formats.
"""
+from __future__ import division
+
import array
import os
import sys
def generate_1d_LUT_image(ramp_1d_path,
resolution=1024,
- min_value=0.0,
- max_value=1.0):
+ min_value=0,
+ max_value=1):
"""
Object description.
Return value description.
"""
- # print('Generate 1d LUT image - %s' % ramp1dPath)
-
- # open image
- format = os.path.splitext(ramp_1d_path)[1]
ramp = oiio.ImageOutput.create(ramp_1d_path)
- # set image specs
spec = oiio.ImageSpec()
spec.set_format(oiio.FLOAT)
# spec.format.basetype = oiio.FLOAT
ramp.close()
-def write_SPI_1d(filename, from_min, from_max, data, entries, channels):
+def write_SPI_1d(filename,
+ from_min,
+ from_max,
+ data,
+ entries,
+ channels,
+ components=3):
"""
Object description.
Return value description.
"""
+ # May want to use fewer components than there are channels in the data
+ # Most commonly used for single channel LUTs
+ components = min(3, components, channels)
+
with open(filename, 'w') as fp:
fp.write('Version 1\n')
fp.write('From %f %f\n' % (from_min, from_max))
fp.write('Length %d\n' % entries)
- fp.write('Components %d\n' % (min(3, channels)))
+ fp.write('Components %d\n' % components)
fp.write('{\n')
for i in range(0, entries):
entry = ''
- for j in range(0, min(3, channels)):
+ for j in range(0, components):
entry = '%s %s' % (entry, data[i * channels + j])
fp.write(' %s\n' % entry)
fp.write('}\n')
+def write_CSP_1d(filename,
+ from_min,
+ from_max,
+ data,
+ entries,
+ channels,
+ components=3):
+ """
+ Object description.
+
+ Parameters
+ ----------
+ parameter : type
+ Parameter description.
+
+ Returns
+ -------
+ type
+ Return value description.
+ """
+
+ # May want to use fewer components than there are channels in the data
+ # Most commonly used for single channel LUTs
+ components = min(3, components, channels)
+
+ with open(filename, 'w') as fp:
+ fp.write('CSPLUTV100\n')
+ fp.write('1D\n')
+ fp.write('\n')
+ fp.write('BEGIN METADATA\n')
+ fp.write('END METADATA\n')
+
+ fp.write('\n')
+
+ fp.write('2\n')
+ fp.write('%f %f\n' % (from_min, from_max))
+ fp.write('0.0 1.0\n')
+ fp.write('2\n')
+ fp.write('%f %f\n' % (from_min, from_max))
+ fp.write('0.0 1.0\n')
+ fp.write('2\n')
+ fp.write('%f %f\n' % (from_min, from_max))
+ fp.write('0.0 1.0\n')
+
+ fp.write('\n')
+
+ fp.write('%d\n' % entries)
+ if components == 1:
+ for i in range(0, entries):
+ entry = ''
+ for j in range(3):
+ entry = '%s %s' % (entry, data[i * channels])
+ fp.write('%s\n' % entry)
+ else:
+ for i in range(entries):
+ entry = ''
+ for j in range(components):
+ entry = '%s %s' % (entry, data[i * channels + j])
+ fp.write('%s\n' % entry)
+ fp.write('\n')
+
+
+def write_CTL_1d(filename,
+ from_min,
+ from_max,
+ data,
+ entries,
+ channels,
+ components=3):
+ """
+ Object description.
+
+ Parameters
+ ----------
+ parameter : type
+ Parameter description.
+
+ Returns
+ -------
+ type
+ Return value description.
