Hey there Dimitri,

Let me know if you are interested in this PR.

I've bundled the modplug library into libnyquist.

Modplug is a PD licensed solution to load multichannel music (mod, s3m, xm, it...) It also can play MIDI files and a wide range of WAVs. MIDI files (and .abc files) require the sound patches located in pat/ folder: these samples are either PD or GPL licensed.

Most if not all of the work came from @JariKomppa 's Soloud repo, so kudos to him :)

Hey there,

Just giving a try on the library. Nice work so far!

It seems to crash here with test_data\ad_hoc\44_16_mono.mpc. Stereo mpc is fine though. Does it work on your machine :tm: ? :)

ADPCM is extremely helpful when streaming hundred of megabytes from disk. mmap()-ing wav files from disk, then decoding adpcm on the fly is one of the cheapest solutions, specially for mobile devices.

The source (MIT license) for a IMA/ADPCM decoder can be found here: https://github.com/afterwise/aw-ima/

When converting audio files on Windows, no problems are observed, the quality is good. But for some reason, when working with UNIX system (Android), the sound quality becomes much worse. Here is an example of what happens when converting the "TestBeat_Float32.wv" file to the .ogg format on UNIX.

TestBeat_Float32.zip

When linking against libnyquist:

• include its public headers in the PUBLIC interface.
• include its libraries dependencies in the PRIVATE interface.

Along the way, update the linkage of the examples, it was made PUBLIC unnecessarily, it is an executable, it won't be linked against another target.

fixes #50 tested with normal linux+make tested with mingw+make from linux seem to work fine in both cases I didn't test MSVC, because I don't have it, though it would be beneficial to test it maybe

There's no need to justify for this feature. Also I think in 2k18 the technology has already come to a point where MP3 is available for ordinary mortals. Whatever people say against it.

So there is libmpg123, and as I see it's LGPLv2, so it can be used commercially and requires dynamic linking. But libnyquist is built as a staticlib. Btw, why exactly libnyquist (and LabSound) only support staticlib build?

May be we could arrange it the other way around. Currently, your AudioDecoder is very rigid. If it allowed some external tampering, I could simply add my own decoder in the middle of runtime.

Taking the "external" approach even further, I could send the decoded MP3 data in some intermediate format... For example with ".mp3dec" extension, and there will be the decoded data and metadata in a specified format. But this is the least desired way.

Hello!

I use LabSound as a backend for Node.js bindings. LabSound uses libnyquist to decode audio. For JS I export WebAudioApi-compliant set of classes.

Among other things it is very important to do in-memory decoding. Because all JS examples and docs provide separate means of loading and decoding for audio files. Namely XHR for webpages, and I use fs.readFile on Node.js. I'm currently trying to reproduce a simple case with file loading. And I found out that LabSound and specifically libnyquist is incapable of doing ogg in memory.

void VorbisDecoder::LoadFromBuffer(AudioData * data, const std::vector<uint8_t> & memory)
{
    throw LoadBufferNotImplEx();
}

But there is, so to speak, a wae. I'll try to fit it into the VorbisDecoder. If I succeed, I'll do a PR.

Any thoughts on this? libnyquist doesn't seem to be developed too actively these days, are you still going to continue maintaining it?

I was pointing someone at libnyquist, but they objected due to needing to link WMF or CoreAudio. I think that comes from the AudioDevice dependency on RtAudio?

I think AudioDevice should be reserved for the examples and tests, but not be in the library itself, as it isn't used anywhere else in the codecs?

Also, for reference, this is the batch file I am using to compile from the command-line:

cl /Zi /Oy- /MDd /DDEBUG /EHsc ^
        -I include ^
        -I include\libnyquist ^
        -I third_party ^
        -I third_party\flac\src\include ^
        -I third_party\libogg\include ^
        -I third_party\libvorbis\include ^
        -I third_party\musepack\include ^
        -I third_party\opus\celt ^
        -I third_party\opus\silk ^
        -I third_party\opus\silk\float ^
        -I third_party\opus\libopus\include ^
        -I third_party\opus\opusfile\include ^
        -I third_party\opus\opusfile\src\include ^
        -I third_party\wavpack\include ^
        src\*.c* ^
        third_party\rtaudio\RtAudio.cpp /D__WINDOWS_DS__ ole32.lib user32.lib ^
        examples\src\main.cpp

Hey there,

Nice lib you have in here :) Please consider musepack decoding as it is very cheap to decode for mobile devices. Also, license is bsd-3.

