54 réponses à ce sujet

[KingEtzel]

...if, perhaps, a developer could lend us a helping hand for this? I would then write the tutorial in the wiki, but I'm not moving anything forward atm.

[sherlockemery]

Unfortunately, I have also run into a brick wall. I would also like some help!!

[EJ42]

You have to be extremely careful how you name things. The game uses parts of the filenames to determine certain things. You can't just name your file bobs_new_file.bob, and expect it to work. If you want a new human male soundset, then your GDA needs to be named sse_ss_pc_mhuman_something.GDA. There may even be a limit to how many characters can be in the "something" part of the filename. A few of us figured out that adding new materials would not work if the GDA's filename had more than two characters after the last underscore, before the .GDA extension. Much of this is semi-trial-and-error.

[Sarevok Anchev]

EJ42 wrote...

You have to be extremely careful how you name things. The game uses parts of the filenames to determine certain things. You can't just name your file bobs_new_file.bob, and expect it to work. If you want a new human male soundset, then your GDA needs to be named sse_ss_pc_mhuman_something.GDA. There may even be a limit to how many characters can be in the "something" part of the filename. A few of us figured out that adding new materials would not work if the GDA's filename had more than two characters after the last underscore, before the .GDA extension. Much of this is semi-trial-and-error.

I ALSO HATE CGI-MORRIGAN!!!
oh..ehm... wrong thread

[KingEtzel]

Thankyou for contributing to this thread  !

(...), then your GDA needs to be named sse_ss_pc_mhuman_something.GDA. There may even be a limit to how many characters can be in the "something" part of the filename.(...)

Okay. Therefore I kept name conventions for the soundset itself - beginning with "ss_*". I also found those GDA files starting with "sst_ss_*" and "sse_ss_*", but what they are and how they are created, I don't know...perhaps that's part of the problem.

[Sarevok Anchev]

I got Fernando Melo. He said, that he will send someone with more understanding of this to the topic.

Lets see....

[sherlockemery]

Well,



I have just about given up guys. Haven't gotten any further on this. The sad part is my Ash from Evil Dead is ready to be used, I just can't get it to work. Files are converted & renamed, I just can't get the toolset side to function.



What a bummer.

[KingEtzel]

Haha, supposedly all the developers who could help here are working on the next DLC that will break the game ;D

[sherlockemery]

But my Ash voicepack could be that DLC!! It'd even be a free DLC to break stuff with!!!

[Dark_Ansem]

could it be that the problem depends on the fact that all audio files need to be put in .fsb file?

[sherlockemery]

Nope. I have them in the .fsb file, & the file does work (tested in a player).



It seems that the issue is we can't get the game to recognize the soundset because we can't get the naming convention in the toolset right.

[KingEtzel]

The DLG file for the soundset does not get associated right to the "Player" (we do this by adding an entry to the playersoundsets.GDA/XLS and then select our soundset during character creation... however, this doesn't work although this should be the way... funny, isn't it?) The soundset itself is a simple DLG that creates FSBs, just like any other dialog file.

Modifié par KingEtzel, 06 février 2010 - 03:15 .

[Dark_Ansem]

Bioware, help. why can't we get to do a custom voiceset?

[Malcroix]

Thanks for all the work you've been doing to make custom voice-sets in DA a reality. It's a really great feature of CRPGs, and was a very important element of character customisation in BG/NWN games. However, it seems DA, despite all of its supposed "openness", is a far cry from the old school CRPGs in terms of voice customisation.

Now that Awakening is upon us, a new interest in customisation is sparking. I can only hope, along with many other gamers, that the online community will be able to "crack" this code and finally give us custom voices (something which SHOULD have been made available from the start by BioWare).

My own measly 2 cps: back in the "old days", one of the more obscure elements of making voicesets was to get the sound characteristics right. Wiki says DA voice-overs are recorded as 24 kHz, 16 bit PCM. Are your custom files fully compliant with this?