+ """
+
+ # May want to use fewer components than there are channels in the data
+ # Most commonly used for single channel LUTs
+ components = min(3, components, channels)
+
+ with open(filename, 'w') as fp:
+ fp.write('// %d x %d LUT generated by "generate_lut"\n' % (
+ entries, components))
+ fp.write('\n')
+ fp.write('const float min1d = %3.9f;\n' % from_min)
+ fp.write('const float max1d = %3.9f;\n' % from_max)
+ fp.write('\n')
+
+ # Write LUT
+ if components == 1:
+ fp.write('const float lut[] = {\n')
+ for i in range(0, entries):
+ fp.write('%s' % data[i * channels])
+ if i != (entries - 1):
+ fp.write(',')
+ fp.write('\n')
+ fp.write('};\n')
+ fp.write('\n')
+ else:
+ for j in range(components):
+ fp.write('const float lut%d[] = {\n' % j)
+ for i in range(0, entries):
+ fp.write('%s' % data[i * channels])
+ if i != (entries - 1):
+ fp.write(',')
+ fp.write('\n')
+ fp.write('};\n')
+ fp.write('\n')
+
+ fp.write('void main\n')
+ fp.write('(\n')
+ fp.write(' input varying float rIn,\n')
+ fp.write(' input varying float gIn,\n')
+ fp.write(' input varying float bIn,\n')
+ fp.write(' input varying float aIn,\n')
+ fp.write(' output varying float rOut,\n')
+ fp.write(' output varying float gOut,\n')
+ fp.write(' output varying float bOut,\n')
+ fp.write(' output varying float aOut\n')
+ fp.write(')\n')
+ fp.write('{\n')
+ fp.write(' float r = rIn;\n')
+ fp.write(' float g = gIn;\n')
+ fp.write(' float b = bIn;\n')
+ fp.write('\n')
+ fp.write(' // Apply LUT\n')
+ if components == 1:
+ fp.write(' r = lookup1D(lut, min1d, max1d, r);\n')
+ fp.write(' g = lookup1D(lut, min1d, max1d, g);\n')
+ fp.write(' b = lookup1D(lut, min1d, max1d, b);\n')
+ elif components == 3:
+ fp.write(' r = lookup1D(lut0, min1d, max1d, r);\n')
+ fp.write(' g = lookup1D(lut1, min1d, max1d, g);\n')
+ fp.write(' b = lookup1D(lut2, min1d, max1d, b);\n')
+ fp.write('\n')
+ fp.write(' rOut = r;\n')
+ fp.write(' gOut = g;\n')
+ fp.write(' bOut = b;\n')
+ fp.write(' aOut = aIn;\n')
+ fp.write('}\n')
+
+
+def write_1d(filename,
+ from_min,
+ from_max,
+ data,
+ data_entries,
+ data_channels,
+ lut_components=3,
+ format='spi1d'):
+ """
+ Object description.
+
+ Parameters
+ ----------
+ parameter : type
+ Parameter description.
+
+ Returns
+ -------
+ type
+ Return value description.
+ """
+
+ ocioFormatsToExtensions = {'cinespace': 'csp',
+ 'flame': '3dl',
+ 'icc': 'icc',
+ 'houdini': 'lut',
+ 'lustre': '3dl',
+ 'ctl': 'ctl'}
+
+ if format in ocioFormatsToExtensions:
+ if ocioFormatsToExtensions[format] == 'csp':
+ write_CSP_1d(filename,
+ from_min,
+ from_max,
+ data,
+ data_entries,
+ data_channels,
+ lut_components)
+ elif ocioFormatsToExtensions[format] == 'ctl':
+ write_CTL_1d(filename,
+ from_min,
+ from_max,
+ data,
+ data_entries,
+ data_channels,
+ lut_components)
+ else:
+ write_SPI_1d(filename,
+ from_min,
+ from_max,
+ data,
+ data_entries,
+ data_channels,
+ lut_components)
+
+
def generate_1d_LUT_from_image(ramp_1d_path,
output_path=None,
- min_value=0.0,
- max_value=1.0):
+ min_value=0,
+ max_value=1,
+ channels=3,
+ format='spi1d'):
"""
Object description.
if output_path is None:
output_path = '%s.%s' % (ramp_1d_path, 'spi1d')
- # open image
ramp = oiio.ImageInput.open(ramp_1d_path)
- # get image specs
- spec = ramp.spec()
- type = spec.format.basetype
- width = spec.width
- height = spec.height
- channels = spec.nchannels
+ ramp_spec = ramp.spec()
+ ramp_width = ramp_spec.width
+ ramp_channels = ramp_spec.nchannels
- # get data
- # Force data to be read as float. The Python API doesn't handle
- # half-floats well yet.
