Dead Space Audio File.

OrangeC · Post by **OrangeC** » Sat Oct 25, 2008 4:15 am

U don't get this audio file, it seems that its compress, its certainly not a mus file.

http://www.megaupload.com/?d=X0JN5VQ2

Here is a sample.

Pepper · Post by **Pepper** » Wed Oct 29, 2008 12:28 pm

Very interested in this aswell, there are refrences to SBNK's in the str files, and I can't figure out if it's compressed or just an uncommon format.

btw, TrID says it's a generic xbase database, but I think it just doesnt have the right definitions for this type of file.

CUE · Post by **CUE** » Wed Oct 29, 2008 6:13 pm

I'm not sure, but looks like a TrueSpeech sound file.

Pepper · Post by **Pepper** » Wed Oct 29, 2008 11:36 pm

I highly doubt they would use true speech, it's primary use is very compact and "emotionless" speech. it does some special thing with tones to make it smaller and sound bland. it also is a very bad choice for a horror game where everybody's screaming and shooting. i think it's more suited for text to speech apps and other things.

OrangeC · Post by **OrangeC** » Thu Oct 30, 2008 6:41 pm

I got info from another member that the file is an interleaved EA ADPCM thats headerless. But again i don't know how to decode it.

Darkfox · Post by **Darkfox** » Mon Nov 03, 2008 8:00 am

STR can refer to "Stream" and "SBNK" is most often a reference to "Sound Bank". I'll give a look-see.

I can hear music when trying to read it as raw music but it is too noisy. A little more would need to be done to convert this properly.

OrangeC · Post by **OrangeC** » Mon Nov 03, 2008 1:50 pm

Okay kataah told me it is basically a version of EA ADPCM_r3 thats headerless, its a codec that was used in Tiberium wars and nobody could decode it.

Pepper · Post by **Pepper** » Mon Nov 03, 2008 11:19 pm

Code: Select all

01252     case CODEC_ID_ADPCM_EA_R1:
01253     case CODEC_ID_ADPCM_EA_R2:
01254     case CODEC_ID_ADPCM_EA_R3: {
01255         /* channel numbering
01256            2chan: 0=fl, 1=fr
01257            4chan: 0=fl, 1=rl, 2=fr, 3=rr
01258            6chan: 0=fl, 1=c,  2=fr, 3=rl,  4=rr, 5=sub */
01259         const int big_endian = avctx->codec->id == CODEC_ID_ADPCM_EA_R3;
01260         int32_t previous_sample, current_sample, next_sample;
01261         int32_t coeff1, coeff2;
01262         uint8_t shift;
01263         unsigned int channel;
01264         uint16_t *samplesC;
01265         const uint8_t *srcC;
01266 
01267         samples_in_chunk = (big_endian ? bytestream_get_be32(&src)
01268                                        : bytestream_get_le32(&src)) / 28;
01269         if (samples_in_chunk > UINT32_MAX/(28*avctx->channels) ||
01270             28*samples_in_chunk*avctx->channels > samples_end-samples) {
01271             src += buf_size - 4;
01272             break;
01273         }
01274 
01275         for (channel=0; channel<avctx->channels; channel++) {
01276             srcC = src + (big_endian ? bytestream_get_be32(&src)
01277                                      : bytestream_get_le32(&src))
01278                        + (avctx->channels-channel-1) * 4;
01279             samplesC = samples + channel;
01280 
01281             if (avctx->codec->id == CODEC_ID_ADPCM_EA_R1) {
01282                 current_sample  = (int16_t)bytestream_get_le16(&srcC);
01283                 previous_sample = (int16_t)bytestream_get_le16(&srcC);
01284             } else {
01285                 current_sample  = c->status[channel].predictor;
01286                 previous_sample = c->status[channel].prev_sample;
01287             }
01288 
01289             for (count1=0; count1<samples_in_chunk; count1++) {
01290                 if (*srcC == 0xEE) {  /* only seen in R2 and R3 */
01291                     srcC++;
01292                     current_sample  = (int16_t)bytestream_get_be16(&srcC);
01293                     previous_sample = (int16_t)bytestream_get_be16(&srcC);
01294 
01295                     for (count2=0; count2<28; count2++) {
01296                         *samplesC = (int16_t)bytestream_get_be16(&srcC);
01297                         samplesC += avctx->channels;
01298                     }
01299                 } else {
01300                     coeff1 = ea_adpcm_table[ *srcC>>4     ];
01301                     coeff2 = ea_adpcm_table[(*srcC>>4) + 4];
01302                     shift = (*srcC++ & 0x0F) + 8;
01303 
01304                     for (count2=0; count2<28; count2++) {
01305                         if (count2 & 1)
01306                             next_sample = (int32_t)((*srcC++ & 0x0F) << 28) >> shift;
01307                         else
01308                             next_sample = (int32_t)((*srcC   & 0xF0) << 24) >> shift;
01309 
01310                         next_sample += (current_sample  * coeff1) +
01311                                        (previous_sample * coeff2);
01312                         next_sample = av_clip_int16(next_sample >> 8);
01313 
01314                         previous_sample = current_sample;
01315                         current_sample  = next_sample;
01316                         *samplesC = current_sample;
01317                         samplesC += avctx->channels;
01318                     }
01319                 }
01320             }
01321 
01322             if (avctx->codec->id != CODEC_ID_ADPCM_EA_R1) {
01323                 c->status[channel].predictor   = current_sample;
01324                 c->status[channel].prev_sample = previous_sample;
01325             }
01326         }
01327 
01328         src = src + buf_size - (4 + 4*avctx->channels);
01329         samples += 28 * samples_in_chunk * avctx->channels;
01330         break;

is this what you're talking about? because according to the ffmpeg docs it supports it.