Also, have you tried editing an existing voiceset instead of making a new one? Are those fsb/fev files openable/editable? If yes, it could cut a lot of corners. As far as I can see, sounds can be extracted from them (http://social.biowar.../index/385958/2), but so far no luck in opening them with the structure intact, right?

Good luck!

Modifié par Malcroix, 14 mars 2010 - 05:35 .

[Malcroix]

The more I think about it, the more I believe that editing/replacing an existing soundset is a much simpler task than creating a new one from scratch and then struggling to make the game recognise it.



After all, the game only uses one PC soundset at a time, so it would be no loss at all to replace one of them with a custom set. It would also be very easy to install - no need to edit the .xls, just put the new fev/feb file into the override and voila. It would also improve compatibility between different voice mods.



So, how about it? Is it possible to replace the sound in an existing voice set?

[Malcroix]

Status: Extracted wavs from a soundset FSB using this fine program:
http://aezay.dk/drag...tor09.11.28.rar
Files retained their original names & format, are now interchangeable & replaceable with custom wavs.

Putting individual wavs into Override didn't do the trick. Apparently they need to be recombined back into an FSB.

The FMOD is as worthless at this task as it is at decompiling FSBs.

However, fellows have apparently found an ingenious solution here:

http://social.biowar...s-455398-1.html

What they're doing is compiling the wavs from the toolset by working with the conversation itself.

Unfortunately, my toolset is for some mysterious reason screwed, and does not display any of the core resources, so I can't open a voice set conversation and put this idea to use.

Anyone wanna try?

UPDATE: Fixed my toolset, opened the voiceset dlg, transmuted the extracted WAVs to proper format with Goldfish, now stuck at this stage: http://social.biowar...437645/2#478051 Voiceover generation fails.

Modifié par Malcroix, 15 mars 2010 - 01:39 .

[Malcroix]

DONE AND DONE.

First ever custom voiceset to be found here:

http://social.biowar...m/project/2300/

As I suspected, it is much easier to work with an existing voice set than to make a new one from scratch.

As a result, we now have a working custom version of the Male Human Violent voice set, with the infamous Ladder comment replaced by a much cooler line from the same voice set.

Creating fully custom voice sets with your own sounds is now possible.

I will post a detailed description of the process at the project, and in this thread.

Modifié par Malcroix, 15 mars 2010 - 09:24 .

[Malcroix]

The simplest way to make a custom voice set is to replace an existing voice set with your own. Since during any single game you only use one voice set, there's hardly any downside; furthermore, it's very easy to install, uninstall, and exchange with the community.
 
Here are the basic steps:
 
1.      Locate the voice set you want replaced (presumably the one used by your PC) in the toolset under Conversations, check it out and open it. Now you can see all topics (e.g. COMBAT ATTACK, COMBAT TAUNT, SPELL FAILED etc.) as well as various sounds for each topic. There are 141 total possible sounds, but only 51 topics.  

2.      Organise your custom sound files. You’ll need 141 sounds for a full set; however, since most are just variations on one of the 51 topics, you can use duplicates (e.g. repeating some sounds for the COMBAT ATTACK list).

3.      Make sure your sounds are in the correct format: 16-bit 24-kHz mono PCM *.WAV. Use a sound program to enforce the right format. I like Goldwave because it’s free, fast, easy to use and allows batch processing (i.e. many files at once).

4.      Locate the voice set FSB file. It should be in the packages\\\\core\\\\audio\\\\vo\\\\en-us\\\\vo directory of your Dragon Age install folder (not the My Documents folder). It will be called like ss_pc_mhuman_violent.fsb.

5.      Use Aezay's fine program to extract the wave files from the FSB.

6.      Now RENAME your custom sound files so that they “replace” their original counterparts, in accordance with the nomenclature you figured out in steps 1 and 2.

7.      Put your RENAMED custom sound files into the /override/toolsetexport directory OF THE MODULE YOU’RE WORKING WITH (and, to be safe, also in the general /override/toolsetexport). These directories would be located in your My Documents/BioWare/Dragon Age folder.

