LAME binary for Mac OS X

This page provides PowerPC specific optimized version of LAME (using AltiVec, etc).
Note for intel Macs: The official version has already some intel-specific optimizations using MMX/SSE. So please build by yourself or get precompiled binary from RareWares (as for version 3.98, there is a precompiled binary here).

What is it?

LAME is a free, open-sourced mp3 encoder. It is said that LAME offers the best quality of all mp3 encoders.

This page provides LAME binaries compiled with G3/G4/G5 specific optimization flag. In addition, I modified the original code to optimize for PowerPC G4/G5 processor. This version is faster than the original version by utilizing AltiVec(aka Velocity Engine, VMX) in those processors. These binaries are built on OSX 10.5.5 with gcc 4.2.1. Archive contains encoder frontend only, not library.

These binaries sometimes output a bit different encoding result compared to the reference one. This is due to difference of internal precision of FP computation between AltiVec VFPU and generic FPU. I think this rarely affects encoding quality. But for a person who worries about this, I recommend to try to invert the output waveform and mix with the reference one with the waveform editor like Audacity.

Changes

• Vectorization
  • Vectorized FFT routine
  • Vectorized MDCT routine
  • Vectorized psychoacoustic analysis routine
  • Vectorized quantization routine
  • Vectorized noise calculation routine
  • Vectorized huffman bit count routine
  • Vectorized PCM input routine
  • Vectorized replaygain calculation routine
• Algorithm change
  • x^3/4 calculation with estimation and Newton-Rhapson refinement
  • Log calculation with approximate polynomial (more accurate than the original table-lookup with lenear interpolation algorithm)
  • Quantization factor calculation with approximate polynomial (more accurate than the original table-lookup algorithm)

Performance

An optimization for VBR(old,new), ABR and CBR routine is almost done. This version is about 85-105%/80-95% faster than the reference one on G4(744x/745x)/G5 when using 3.98b8.

fig.: encoder speedups (lame 3.98b8)

LAME 3.98.2 (updated on 2008/9/23)

• gcc4.2 compiled binary for G5/G4(744x/745x)/G4(7400/7410)/G3/Intel
Compiled with gcc 4.2.1. Includes a universal binary for all machines. The optimum one is chosen automatically.
• source diffs
• Now the diff is compatible with non-OSX (like Linux) environments.
• Define ALTIVEC and ALTIVEC_970 macro to build an AltiVec-enabled binary for G5.
• Define ALTIVEC macro to build an AltiVec-enabled binary for G4.
• Define PPC_FRSQRTE macro to build a binary for G3.

LAME 3.97 (updated on 2008/5/6)

Patch is included in the archive.

• for G5
• for G4(744x/745x)
• for G4(7400/7410)
• for G3

This page is link free.
If you have any comment with this page, please contact me.