OrangeC · Post by **OrangeC** » Mon Nov 03, 2008 11:25 pm

Yes but the problem is that these type of R3 files are headerless, FFMPG only supports EA ADPCM with a header, but im not to sure,

Code: Select all

00000000 03 00 00 08 00 FA 00 00 20 00 00 00 00 00 00 00 ........ .......
00000010 04 1C BB 80 60 2E E0 00 00 00 00 00 00 00 00 00 ....`...........
00000020 00 00 07 28 00 00 01 80 A1 F8 05 00 A2 07 05 08 ...(............
00000030 B2 EF 26 EE 12 F8 16 FA 0F 10 3E 13 00 10 EF 22 ..&.......>...."
00000040 22 1F 00 FE 10 F1 DD 3E 0E 00 0E D0 FE 10 C0 22 "......>......."
00000050 DD FE 35 01 EE 01 FC 14 FF 0F DD 0C 01 13 00 CE ..5.............
00000060 12 EE FE 21 32 0F CE 0D 12 FF 43 24 32 0B 21 53 ...!2.....C$2.!S
00000070 34 F0 FF 33 F1 FC 06 00 32 07 76 08 02 F4 16 F3 4..3....2.v.....
00000080 82 04 F5 05 0F FF 04 0C 00 00 1E E3 22 1F 10 00 ............"...
00000090 10 D1 76 03 FF 50 1E 33 FF DB 43 FD FD 0E DE C2 ..v..P.3..C.....
000000A0 ED BC DF 32 DC FE 00 FE F2 FF 02 00 22 CD 11 21 ...2........"..!
000000B0 22 32 CD FF 13 50 35 03 56 E0 20 4E 47 2F CF F1 "2...P5.V. NG/..
000000C0 D0 FF 55 00 A1 00 C4 FF B1 FB F4 FE F1 FC 14 FF ..U.............
000000D0 0F 11 1E 12 E0 0D 21 FC F3 33 EF 42 5F D1 10 D1 ......!..3.B_...
000000E0 C0 00 E1 00 1F 0E 1E 10 0F F1 10 02 13 FF E2 1E ................
000000F0 DE 40 0B 10 2C E1 24 C2 D0 00 E0 2F F0 1F 1D 12 .@..,.$..../....
00000100 0B 02 C3 FF EE 1D 0E 20 F2 FE 10 01 81 FF 65 01 ....... ......e.
00000110 21 01 C4 00 41 00 75 FC E1 00 45 00 C1 21 E1 13 !...A.u...E..!..
00000120 D2 D1 1E E5 40 5C 03 3B FF E4 0B 01 10 FE 05 3F ....@\.;.......?
00000130 E0 0F C1 E0 2F 11 2B 30 0C F0 40 C4 11 0F B3 ED ..../.+0..@.....
00000140 B2 01 4E 50 01 11 01 F2 2C F0 D1 31 24 1F 0D F1 ..NP....,..1$...
00000150 E5 E0 22 02 DF D0 10 0B B1 FF D6 00 C2 00 E7 00 ..".............
00000160 02 01 78 00 22 F9 77 F9 DC 02 4E 20 02 E0 F7 31 ..x.".w...N ...1
00000170 23 0E C7 20 02 1E BE 02 FF 01 BC 0D FE 22 FF FC #.. ........."..
00000180 0D 11 20 D0 FE 30 FF D2 FF FF 11 00 01 C0 EA 2D .. ..0.........-
00000190 10 B2 CF 1E 10 12 A1 01 12 11 D0 2E 00 1F CC 2D ...............-
000001A0 11 1F E1 20 81 FF 36 01 42 FF 78 FF 82 FE F7 FD ... ..6.B.x.....

Pepper · Post by **Pepper** » Tue Nov 04, 2008 12:01 am

I know, but is there any way to rebuild the header? or any info on what it normally looks like?

Pepper · Post by **Pepper** » Tue Nov 04, 2008 12:25 am

http://wiki.multimedia.cx/index.php?tit ... udio_Codec

appearently it's XAS. and headerless versions have the headers stored elsewhere in the data.

OrangeC · Post by **OrangeC** » Tue Nov 04, 2008 1:30 am

HMM interesting.

Kataah a member that studied the file says that it has certain properties to ogg codec. and that the stream is also interleaved, its very complicated.

Pepper · Post by **Pepper** » Tue Nov 04, 2008 5:01 am

should be similar to mp3, the files are normally ea layer3 (cant find much on it) when headerless.

Kataah · Post by **Kataah** » Tue Nov 04, 2008 10:08 pm

The contents of this post was deleted because of possible forum rules violation.

Pepper · Post by **Pepper** » Wed Nov 05, 2008 2:08 am

lately i've just been looking through the game files for the headers, I've had shit luck, but I havent covered nearly all of it. I'm sure the game has a method for loading the 16byte headers somewhere.

edit: I know what you mean, XAS has 16 byte headers then 76byte data chunks alternating between the channels or something.

XeNTaX

XeNTaX

Dead Space Audio File.

Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.

Re: Dead Space Audio File.