"FSBank Error: The format of the source file is invalid."

FMOD Engine
1

Hi, I have been trying to automate a process for building FSBs using the FSBANK api but I am running into an error when trying to build some banks “The format of the source file is invalid.”:

std::vector<const char*> fileNames; // absolute paths to .wavs

FSBANK_SUBSOUND subSound = {};
subSound.fileNames = fileNames.data();
subSound.numFiles = static_cast<unsigned int>(fileNames.size());

const FSBANK_SUBSOUND* subSounds = &subSound;

fs::path fsbPath = dir / (stem + ".fsb");

std::cout << "Building FSB: " << fsbPath << "\n";
ERRCHECK(FSBank_Build(subSounds, 1, FSBANK_FORMAT_PCM, FSBANK_BUILD_DEFAULT, 0, nullptr, fsbPath.string().c_str()));

Pasted image 20250627092806
Pasted image 20250627092806977×197 20.4 KB

It’s only some of my banks that get this error too.

But what gets me is that using fsbank.exe seems to happily build the bank with the same .wavs!

Pasted image 20250627093254
Pasted image 20250627093254882×485 60.9 KB

.

I also get the error “An invalid parameter has been passed to this function.” on FSBank_Build for some other banks – while these .wav files also build successfully in fsbank.exe.

What am I missing?

2

Hi! If possible, could I get you to upload the problem wavs to your FMOD user profile for me to take a look at on my end?

3

Hi! Apologies for the late response - I don’t think I am able to provide these files publicly, but I can show details of the file.

This is one of the problem wav files:

image
image1297×113 5.66 KB

This has actually been converted from a Playstation 2 .vag file using FFMPEG.
image

I don’t have a great deal of experience with audio, so I may have missed something fairly obvious here - I appreciate the help.

4

Something I’ve actually noticed is that multiple files that I am trying to build into the same .FSB do have different sample rates.
For example,
File 1 - 22042 Hz
File 2 - 11015 Hz
File 3 - 24000 Hz

5

I’ve identified the error here and have been able to fix it.

I misunderstood the documentation of FSBANK_SUBSOUND – Specifically, “fileNames - List of file names used to produce an interleaved sound.”.

I assumed this was where I was to provide the paths to all of my waves that I wanted to build into the .fsb, but the correct way to do it is to have an FSBANK_SUBSOUND for each sub sound (should be obvious, right? My mistake…)

My updated, correct code:

std::vector<FSBANK_SUBSOUND> subSounds;
std::vector<std::vector<const char *>>
    subSoundFileNames; // Need to keep strings in scope...
for (const auto &fileName : wavFileNames)
{
    subSoundFileNames.push_back({fileName.c_str()});

    FSBANK_SUBSOUND subSound = {0};
    subSound.fileNames = subSoundFileNames.back().data();
    subSound.numFiles = 1;

    subSounds.push_back(subSound);
}

ERRCHECK(FSBank_Build(subSounds.data(), static_cast<unsigned int>(subSounds.size()),
                      FSBANK_FORMAT_PCM, FSBANK_BUILD_DEFAULT, 0, nullptr, outputPath);

This actually builds a set of wavs into a single .fsb, which was my intention.

1 Like
6

Happy to hear that you resolved the issue!

The general use of a subsound is indeed to create a single subsound per audio file. This is what the FSBank executable does. Providing multiple files to a single subsound is done when you want to the audio data from each file to be interleaved in the encoded output - i.e. providing 6 mono files to be interleaved into a single 6ch subsound.

As you’ve noted, the FSBank API docs aren’t the clearest on this point, so I’ve flagged it for improvement internally.