Improve VLC samplerate conversion by asking float samples

Asking floats from VLC instead of signed 16-bit samples allows
VLC to use the libsamplerate resampler instead of the ugly_resampler.

Otherwise it might convert to s16l first, and refuse to use
libsamplerate which is unable to work with that data format.

1 files changed, 2 insertions, 2 deletions

diff --git a/src/VLCInput.h b/src/VLCInput.h
index 5ef970b..16c6cbf 100644
--- a/src/VLCInput.h
+++ b/src/VLCInput.h
@@ -163,12 +163,12 @@ class VLCInput
         ssize_t m_read(uint8_t* buf, size_t length);
 
         /*! Buffer used in the callback functions for VLC */
-        std::vector<uint8_t> m_current_buf;
+        std::vector<float> m_current_buf;
 
         std::mutex m_queue_mutex;
 
         /*! Buffer containing all available samples from VLC */
-        std::deque<uint8_t> m_queue;
+        std::deque<float> m_queue;
 
         std::string m_uri;
         unsigned m_verbosity;