+ # Forcibly read data as float, the Python API doesn't handle half-float
+ # well yet.
type = oiio.FLOAT
- data = ramp.read_image(type)
+ ramp_data = ramp.read_image(type)
- write_SPI_1d(output_path, min_value, max_value, data, width, channels)
+ write_1d(output_path, min_value, max_value,
+ ramp_data, ramp_width, ramp_channels, channels, format)
def generate_3d_LUT_image(ramp_3d_path, resolution=32):
lut_extract.execute()
-def generate_3d_LUT_from_image(ramp_3d_path, output_path=None, resolution=32):
+def generate_3d_LUT_from_image(ramp_3d_path,
+ output_path=None,
+ resolution=32,
+ format='spi3d'):
"""
Object description.
"""
if output_path is None:
- output_path = '%s.%s' % (ramp_3d_path, 'spi1d')
+ output_path = '%s.%s' % (ramp_3d_path, 'spi3d')
+
+ ocioFormatsToExtensions = {'cinespace': 'csp',
+ 'flame': '3dl',
+ 'icc': 'icc',
+ 'houdini': 'lut',
+ 'lustre': '3dl'}
+
+ if format == 'spi3d' or not (format in ocioFormatsToExtensions):
+ # Extract a spi3d LUT
+ args = ['--extract',
+ '--cubesize',
+ str(resolution),
+ '--maxwidth',
+ str(resolution * resolution),
+ '--input',
+ ramp_3d_path,
+ '--output',
+ output_path]
+ lut_extract = Process(description='extract a 3d LUT',
+ cmd='ociolutimage',
+ args=args)
+ lut_extract.execute()
- args = ['--extract',
- '--cubesize',
- str(resolution),
- '--maxwidth',
- str(resolution * resolution),
- '--input',
- ramp_3d_path,
- '--output',
- output_path]
- lut_extract = Process(description='extract a 3d LUT',
- cmd='ociolutimage',
- args=args)
- lut_extract.execute()
+ else:
+ output_path_spi3d = '%s.%s' % (output_path, 'spi3d')
+
+ # Extract a spi3d LUT
+ args = ['--extract',
+ '--cubesize',
+ str(resolution),
+ '--maxwidth',
+ str(resolution * resolution),
+ '--input',
+ ramp_3d_path,
+ '--output',
+ output_path_spi3d]
+ lut_extract = Process(description='extract a 3d LUT',
+ cmd='ociolutimage',
+ args=args)
+ lut_extract.execute()
+
+ # Convert to a different format
+ args = ['--lut',
+ output_path_spi3d,
+ '--format',
+ format,
+ output_path]
+ lut_convert = Process(description='convert a 3d LUT',
+ cmd='ociobakelut',
+ args=args)
+ lut_convert.execute()
def apply_CTL_to_image(input_image,
output_image,
- ctl_paths=[],
- input_scale=1.0,
- output_scale=1.0,
- global_params={},
- aces_CTL_directory=None):
+ ctl_paths=None,
+ input_scale=1,
+ output_scale=1,
+ global_params=None,
+ aces_ctl_directory=None):
"""
Object description.
Return value description.
