Try different approaches to lower CPU pressure for AudioBuffer & WaveformImage rendering

[fengalin]

Here are the identified approaches:

• Use i16 to store the sample values (actually the channel values). The aim is to use a more compact buffer in order to improve cache locality. It also pushes the actual conversion to another format (namely f64 for WaveformImage) to the caller.
• Only store up to 5+1 channels. This is to limit the buffer length. Rendering more than 6 channels would cause more CPU stress for a visually confusing result.
• Gather channels of a sample together and store them as a SmallVec in the AudioBuffer. The idea is to get all channels from a sample Index at once from the the ring buffer.
• Draw each channel from the iter completely instead of drawing each channel from a particular sample and then moving to the next one. This would save color changing and moving to previous position for each channel at each iteration.

Thanks to the recent addition of benches, we can now verify the gains from these approaches.

[fengalin]

Before any changes:

test bench_append_samples ... bench:       4,660 ns/iter (+/- 30)
test bench_render_buffers ... bench:  20,699,294 ns/iter (+/- 656,395)

[fengalin]

Storage as i16:

test bench_append_samples ... bench:       4,669 ns/iter (+/- 20)
test bench_render_buffers ... bench:  18,788,012 ns/iter (+/- 1,398,147)

[fengalin]

Storing samples a SmallVecs of channels:

test bench_append_samples ... bench:       4,346 ns/iter (+/- 42)
test bench_render_buffers ... bench:  18,975,297 ns/iter (+/- 1,482,715)

[fengalin]

Drawing each channel completely:

test bench_append_samples ... bench:       4,674 ns/iter (+/- 84)
test bench_render_buffers ... bench:  12,478,807 ns/iter (+/- 423,996)