Brtdku

Reading the wav_to_spectrogram example from TensorFlow, it is said that To load your own audio, you need to supply a .wav file in LIN16 format. I have a wav file with the following information

$ mplayer -identify wrecking_crew.wav 

MPlayer 1.3.0 (Debian), built with gcc-7 (C) 2000-2016 MPlayer Team

do_connect: could not connect to socket

connect: No such file or directory

Failed to open LIRC support. You will not be able to use your remote control.



Playing wrecking_crew.wav.

libavformat version 57.83.100 (external)

ID_AUDIO_ID=0

Audio only file format detected.

Load subtitles in ./

ID_FILENAME=wrecking_crew.wav

ID_DEMUXER=audio

ID_AUDIO_FORMAT=1

ID_AUDIO_BITRATE=352800

ID_AUDIO_RATE=0

ID_AUDIO_NCH=1

ID_START_TIME=0.00

ID_LENGTH=48694.00

ID_SEEKABLE=1

ID_CHAPTERS=0

==========================================================================

Opening audio decoder: [pcm] Uncompressed PCM audio decoder

AUDIO: 22050 Hz, 1 ch, s16le, 352.8 kbit/100.00% (ratio: 44100->44100)

ID_AUDIO_BITRATE=352800

ID_AUDIO_RATE=22050

ID_AUDIO_NCH=1

Selected audio codec: [pcm] afm: pcm (Uncompressed PCM)

=========================================================================

AO: [pulse] 22050Hz 1ch s16le (2 bytes per sample)

ID_AUDIO_CODEC=pcm

Video: no video

Starting playback...

A:   8.4 (08.3) of 48694.0 (13:31:34.0)  0.0% 

Audio output truncated at end.

A:   8.4 (08.4) of 48694.0 (13:31:34.0)  0.0% 





Exiting... (End of file)

ID_EXIT=EOF

However, TensorFlow aborts the execution with the following error

$ bazel-bin/tensorflow/examples/wav_to_spectrogram/wav_to_spectrogram --input_wav=../wrecking_crew.wav

2019-01-19 21:38:51.292782: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA

2019-01-19 21:38:51.294078: I tensorflow/stream_executor/platform/default/dso_loader.cc:154] successfully opened CUDA library libcuda.so.1 locally

2019-01-19 21:38:51.361657: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1003] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero

2019-01-19 21:38:51.362134: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1434] Found device 0 with properties: 

name: Quadro M2000 major: 5 minor: 2 memoryClockRate(GHz): 1.1625

pciBusID: 0000:26:00.0

totalMemory: 3.95GiB freeMemory: 3.78GiB

2019-01-19 21:38:51.362153: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1513] Adding visible gpu devices: 0

2019-01-19 21:38:51.362752: I tensorflow/core/common_runtime/gpu/gpu_device.cc:985] Device interconnect StreamExecutor with strength 1 edge matrix:

2019-01-19 21:38:51.362762: I tensorflow/core/common_runtime/gpu/gpu_device.cc:991]      0 

2019-01-19 21:38:51.362767: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1004] 0:   N 

2019-01-19 21:38:51.362900: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1116] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 3566 MB memory) -> physical GPU (device: 0, name: Quadro M2000, pci bus id: 0000:26:00.0, compute capability: 5.2)

2019-01-19 21:38:51.411588: W tensorflow/core/framework/op_kernel.cc:1412] OP_REQUIRES failed at decode_wav_op.cc:55 : Invalid argument: Data too short when trying to read string

2019-01-19 21:38:51.412591: W tensorflow/core/framework/op_kernel.cc:1412] OP_REQUIRES failed at decode_wav_op.cc:55 : Invalid argument: Data too short when trying to read string

2019-01-19 21:38:51.412970: E tensorflow/examples/wav_to_spectrogram/main.cc:61] WavToSpectrogram failed with Invalid argument: Data too short when trying to read string

     [[{{node wav_decoder}}]]

     [[spectrogram/_1]]

Is that related to LIN16 format? What should I do?