8.      Still have the voice set conversation open in your toolset as per Step One? Good; now select Root in the conversation, then select Tools -> Generate VO -> Generate VO Local.

9.      At this point you’ll likely have some error messages pop up in your Log Window. Examine them carefully. If you see something like this:

Failed to copy template 'C:/Program Files/Dragon Age/Tools/ResourceBuild/Processors\\\\VO_template_project.fdp' to 'C:\\\\DOCUME~1\\\\[Your Account Name]\\\\LOCALS~1\\\\Temp\\\\DragonAgePlaceholderVO\\\\VO_template_project.fdp'

then go to your Dragon Age/Tools directory in your install folder, and copy the entire ResourceBuild folder into the directory which corresponds to the one the Toolset is trying to copy files from. In my case, I had to create Dragon Age/Tools in my C:/Program Files, and put ResourceBuild there. This is most likely a bug, with the Toolset not recognizing any other path than the default DA:O install path.

If you see other error messages, e.g. "could not speak text", you will want to check out this thread:

http://social.biowar.../index/592387/1

10.    Once you have ResourceBuild in this fake DA:O directory, the VoiceOver generation should proceed smoothly. Watch the Log Window for messages about the OUTPUT DIRECTORY. In my case it was C:\\\\Documents and Settings\\\\[Account Name]\\\\Local Settings\\\\Temp\\\\DragonAgePlaceholderVO. Open this folder. You should have a lot of *.wav and *.txt files there, along with a single *.FDP file. This file is what you’re looking for.

11.   Open the FDP file in FMOD Designer (located in your Dragon Age/Tools directory). Don’t change any settings, just select Build -> Build Project. This will compile all the *.wavs and *.txts and create an *.FSB and a *.FEV file which will be called like your voice set.

12.   Put these two files into your Override directory and voila! Your character will start using the custom voice in game!

Now share your custom voice with the community! It's very easy to install - just put the custom files in your Override, and rename them to match the voice set you want replaced!

Thanks to
Lady Olivia ,  sherlockemery ,  KingEtzel , Avaraen , Nitro9a  and all others who posted valuable insights which helped me get this done.

Modifié par Malcroix, 15 mars 2010 - 10:37 .

[Malcroix]

Easier way to do steps 4 - 6 above:



Look through the Conversation in the Toolset and note the ID numbers for every comment (bottom left corner of the Dialogue tab). Rename your custom sounds accordingly:



[6-digit ID number]_m.wav



This way you don't even have to bother with extracting the original FSB.

[KingEtzel]

Now we've gotten a big step further I think. Thankyou! But still, I want to add new soundsets instead of replacing existing ones. Maybe I can fiddle something with this DAextractor and finally find the problem, that prevents me from implementing new soundsets (using extending 2DAs). Once I am back from Africa. That's next week. Let's see what happens then. Still it should be possible, I have not given up.

[Malcroix]

Well, the aim of having a custom soundset is achieved. You can now make your character sound like whatever you want, and it's not even hard to do.

Having the soundset as a separate selection would complicate the installation process (as you'd have to replace some core files), and would make it hard for different soundset mods to coexist without manually altering these files. It would also require either starting a new game, or editing the saved game, which is often impossible at later stages of the game due to the enormous size of the unpacked saves.

Of course, replacing original voicesets is not a perfect solution, and separate selection definitely helps make the sound set feel more "proper", inserting it seamlessly into the character generation process.As a feature, separate selection is certainly valuable.

Have you been using existing DLG/FEV/FSB files in your attempts to create a new selectable soundset, or did you work with custom DLG/FEV/FSBs? For "proof of concept" it would be enough to make the selection work with an existing voice set - i.e. same DLG/FEV/FSB but different name. Once that is working, making custom files is a technicality since we can already do that (cf. my guide).

Modifié par Malcroix, 17 mars 2010 - 01:21 .

