This topic was archived here:

https://web.archive.org/web/20230000000000fw_/https://www.zenhax.com/viewtopic.php?t=2853

egab, posted Sat Jul 30, 2016 10:56 am (15915)

---

If I am not mistaken, the textures are in a file format called .ntx (or NTXR when viewed in a hex editor). I've added some samples down below if anyone is interested.

Acewell, posted Sun Jul 31, 2016 2:32 am (15964)

---

this thread should probably be in "Graphic file formats"

run this bms script on all your ntx files
*removed*
see this post
viewtopic.php?p=16241#p16241

then use this Noesis script to open the resulting ntx textures
*removed*
see this post
viewtopic.php?p=16241#p16241

your samples are actually archives, its just that 3 out of 4 only have one texture,
the 000000ad.ntx sample has 23 textures, the bms script extracts these and
strips the first 16 bytes from the others and will unify them so the Noesis script opens them all.

egab, posted Sun Jul 31, 2016 3:35 am (15966)

---

Acewell wrote:

this thread should probably be in "Graphic file formats"

My bad. Hopefully someone can move it over?

egab, posted Sun Jul 31, 2016 9:04 am (15977)

---

Acewell wrote:

this thread should probably be in "Graphic file formats"

run this bms script on all your ntx files

Code:

-snip-

Is it normal that I only get this:


I seem to get 0 byte .ntx files, should I wait for this to finish? Or did I do something wrong?

Acewell, posted Sun Jul 31, 2016 4:01 pm (15998)

---

i don't know, you'll have to upload a sample that doesn't work with it
i only had the 4 samples you shared before and it worked fine on all of them

egab, posted Mon Aug 01, 2016 9:38 am (16034)

---

Well, I used the script on the same four samples I gave you and it seemed to just spam 0 byte .ntx files.

Should I select each .ntx file individually(like using the script on one .ntx file at a time) or should I select all of them when using the script?

Acewell, posted Mon Aug 01, 2016 11:14 pm (16057)

---

you can run the bms script on each one at a time, or all at the same time, doesn't really matter
i had no problems with your original samples working with the scripts already shared here though

are all of these samples from the same game and platform?
your old samples had a 112 byte header
your new samples have a 144 byte header

naturally this difference will affect the success of the scripts
use this bms script on the new samples
*removed*
see this post
viewtopic.php?p=16241#p16241

and use the Noesis script to open those resulting ntx files
*removed*
see this post
viewtopic.php?p=16241#p16241

egab, posted Tue Aug 02, 2016 1:17 pm (16078)

---

I've been trying multiple times, and this is the result no matter which .ntx file I use on both versions of the scripts.



All options will export a 0 byte .ntx file and will go on for as long as it can.

I used the 2nd version of the script on the same samples I gave you last time.

Acewell, posted Tue Aug 02, 2016 8:24 pm (16087)

---

well, you gave 2 different types of samples and you didn't answer my question:
are all of your shared samples from Tekken 6 on Xbox 360?

your old samples have a 112 byte header and your new samples have a 144 byte header
the current scripts are not interchangeable because of this
you must use the right scripts with the correct sample set

i don't know why you having problems with the scripts, everything works for me,
maybe your paths are too long or something,
try dropping the samples and QuickBMS in the root of your drive and running from there.

egab, posted Tue Aug 02, 2016 10:14 pm (16100)

---

Yes, both samples are from Tekken 6 for the Xbox 360 version. But from different .bin files.

Acewell, posted Wed Aug 03, 2016 9:21 pm (16147)

---

egab wrote:

All options will export a 0 byte .ntx file and will go on for as long as it can.

I suppose this might be causing your zero byte files problem if you are using the latest version of QuickBMS
viewtopic.php?p=16116#p16116

i'm still using 0.7.3, so maybe you can use a previous version of QuickBMS to run these scripts
or remove the quotation marks at the end of the second findloc line until aluigi fixes the issue.

egab, posted Thu Aug 04, 2016 2:38 pm (16170)

---

Acewell wrote:

egab wrote:

All options will export a 0 byte .ntx file and will go on for as long as it can.