Original source download at: https://www.musepack.net/index.php?pg=src (see sv8 download, libmpcdec/ subfolder)

Totally optional, I hacked a mpc decoder if you need some guidance (which I doubt :) https://github.com/r-lyeh/wave/blob/master/wave.cpp#L17

Hi, I was curious about it, isn't this library only meant to help load and read audio files ? in this case why are they needed since from what I know these libraries are meant to handle interfacing with the OS on linux to output sound or get input from microphones etc

atm i could compile my program without specifying one of them but i still get the error on cmake FATAL,On Linux, one of LIBNYQUIST_JACK, LIBNYQUIST_PULSE, or LIBNYQUIST_ASOUND must be set.

This is a draft to add a CMake config file to allow the library to be found using find_package in config mode.

A necessary part of this is adding a version number, discussed in #61.

I used the version 0.1.0 but this can be replaced with whatever version you want to use.

Hi,

I am running some experiments for AFLAPI and it has found a segmentation fault in FlacDecoderInternal::s_writeCallback when nqr::NyquistIO::Load a bad .flac file. This bug may allows attackers to cause DoS, so I report it here.

Environment: Ubuntu 20.04 + g++ 9.6.0

Test target: https://github.com/ddiakopoulos/libnyquist/blob/master/examples/src/Main.cpp

Poc: segv1.zip

To reproduce:

1. Complie the hole project with ASAN
2. Complie the example with ASAN:

[email protected]:~/test/libnyquist/build/CMakeFiles/libnyquist-examples.dir/examples/src$ g++ -fsanitize=address -o example Main.cpp.o AudioDevice.cpp.o -llibnyquist -lrtaudio

3. Run:

$ ./example ./segv1.flac

ASAN says:

[email protected]:~/test/libnyquist/build/CMakeFiles/libnyquist-examples.dir/examples/src$ ./example segv1.flac 
[rtaudio] Found: 3 device(s)
	Device: 0 - hw:Ensoniq AudioPCI,0
	Device: 1 - hw:Ensoniq AudioPCI,1
	Device: 2 - default

AddressSanitizer:DEADLYSIGNAL
=================================================================
==1646422==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000001 (pc 0x7efc7260ec30 bp 0x7ffd0d1b4a80 sp 0x7ffd0d1b41e8 T0)
==1646422==The signal is caused by a READ memory access.
==1646422==Hint: address points to the zero page.
    #0 0x7efc7260ec2f  (/lib/x86_64-linux-gnu/libc.so.6+0xbbc2f)
    #1 0x7efc72b4a37e in __interceptor_memcpy ../../../../src/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:790
    #2 0x563605538c2d in FlacDecoderInternal::s_writeCallback(FLAC__StreamDecoder const*, FLAC__Frame const*, int const* const*, void*) /home/ubuntu/test/libnyquist/src/FlacDecoder.cpp:164
    #3 0x5636055ac8ae in write_audio_frame_to_client_ /home/ubuntu/test/libnyquist/third_party/FLAC/src/stream_decoder.c:2972
    #4 0x56360559e1ac in read_frame_ /home/ubuntu/test/libnyquist/third_party/FLAC/src/stream_decoder.c:2146
    #5 0x56360558cf7b in FLAC__stream_decoder_process_until_end_of_stream /home/ubuntu/test/libnyquist/third_party/FLAC/src/stream_decoder.c:1101
    #6 0x5636055373b3 in FlacDecoderInternal::FlacDecoderInternal(nqr::AudioData*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) /home/ubuntu/test/libnyquist/src/FlacDecoder.cpp:66
    #7 0x5636055361f0 in nqr::FlacDecoder::LoadFromPath(nqr::AudioData*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) /home/ubuntu/test/libnyquist/src/FlacDecoder.cpp:247
    #8 0x5636054e7e4d in nqr::NyquistIO::Load(nqr::AudioData*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) /home/ubuntu/test/libnyquist/src/Common.cpp:47
    #9 0x5636054cfaed in main /home/ubuntu/test/libnyquist/examples/src/Main.cpp:34
    #10 0x7efc72577082 in __libc_start_main ../csu/libc-start.c:308
    #11 0x5636054cf6dd in _start (/home/ubuntu/test/libnyquist/build/CMakeFiles/libnyquist-examples.dir/examples/src/example+0x826dd)

AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV (/lib/x86_64-linux-gnu/libc.so.6+0xbbc2f) 
==1646422==ABORTING

Impact: An attacker can exploit this vulnerability by submitting a malicious elf file that exploits this bug which will result in a Denial of Service attack.

