Skip to content

utils

Classes:

  • DitherType

    Enum for zimg_dither_type_e and fmtc dmode.

Functions:

  • depth

    A convenience bitdepth conversion function using only internal plugins if possible.

  • expect_bits

    Expected output bitdepth for a clip.

  • flatten_vnodes

    Flatten an array of VideoNodes.

  • frame2clip

    Convert a VideoFrame to a VideoNode.

  • get_b

    Extract the blue plane of the given clip.

  • get_g

    Extract the green plane of the given clip.

  • get_r

    Extract the red plane of the given clip.

  • get_u

    Extract the first chroma (U) plane of the given clip.

  • get_v

    Extract the second chroma (V) plane of the given clip.

  • get_y

    Extract the luma (Y) plane of the given clip.

  • insert_clip

    Replace frames of a longer clip with those of a shorter one.

  • join

    General interface to combine clips or planes into a single clip.

  • limiter
  • plane

    Extract a plane from the given clip.

  • split

    Split a clip into a list of individual planes.

  • stack_clips

    Recursively stack clips in alternating directions: horizontal → vertical → horizontal → ...

Attributes:

  • EXPR_VARS

    Variables to access clips in Expr.

EXPR_VARS module-attribute

EXPR_VARS = [
    "x",
    "y",
    "z",
    "a",
    "b",
    "c",
    "d",
    "e",
    "f",
    "g",
    "h",
    "i",
    "j",
    "k",
    "l",
    "m",
    "n",
    "o",
    "p",
    "q",
    "r",
    "s",
    "t",
    "u",
    "v",
    "w",
]

Variables to access clips in Expr.

DitherType

Bases: CustomStrEnum

Enum for zimg_dither_type_e and fmtc dmode.

Methods:

  • apply

    Apply the given DitherType to a clip.

  • is_fmtc

    Whether the DitherType is applied through fmtc.

  • should_dither

    Automatically determines whether dithering is needed for a given depth/range/sample type conversion.

Attributes:

  • ATKINSON

    Another error diffusion kernel.

  • AUTO

    Choose automatically.

  • ERROR_DIFFUSION

    Floyd-Steinberg error diffusion.

  • ERROR_DIFFUSION_FMTC

    Floyd-Steinberg error diffusion.

  • NONE

    Round to nearest.

  • ORDERED

    Bayer patterned dither.

  • OSTROMOUKHOV

    Another error diffusion kernel.

  • QUASIRANDOM

    Dither using quasirandom sequences.

  • RANDOM

    Pseudo-random noise of magnitude 0.5.

  • SIERRA_2_4A

    Another type of error diffusion.

  • STUCKI

    Another error diffusion kernel.

  • VOID

    A way to generate blue-noise dither and has a much better visual aspect than ordered dithering.

ATKINSON class-attribute instance-attribute

ATKINSON = ('atkinson', 5)

Another error diffusion kernel. Generates distinct patterns but keeps clean the flat areas (noise modulation).

AUTO class-attribute instance-attribute

AUTO = 'auto'

Choose automatically.

ERROR_DIFFUSION class-attribute instance-attribute

ERROR_DIFFUSION = 'error_diffusion'

Floyd-Steinberg error diffusion.

ERROR_DIFFUSION_FMTC class-attribute instance-attribute

ERROR_DIFFUSION_FMTC = ('error_diffusion_fmtc', 6)

Floyd-Steinberg error diffusion. Modified for serpentine scan (avoids worm artifacts).

NONE class-attribute instance-attribute

NONE = 'none'

Round to nearest.

ORDERED class-attribute instance-attribute

ORDERED = 'ordered'

Bayer patterned dither.

OSTROMOUKHOV class-attribute instance-attribute

OSTROMOUKHOV = ('ostromoukhov', 7)

Another error diffusion kernel. Slow, available only for integer input at the moment. Avoids usual F-S artifacts.

QUASIRANDOM class-attribute instance-attribute

QUASIRANDOM = ('quasirandom', 9)

Dither using quasirandom sequences. Good intermediary between void, cluster, and error diffusion algorithms.

RANDOM class-attribute instance-attribute

RANDOM = 'random'

Pseudo-random noise of magnitude 0.5.

