Changeset 8924bdf


Ignore:
Timestamp:
Nov 21, 2018, 9:21:49 PM (13 months ago)
Author:
Paul Brossier <piem@piem.org>
Branches:
feature/autosink, feature/constantq, feature/pitchshift, feature/pydocstrings, feature/timestretch, master
Children:
08db885
Parents:
313eb2b
Message:

[doc] describe single-/double-precision modes

File:
1 edited

Legend:

Unmodified
Added
Removed
  • doc/requirements.rst

    r313eb2b r8924bdf  
    303303................
    304304
     305The datatype used to store real numbers in aubio is named `smpl_t`. By default,
     306`smpl_t` is defined as `float`, a `single-precision format
     307<https://en.wikipedia.org/wiki/Single-precision_floating-point_format>`_
     308(32-bit).  Some algorithms require a floating point representation with a
     309higher precision, for instance to prevent arithmetic underflow in recursive
     310filters.  In aubio, these special samples are named `lsmp_t` and defined as
     311`double` by default (64-bit).
     312
     313Sometimes it may be useful to compile aubio in `double-precision`, for instance
     314to reproduce numerical results obtained with 64-bit routines. In this case,
     315`smpl_t` will be defined as `double`.
     316
     317The following table shows how `smpl_t` and `lsmp_t` are defined in single- and
     318double-precision modes:
     319
     320.. list-table:: Single and double-precision modes
     321   :align: center
     322
     323   * -
     324     - single
     325     - double
     326   * - `smpl_t`
     327     - ``float``
     328     - ``double``
     329   * - `lsmp_t`
     330     - ``double``
     331     - ``long double``
     332
    305333To compile aubio in double precision mode, configure with ``--enable-double``.
    306334
    307 To compile aubio in single precision mode, use ``--disable-double`` (default,
    308 recommended).
     335To compile in single-precision mode (default), use ``--disable-double`` (or
     336simply none of these two options).
    309337
    310338Disabling the tests
Note: See TracChangeset for help on using the changeset viewer.