Hi,

I am running some experiments for AFLAPI and it has found a segmentation fault in nqr::linear_resample. This bug may allows attackers to cause DoS, so I report it here.

Environment: Ubuntu 20.04 + g++ 9.6.0

Test target: https://github.com/ddiakopoulos/libnyquist/blob/master/examples/src/Main.cpp

Poc: segv.zip

To reproduce:

1. Complie the hole project with ASAN
2. Complie the example with ASAN:

[email protected]:~/test/libnyquist/build/CMakeFiles/libnyquist-examples.dir/examples/src$ g++ -fsanitize=address -o example Main.cpp.o AudioDevice.cpp.o -llibnyquist -lrtaudio

3. Run:

$ ./example ./segv.flac

ASAN says:

[email protected]:~/test/libnyquist/build/CMakeFiles/libnyquist-examples.dir/examples/src$ ./example ./segv.flac 
[rtaudio] Found: 3 device(s)
	Device: 0 - hw:Ensoniq AudioPCI,0
	Device: 1 - hw:Ensoniq AudioPCI,1
	Device: 2 - default

[Warning - Sample Rate Mismatch] - file is sampled at 0 and output is 44100
Input Samples: 0
Playing STEREO for: -nan seconds...
Output Samples: 0
AddressSanitizer:DEADLYSIGNAL
=================================================================
==1091132==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x556a7874ef66 bp 0x7ffecea08010 sp 0x7ffecea07f40 T0)
==1091132==The signal is caused by a READ memory access.
==1091132==Hint: address points to the zero page.
    #0 0x556a7874ef65 in nqr::linear_resample(double, std::vector<float, std::allocator<float> > const&, std::vector<float, std::allocator<float> >&, unsigned int) /home/ubuntu/test/libnyquist/include/libnyquist/Common.h:229
    #1 0x556a7874d358 in main /home/ubuntu/test/libnyquist/examples/src/Main.cpp:143
    #2 0x7f49d80eb082 in __libc_start_main ../csu/libc-start.c:308
    #3 0x556a7874c6dd in _start (/home/ubuntu/test/libnyquist/build/CMakeFiles/libnyquist-examples.dir/examples/src/example+0x826dd)

AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /home/ubuntu/test/libnyquist/include/libnyquist/Common.h:229 in nqr::linear_resample(double, std::vector<float, std::allocator<float> > const&, std::vector<float, std::allocator<float> >&, unsigned int)
==1091132==ABORTING

Issue code snippet:

// libnyquist/include/libnyquist/Common.h:229
220  inline void linear_resample(const double rate, const std::vector<float> & input, std::vector<float> & output, const uint32_t samplesToProcess)
221  {   // <---- It seems that check samplesToProcess == 0 can solve this bug?
222      double virtualReadIndex = 0;
223      double a, b, i, sample;
224      uint32_t n = samplesToProcess - 1;  // samplesToProcess =0 --> n = -1 
225      while (n--)  // --->overflow here!
226      {
227          uint32_t readIndex = static_cast<uint32_t>(virtualReadIndex);
228          i = virtualReadIndex - readIndex;
229          a = input[readIndex + 0];
230          b = input[readIndex + 1];
231          sample = (1.0 - i) * a + i * b; // linear interpolate
232          output.push_back(static_cast<float>(sample));
233          virtualReadIndex += rate;
234      }
235  }

Hope that helps.

Hi,

first: Thanks for providing such a useful and easy to use lib. This has been some awesome help for my own projects.

It would be great to have some versioning of libnyquist to have some stable API, ABI and flags to build against. Are there any plans for versioning libnyquist? As e.g. one of my builds was broken with #60, it would be great to have some stable tags/version to build against.

Hi.

I'm using libnyquist. And found that while decoding wavs of type IMA ADPCM of two channels the length of the decoded file is reduced to half. If I add

        if (wavHeader.channel_count == 2)
          totalSamples *= 2;

On WavDecoder.cpp line 309 the problem seems to be fixed.

But I'm not sure if it is the right solution.

Thanks.