SIERRA_2_4A class-attribute instance-attribute

SIERRA_2_4A = ('sierra_2_4a', 3)

Another type of error diffusion. Quick and excellent quality, similar to Floyd-Steinberg.

STUCKI class-attribute instance-attribute

STUCKI = ('stucki', 4)

Another error diffusion kernel. Preserves delicate edges better but distorts gradients.

VOID class-attribute instance-attribute

VOID = ('void', 8)

A way to generate blue-noise dither and has a much better visual aspect than ordered dithering.

apply

apply(
    clip: VideoNode,
    out_fmt: VideoFormat,
    range_in: ColorRange,
    range_out: ColorRange,
) -> VideoNode

Apply the given DitherType to a clip.

Source code in vstools/functions/utils.py
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
def apply(
    self, clip: vs.VideoNode, out_fmt: vs.VideoFormat, range_in: ColorRange, range_out: ColorRange
) -> vs.VideoNode:
    """
    Apply the given DitherType to a clip.
    """
    if self is DitherType.AUTO:
        self = DitherType.VOID if DitherType.should_dither(clip, out_fmt, range_in, range_out) else DitherType.NONE

    fmt = get_video_format(clip)
    clip = ColorRange.ensure_presence(clip, range_in)

    if not self.is_fmtc():
        return clip.resize.Point(
            format=out_fmt.id,
            dither_type=self.value.lower(),
            range_in=range_in.value_zimg,
            range=range_out.value_zimg,
        )

    # Workaround because fmtc doesn't support FLOAT 16 input
    if fmt.sample_type is vs.FLOAT and fmt.bits_per_sample == 16:
        clip = DitherType.NONE.apply(clip, fmt.replace(bits_per_sample=32), range_in, range_out)

    return clip.fmtc.bitdepth(
        dmode=self._fmtc_dmode,
        bits=out_fmt.bits_per_sample,
        fulls=range_in is ColorRange.FULL,
        fulld=range_out is ColorRange.FULL,
    )

is_fmtc

is_fmtc() -> bool

Whether the DitherType is applied through fmtc.

Source code in vstools/functions/utils.py
184
185
186
187
188
def is_fmtc(self) -> bool:
    """
    Whether the DitherType is applied through fmtc.
    """
    return self._fmtc_dmode is not None

should_dither staticmethod

should_dither(
    in_fmt: VideoFormatLike | HoldsVideoFormat,
    out_fmt: VideoFormatLike | HoldsVideoFormat,
    /,
    in_range: ColorRangeLike | None = None,
    out_range: ColorRangeLike | None = None,
) -> bool
should_dither(
    in_bits: int,
    out_bits: int,
    /,
    in_range: ColorRangeLike | None = None,
    out_range: ColorRangeLike | None = None,
    in_sample_type: SampleType | None = None,
    out_sample_type: SampleType | None = None,
) -> bool
should_dither(
    in_bits_or_fmt: int | VideoFormatLike | HoldsVideoFormat,
    out_bits_or_fmt: int | VideoFormatLike | HoldsVideoFormat,
    /,
    in_range: ColorRangeLike | None = None,
    out_range: ColorRangeLike | None = None,
    in_sample_type: SampleType | None = None,
    out_sample_type: SampleType | None = None,
) -> bool

Automatically determines whether dithering is needed for a given depth/range/sample type conversion.

If an input range is specified, an output range should be specified, otherwise it assumes a range conversion.

For an explanation of when dithering is needed:

  • Dithering is NEVER needed if the conversion results in a float sample type.
  • Dithering is ALWAYS needed for a range conversion (i.e. full to limited or vice-versa).
  • Dithering is ALWAYS needed to convert a float sample type to an integer sample type.
  • Dithering is needed when upsampling full range content except when one depth is a multiple of the other, when the upsampling is a simple integer multiplication, e.g. for 8 -> 16: (0-255) * 257 -> (0-65535).
  • Dithering is needed when downsampling limited or full range.

Dithering is theoretically needed when converting from an integer depth greater than 10 to half float, despite the higher bit depth, but zimg's internal resampler currently does not dither for float output.

Parameters:

  • in_bits_or_fmt

    (int | VideoFormatLike | HoldsVideoFormat) –

    Input bitdepth, clip, frame or video format.