I suppose this might be causing your zero byte files problem if you are using the latest version of QuickBMS
viewtopic.php?p=16116#p16116

i'm still using 0.7.3, so maybe you can use a previous version of QuickBMS to run these scripts
or remove the quotation marks at the end of the second findloc line until aluigi fixes the issue.

The script works now on 0.7.3, I guess the problem is with Quickbms after all.

aluigi, posted Thu Aug 04, 2016 5:15 pm (16173)

---

Yes that's the problem.
People using 0.7.5 can simply add a -1 to the second findloc command: findloc NEXT_OFFSET binary "\x47\x49\x44\x58" 0 "" -1

egab, posted Sat Aug 06, 2016 1:57 am (16217)

---

The script (v2 one) doesn't seem to work with the attached .ntx files. These are all from the same .bin file as last time.

Acewell, posted Sat Aug 06, 2016 8:37 am (16223)

---

okay looks like a another header size and format we have to deal with
these scripts tested with the "samples 1.rar" and work for all
use this bms script on your newest samples
*removed*
see this post
viewtopic.php?p=16241#p16241

and use this v3 Noesis script to open the resulting ntx textures
*removed*
see this post
viewtopic.php?p=16241#p16241

i'll check the other samples against these new scripts later but i think we got it now
just need to find a way to combine them all into the same scripts and add auto-detection
for the header types so you don't need a bunch of them

egab, posted Sun Aug 07, 2016 4:22 am (16240)

---

Some textures are in an odd color when viewed. Could you take a look?

Acewell, posted Sun Aug 07, 2016 7:01 am (16241)

---

okay here are just 2 scripts you need instead of 6

*updated Feb 24, 2017*
use this bms scripts on all your ntx archives


*updated Sep 19, 2017*
then use this Noesis python script on the extracted xtex textures



both of those scripts are easily extendable if we run into more types of ntx files
tested on all your samples and only about 16 out of ~900 textures
wouldn't open in Noesis, i would still call that is a success


edit
i didn't see your new post until after i posted this one, but i saw those
green tinted images already and i played around with different combinations
of the rgb channels and i didn't see anything better looking than what you
see there. What are they supposed to look like?

edit2
i updated the Noesis script, see what you think of those green textures now

Tekken6_X360_ntx.zip

tex_Tekken6_X360_xtex.zip

egab, posted Sun Aug 07, 2016 7:28 am (16242)

---

I found an alternate file but with proper colors.



I messed with the RGB channels of those with incorrect color in a photo editing software, and had no success.

EDIT 1:



Now it's in a reddish color?

Acewell, posted Sun Aug 07, 2016 8:52 am (16243)

---

i updated the Noesis script again in my previous post, give that one a try
i disabled the byte swap for the rgba images and now they look better

egab, posted Sun Aug 07, 2016 8:59 am (16244)

---

Acewell wrote:

i updated the Noesis script again in my previous post, give that one a try
i disabled the byte swap for the rgba images and now they look better

Amazing! They all look proper now Could you take a look with the following samples, if there are textures that can be possibly extracted from these?

Acewell, posted Sun Aug 07, 2016 10:38 pm (16266)

---

Your last 3 samples are just 1280x716 dxt5 big-endian textures,
but the the height stored in the file is 720 and this is what is read by the script.
the others that won't open are dxt5 with probably incorrect dimensions too.
Noesis needs the exact amount of data to display the image or it will error.
I could change the script to subtract 4 from the height dimension to correct them,
but this will throw off other textures that already are correct, maybe there is a flag
or it might be best easier to just convert those by hand.

here are those 3 textures converted to png




edit
and here is those other samples that wouldn't open with the current Noesis script




edit2
i updated the Noesis script to open all of the valid textures extracted from your samples in this thread
viewtopic.php?p=16241#p16241

egab, posted Sat Sep 03, 2016 8:27 am (17187)

---

Hey! Thanks, I'm back again after encountering another problem (I think):



Looks like an alpha problem.

Acewell, posted Sat Sep 03, 2016 10:29 am (17191)

---

okay i updated the Noesis python script to support dxt3 textures now
viewtopic.php?p=16241#p16241

egab, posted Tue Feb 21, 2017 3:21 pm (20930)

---

There's a different format for the PS3 version it seems, attached a sample below.

It's an archive rather than an individual texture file like the .ntx