[KingEtzel]

The problem is, that I don't get the original DLG files for the voicesets. They're encrypted somewhere, or at least very well hidden.



You know, it looks like this at the moment:



1) playersoundsets.xls/gda POINTS TO a certain DLG-File

2) The DLG File points to the sounds with the ID numbers in the FSB.



The problem is in step 1. DA knows how the DLG file is named, but where does it look for it? In which folder do I have to put it (I already tried many overrides and the like). Perhaps it has to be converted to something or stored within another file or archive. I don't know, since I'm not able to track the corresponding soundset - DLG files of the core resources. Perhaps you have an idea?

[Malcroix]

FIRST FULLY CUSTOMISED VOICE SET CREATED.

Based on the character "The Transcendent One" from Planescape: Torment, it is available on Dragon Age Nexus and in my project.

I believe the title of this thread should indicate that the PC Soundset status is now [Solved].

We can continue working on separate selection as a further improvement.

Other improvements to be sought:

Currently in DA:O the player character does not respond when his/her portrait is clicked, or when given orders to move/use ability, when outside of combat. Tweaking the ss_types.2DA does not help - it is possible to regulate the frequency of replies given by party members on click, but the player character stays silent in these cases. This reduces the immersion and decreases the benefits of having custom voice sets (or voice sets per se, for that matter). If anyone has any ideas about possible resolution, please share.

I had a thought that maybe this is connected to the Approval parameter: all *on selection* comments are grouped by approval towards the PC (neutral, love, hate etc). Party members have full complements of those; PC soundsets, however, have only the "neutral" one. The game apparently chooses which batch of replies to use based on the character's attitude towards the PC, and the PC him/herself probably does not HAVE an Approval rating at all, not even a "neutral" one, so the game doesn't use the existing "neutral" batch of on-click replies.

However, the on-select replies IN COMBAT for some reason WORK. Maybe, since all combat replies are classed as "neutral", the game doesn't check for approval, and thus allows PC replies?

Sarevok Anchev wrote...

I was just curious to know, if it is possible to "import" the Infinity-Engine games Soundfiles for your PC.
Is that possible?

My Sarevok just sounds gay without some proper "FEEL MY UNHOLY WRATH!!!" ^^

It is certainly possible and it works, as proven by this mod.

Now you can unleash proper unholy wrath on your enemies 8=)

First extract the sounds from the game using NearInfinity and ACM2WAV converter, then follow the guide above to make your own voice set for DA:O.

KingEtzel wrote...

The problem is, that I don't get the original DLG files for the voicesets. They're encrypted somewhere, or at least very well hidden.

You know, it looks like this at the moment:

1) playersoundsets.xls/gda POINTS TO a certain DLG-File
2) The DLG File points to the sounds with the ID numbers in the FSB.

The problem is in step 1. DA knows how the DLG file is named, but where does it look for it? In which folder do I have to put it (I already tried many overrides and the like). Perhaps it has to be converted to something or stored within another file or archive. I don't know, since I'm not able to track the corresponding soundset - DLG files of the core resources. Perhaps you have an idea?

I don't understand your problem. The original DLGs can be easily located through the toolset (they are under Conversation), and extracted as GDAs. In fact, I've just done it.

Perhaps you could try using an existing soundset as a basis, and make it appear (and work) under a custom name in the character generation screen? If this works, then the issue is basically resolved.

Modifié par Malcroix, 20 mars 2010 - 09:52 .

[KingEtzel]

I have put much time into this already and of course I know where the DLGs are inside of the toolset. My point is, that the game has to recognize the additional DLG/the new soundset. And just by putting an exported DLG into an override directory of your choice (I think I tried all out that there are), this does not work.



Perhaps I try something on Tuesday, then I have access to my computer again...

[Torilund]

@Malcroix and everyone else:

Thanks for all this info. I have been thinking about doing a voice set since I my first character creation when I listened to the voices and was all, "Meh...". So psyched for this now!!