  • out_bits_or_fmt

    (int | VideoFormatLike | HoldsVideoFormat) –

    Output bitdepth, clip, frame or video format.

  • in_range

    (ColorRangeLike | None, default: None ) –

    Input color range.

  • out_range

    (ColorRangeLike | None, default: None ) –

    Output color range.

  • in_sample_type

    (SampleType | None, default: None ) –

    Input sample type.

  • out_sample_type

    (SampleType | None, default: None ) –

    Output sample type.

Returns:

  • bool

    Whether the clip should be dithered.

Source code in vstools/functions/utils.py
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
@staticmethod
def should_dither(
    in_bits_or_fmt: int | VideoFormatLike | HoldsVideoFormat,
    out_bits_or_fmt: int | VideoFormatLike | HoldsVideoFormat,
    /,
    in_range: ColorRangeLike | None = None,
    out_range: ColorRangeLike | None = None,
    in_sample_type: vs.SampleType | None = None,
    out_sample_type: vs.SampleType | None = None,
) -> bool:
    """
    Automatically determines whether dithering is needed for a given depth/range/sample type conversion.

    If an input range is specified, an output range *should* be specified, otherwise it assumes a range conversion.

    For an explanation of when dithering is needed:

    - Dithering is NEVER needed if the conversion results in a float sample type.
    - Dithering is ALWAYS needed for a range conversion (i.e. full to limited or vice-versa).
    - Dithering is ALWAYS needed to convert a float sample type to an integer sample type.
    - Dithering is needed when upsampling full range content except when one depth is a multiple of the other,
        when the upsampling is a simple integer multiplication, e.g. for 8 -> 16: (0-255) * 257 -> (0-65535).
    - Dithering is needed when downsampling limited or full range.

    Dithering is theoretically needed when converting from an integer depth greater than 10 to half float,
    despite the higher bit depth, but zimg's internal resampler currently does not dither for float output.

    Args:
        in_bits_or_fmt: Input bitdepth, clip, frame or video format.
        out_bits_or_fmt: Output bitdepth, clip, frame or video format.
        in_range: Input color range.
        out_range: Output color range.
        in_sample_type: Input sample type.
        out_sample_type: Output sample type.

    Returns:
        Whether the clip should be dithered.
    """
    in_fmt = get_video_format(in_bits_or_fmt, sample_type=in_sample_type)
    out_fmt = get_video_format(out_bits_or_fmt, sample_type=out_sample_type)

    in_range = ColorRange.from_param_with_fallback(in_range)
    out_range = ColorRange.from_param_with_fallback(out_range)

    if out_fmt.sample_type is vs.FLOAT:
        return False

    if in_fmt.sample_type is vs.FLOAT:
        return True

    if in_range != out_range:
        return True

    in_bits = in_fmt.bits_per_sample
    out_bits = out_fmt.bits_per_sample

    if in_bits == out_bits:
        return False

    if in_bits > out_bits:
        return True

    return in_range == ColorRange.FULL and bool(out_bits % in_bits)

depth

depth(
    clip: VideoNode,
    bitdepth: VideoFormatLike | HoldsVideoFormat | int | None = None,
    /,
    sample_type: int | SampleType | None = None,
    *,
    range_in: ColorRangeLike | None = None,
    range_out: ColorRangeLike | None = None,
    dither_type: str | DitherType = AUTO,
) -> VideoNode

A convenience bitdepth conversion function using only internal plugins if possible.

This uses exclusively internal plugins except for specific dither_types. To check whether your DitherType uses fmtc, use DitherType.is_fmtc.

Example
rc_8 = vs.core.std.BlankClip(format=vs.YUV420P8)
rc_10 = depth(src_8, 10)
print(rc_10.format.name)  # YUV420P10

rc2_10 = vs.core.std.BlankClip(format=vs.RGB30)
rc2_8 = depth(src2_10, 8, dither_type=Dither.RANDOM)  # override default dither behavior
print(rc2_8.format.name)  # RGB24

Parameters:

  • clip

    (VideoNode) –

    Input clip.

  • bitdepth

    (VideoFormatLike | HoldsVideoFormat | int | None, default: None ) –

    Desired bitdepth of the output clip.

