrote

There's no golden bullet. It's simply a lot of patience and luck
For MP3 see:
http://old.nabble.com/Arirang-MIDI-Karaoke-DVD-storage-file-struct-and-MP3-Extraction-p29129944.html

[bigboss97]

NOT SHARING???
All ideas were publicly discussed, even source codes were posted. Your questions got answered (within hours) and advised to existing posts. Usually you should be happy if you get an answer in the forum within one week.
If all these are not sharing knowledge why do I waste my time here?

bekon wrote

Why not share the knowledge so others be-able to accomplish the same goal and benefit from it. Isn't that the whole purpose of public forum? I'm just disappointed that someone would just ignored and decided to abandon it. I do have reasons to doubt but that is not even the point.

[bekon]

obvious this is a misunderstanding, i was speaking directly to the other mate, henry123. I believe i've shown my appreciation toward your consistent responses which is something i wouldn't expect.

your works are extremely amazing with tirelessly feedback posted base on your observations. So allow me to be cleared on that regard. I'm in the process of analyzing your notes and learning what i can. I'm rather surprised that there isn't a devoted community working on these karaoke systems since they are so popular other than what we see here.

[mrjenki]

Is anyone still discussing this topic?
Since last post was on Aug, I'm not sure if I post this on the right place.Anyway, I still wonder if has any one succeed trying to extract any MIDI from MULTAK.DAT? I've been trying to mess around with that file for a moment already and what I've found is not much.
I've found that the MIDI stored in the file does not have the same format with standard MIDI and by looping some with masks, I could reconstructed some MIDI parts for particular songs only.Till now I haven't got any clue where is mask actually is in each song and looping masks for all of the songs seems to be a painful process.
The MIDI data in this file seems to have a different structure which takes less space to store and uses extensively whole byte Delta-time not variable-length delta-time like in standard MIDI format.
Some of the songs have unmasked MIDI part, which may be because of coincidence when mask is a specific number.
All of the song I've reconstructed have complete MIDI data but I could find any MIDI event for Patch change , tempo and other variables.
 

[bigboss97]

mrjenki wrote

Is anyone still discussing this topic?
Since last post was on Aug, I'm not sure if I post this on the right place.Anyway, I still wonder if has any one succeed trying to extract any MIDI from MULTAK.DAT? I've been trying to mess around with that file for a moment already and what I've found is not much.

a big YES, you are right here.
What I can see... you already found a lot. I couldn't get a single note  :-(
Can you please explain your ideas and give all the details about what you already know. That's the only way to progress to next level.

[bht67]

Anyone interested in continuing?

[bht67]

Are you sure the mp3 files (nhac be) you posted were extracted from vol. 34?  I could not locate them in either the MULTAK.DAT or MULTAK.DA1 files.  However, the two files voice1.mp3 and voice2.mp3 you posted are there.

[Bho1668]

Hi Please update us on whether this method work?

[Bho1668]

HI,
I just wonder
1. How did you identify the MIDI block?
2. I understand you mentioned running some masks. But my question is how could you tell it is the masked data or the real MIDI data?
3. Is it a particular data structure for the MIDI data?

Thanks

[bigboss97]

Bho1668 wrote

2. I understand you mentioned running some masks. But my question is how could you tell it is the masked data or the real MIDI data?
3. Is it a particular data structure for the MIDI data?

If MIDI is not encrypted you should be able to see lyrics as plain text somewhere. That's usually not the case.
The MIDI structure was discussed in:
http://karaoke-engineering.44.s1.nabble.com/Understanding-the-HOTDOG-files-on-DVD-of-California-electronics-td16.html
It's not the MIDI file format.

You can analyze the data. But you can't do much until your suspicion confirmed:
1) Find a simple song, e.g. Happy Birthday. Try to locate it.
2) Get a DVD-RW and make a copy via ISO image and ensure the copy is working.
3) Locate the song in the ISO image and use HEX editor to patch one or two bytes based on the finding in (1).
4) Burn the disk and try to play the song. It might CRASH the device!!!! If you are lucky you may patch into lyrics and it would be visible on screen.
5) After you find the way to locate MIDI, you can start trying on song with lyrics like lalala...

These all can be found under:
http://karaoke-engineering.44.s1.nabble.com/Understanding-the-HOTDOG-files-on-DVD-of-California-electronics-td16.html

When you're more confident you can patch more songs in each patch.

[Ishisashi]

I have been trying to decipher karaoke DVDs for Kumyoung KDVD-1100 recently, but I can only extract the title information (the back of TLIST.DAT, but I don't even know what the front is) and have no clue about the MIDI and MP3 parts.
(It's obvious that Korean titles are encoded as EUC-KR and Japanese titles are encoded as EUC-JP.)
The file distribution is very similar to Arirang Karaoke DVD Vol. 34, so I believe that studying either one is helpful for the other.
(You can search for 'Kumyoung Vol. 66' on Internet Archive to get the image. I don't have the KDVD-1100 itself, so unfortunately I can't do real machine debugging.)

[Ishisashi]

OK I can see some SMFs (by searching for MThd) and MP3s (by searching for LAME)—unencrypted—in MUN****.DAT. But quantity is very small, I can only find 123 results in Disc B, MUN99603.DAT. Maybe there are masks.
I think that's also why henry123 was able to obtain data but didn't know the order; those are indeed SMFs, so he may have simply separated files through headers?

[Ishisashi]

I could clearly see the lyric with title before SMF block, so if the mask could be figured out, I don't think it would be a problem that pairing SMF and song…
It seem like a 24-bit integer is used to indicate the length of a content (lyric, or SMF)?

[bigboss97]

Ishisashi wrote

I could clearly see the title before SMF block, so if the mask could be figured out, I don't think it would be a problem that pairing SMF and song…

My problem is that I don't have the player. So I can't patch the content to figure out the mask.
Please share some screenshots of the HEX dump and also add information in the subject. That helps people to find your post.