"""
+ if ctl_paths is None:
+ ctl_paths = []
+ if global_params is None:
+ global_params = {}
+
if len(ctl_paths) > 0:
ctlenv = os.environ
- if aces_CTL_directory != None:
- if os.path.split(aces_CTL_directory)[1] != 'utilities':
- ctl_module_path = os.path.join(aces_CTL_directory, 'utilities')
+ if aces_ctl_directory is not None:
+ if os.path.split(aces_ctl_directory)[1] != 'utilities':
+ ctl_module_path = os.path.join(aces_ctl_directory, 'utilities')
else:
- ctl_module_path = aces_CTL_directory
+ ctl_module_path = aces_ctl_directory
ctlenv['CTL_MODULE_PATH'] = ctl_module_path
args = []
for ctl in ctl_paths:
args += ['-ctl', ctl]
args += ['-force']
- # args += ['-verbose']
args += ['-input_scale', str(input_scale)]
args += ['-output_scale', str(output_scale)]
args += ['-global_param1', 'aIn', '1.0']
args += [input_image]
args += [output_image]
- # print('args : %s' % args)
-
ctlp = Process(description='a ctlrender process',
cmd='ctlrender',
args=args, env=ctlenv)
ctl_paths,
lut_resolution=1024,
identity_LUT_bit_depth='half',
- input_scale=1.0,
- output_scale=1.0,
- global_params={},
+ input_scale=1,
+ output_scale=1,
+ global_params=None,
cleanup=True,
- aces_CTL_directory=None,
- min_value=0.0,
- max_value=1.0):
+ aces_ctl_directory=None,
+ min_value=0,
+ max_value=1,
+ channels=3,
+ format='spi1d'):
"""
Object description.
Return value description.
"""
- # print(lutPath)
- # print(ctlPaths)
+ if global_params is None:
+ global_params = {}
lut_path_base = os.path.splitext(lut_path)[0]
min_value,
max_value)
- if identity_LUT_bit_depth != 'half':
+ if identity_LUT_bit_depth not in ['half', 'float']:
identity_LUT_image = '%s.%s.%s' % (lut_path_base, 'uint16', 'tiff')
convert_bit_depth(identity_LUT_image_float,
identity_LUT_image,
input_scale,
output_scale,
global_params,
- aces_CTL_directory)
+ aces_ctl_directory)
generate_1d_LUT_from_image(transformed_LUT_image,
lut_path,
min_value,
- max_value)
+ max_value,
+ channels,
+ format)
if cleanup:
os.remove(identity_LUT_image)
Return value description.
"""
- # open image
transformed = oiio.ImageInput.open(transformed_LUT_image)
- # get image specs
transformed_spec = transformed.spec()
- type = transformed_spec.format.basetype
width = transformed_spec.width
height = transformed_spec.height
channels = transformed_spec.nchannels
- # rotate or not
if width != lut_resolution * lut_resolution or height != lut_resolution:
print(('Correcting image as resolution is off. '
'Found %d x %d. Expected %d x %d') % (
lut_resolution))
print('Generating %s' % corrected_LUT_image)
- #
- # We're going to generate a new correct image
- #
-
- # Get the source data
- # Force data to be read as float. The Python API doesn't handle
- # half-floats well yet.
+ # Forcibly read data as float, the Python API doesn't handle half-float
+ # well yet.
type = oiio.FLOAT
source_data = transformed.read_image(type)
- format = os.path.splitext(corrected_LUT_image)[1]
correct = oiio.ImageOutput.create(corrected_LUT_image)
- # set image specs
correct_spec = oiio.ImageSpec()
correct_spec.set_format(oiio.FLOAT)
correct_spec.width = height
for j in range(0, correct_spec.height):
for i in range(0, correct_spec.width):
for c in range(0, correct_spec.nchannels):
- # print(i, j, c)
dest_data[(correct_spec.nchannels *
correct_spec.width * j +
correct_spec.nchannels * i + c)] = (
ctl_paths,
lut_resolution=64,
identity_LUT_bit_depth='half',
- input_scale=1.0,
- output_scale=1.0,
- global_params={},
+ input_scale=1,
+ output_scale=1,
+ global_params=None,
cleanup=True,
- aces_CTL_directory=None):
+ aces_ctl_directory=None,
+ format='spi3d'):
"""
Object description.
Return value description.
"""
- # print(lutPath)
- # print(ctlPaths)
+ if global_params is None:
+ global_params = {}
lut_path_base = os.path.splitext(lut_path)[0]
identity_LUT_image_float = '%s.%s.%s' % (lut_path_base, 'float', 'tiff')
generate_3d_LUT_image(identity_LUT_image_float, lut_resolution)
- if identity_LUT_bit_depth != 'half':
+ if identity_LUT_bit_depth not in ['half', 'float']:
identity_LUT_image = '%s.%s.%s' % (lut_path_base,
identity_LUT_bit_depth,
'tiff')
input_scale,
output_scale,
global_params,
- aces_CTL_directory)
+ aces_ctl_directory)
corrected_LUT_image = '%s.%s.%s' % (lut_path_base, 'correct', 'exr')
corrected_LUT_image = correct_LUT_image(transformed_LUT_image,
corrected_LUT_image,
lut_resolution)
- generate_3d_LUT_from_image(corrected_LUT_image, lut_path, lut_resolution)
+ generate_3d_LUT_from_image(corrected_LUT_image,
+ lut_path,
+ lut_resolution,
+ format)
if cleanup:
os.remove(identity_LUT_image)
os.remove(transformed_LUT_image)
if corrected_LUT_image != transformed_LUT_image:
os.remove(corrected_LUT_image)
- # os.remove(correctedLUTImage)
+ if format != 'spi3d':
+ lut_path_spi3d = '%s.%s' % (lut_path, 'spi3d')
+ os.remove(lut_path_spi3d)
def main():
usage='%prog [options]')
p.add_option('--lut', '-l', type='string', default='')
+ p.add_option('--format', '-f', type='string', default='')
p.add_option('--ctl', '-c', type='string', action='append')
p.add_option('--lutResolution1d', '', type='int', default=1024)
p.add_option('--lutResolution3d', '', type='int', default=33)
p.add_option('--ctlReleasePath', '-r', type='string', default='')
p.add_option('--bitDepth', '-b', type='string', default='float')
p.add_option('--keepTempImages', '', action='store_true')
- p.add_option('--minValue', '', type='float', default=0.0)
- p.add_option('--maxValue', '', type='float', default=1.0)
- p.add_option('--inputScale', '', type='float', default=1.0)
- p.add_option('--outputScale', '', type='float', default=1.0)
+ p.add_option('--minValue', '', type='float', default=0)
+ p.add_option('--maxValue', '', type='float', default=1)
+ p.add_option('--inputScale', '', type='float', default=1)
+ p.add_option('--outputScale', '', type='float', default=1)
p.add_option('--ctlRenderParam', '-p', type='string', nargs=2,
action='append')
options, arguments = p.parse_args()
- #
- # Get options
- #
lut = options.lut
+ format = options.format
ctls = options.ctl
- lut_resolution_1d = options.lut_resolution_1d
- lut_resolution_3d = options.lut_resolution_3d
+ lut_resolution_1d = options.lutResolution1d
+ lut_resolution_3d = options.lutResolution3d
min_value = options.minValue
max_value = options.maxValue
input_scale = options.inputScale
cleanup = not options.keepTempImages
params = {}
- if options.ctlRenderParam != None:
+ if options.ctlRenderParam is not None:
for param in options.ctlRenderParam:
params[param[0]] = float(param[1])
args_start = len(sys.argv) + 1
args = []
- # print('command line : \n%s\n' % ' '.join(sys.argv))
-
- #
- # Generate LUTs
- #
if generate_1d:
print('1D LUT generation options')
else:
print('3D LUT generation options')
print('lut : %s' % lut)
+ print('format : %s' % format)
print('ctls : %s' % ctls)
print('lut res 1d : %s' % lut_resolution_1d)
print('lut res 3d : %s' % lut_resolution_3d)
cleanup,
ctl_release_path,
min_value,
- max_value)
+ max_value,
+ format=format)
elif generate_3d:
generate_3d_LUT_from_CTL(lut,
output_scale,
params,
cleanup,
- ctl_release_path)
+ ctl_release_path,
+ format=format)
else:
print(('\n\nNo LUT generated. '
'You must choose either 1D or 3D LUT generation\n\n'))
if __name__ == '__main__':
main()
-
projects / OpenColorIO-Configs.git / blobdiff