  • sample_type

    (int | SampleType | None, default: None ) –

    Desired sample type of output clip. Allows overriding default float/integer behavior.

  • range_in

    (ColorRangeLike | None, default: None ) –

    Input pixel range (defaults to input clip's range).

  • range_out

    (ColorRangeLike | None, default: None ) –

    Output pixel range (defaults to input clip's range).

  • dither_type

    (str | DitherType, default: AUTO ) –

    Dithering algorithm. Allows overriding default dithering behavior. See DitherType.

    When integer output is desired but the conversion may produce fractional values, defaults to VOID.

    In other cases, defaults to no dither.

    See DitherType.should_dither for more information.

Returns:

  • VideoNode

    Converted clip with desired bit depth and sample type. ColorFamily will be same as input.

Source code in vstools/functions/utils.py
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
def depth(
    clip: vs.VideoNode,
    bitdepth: VideoFormatLike | HoldsVideoFormat | int | None = None,
    /,
    sample_type: int | vs.SampleType | None = None,
    *,
    range_in: ColorRangeLike | None = None,
    range_out: ColorRangeLike | None = None,
    dither_type: str | DitherType = DitherType.AUTO,
) -> vs.VideoNode:
    """
    A convenience bitdepth conversion function using only internal plugins if possible.

    This uses exclusively internal plugins except for specific dither_types.
    To check whether your DitherType uses fmtc, use [DitherType.is_fmtc][vstools.DitherType.is_fmtc].

    Example:
        ```py
        rc_8 = vs.core.std.BlankClip(format=vs.YUV420P8)
        rc_10 = depth(src_8, 10)
        print(rc_10.format.name)  # YUV420P10

        rc2_10 = vs.core.std.BlankClip(format=vs.RGB30)
        rc2_8 = depth(src2_10, 8, dither_type=Dither.RANDOM)  # override default dither behavior
        print(rc2_8.format.name)  # RGB24
        ```

    Args:
        clip: Input clip.
        bitdepth: Desired bitdepth of the output clip.
        sample_type: Desired sample type of output clip. Allows overriding default float/integer behavior.
        range_in: Input pixel range (defaults to input `clip`'s range).
        range_out: Output pixel range (defaults to input `clip`'s range).
        dither_type: Dithering algorithm. Allows overriding default dithering behavior.
            See [DitherType][vstools.DitherType].

            When integer output is desired but the conversion may produce fractional values,
            defaults to [VOID][vstools.DitherType.VOID].

            In other cases, defaults to no dither.

            See [DitherType.should_dither][vstools.DitherType.should_dither] for more information.

    Returns:
        Converted clip with desired bit depth and sample type. `ColorFamily` will be same as input.
    """
    in_fmt = get_video_format(clip)
    out_fmt = get_video_format(bitdepth or clip, sample_type=sample_type)

    range_out = ColorRange.from_param_or_video(range_out, clip)
    range_in = ColorRange.from_param_or_video(range_in, clip)

    if (in_fmt.bits_per_sample, in_fmt.sample_type, range_in) == (
        out_fmt.bits_per_sample,
        out_fmt.sample_type,
        range_out,
    ):
        return clip

    new_format = in_fmt.replace(bits_per_sample=out_fmt.bits_per_sample, sample_type=out_fmt.sample_type)

    return DitherType.from_param(dither_type, depth).apply(clip, new_format, range_in, range_out)

expect_bits

expect_bits(
    clip: VideoNode, /, expected_depth: int = 16, **kwargs: Any
) -> tuple[VideoNode, int]

Expected output bitdepth for a clip.

This function is meant to be used when a clip may not match the expected input bitdepth. Both the dithered clip and the original bitdepth are returned.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

  • expected_depth

    (int, default: 16 ) –

    Expected bitdepth. Default: 16.

Returns:

  • tuple[VideoNode, int]

    Tuple containing the clip dithered to the expected depth and the original bitdepth.

Source code in vstools/functions/utils.py
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
def expect_bits(clip: vs.VideoNode, /, expected_depth: int = 16, **kwargs: Any) -> tuple[vs.VideoNode, int]:
    """
    Expected output bitdepth for a clip.

    This function is meant to be used when a clip may not match the expected input bitdepth.
    Both the dithered clip and the original bitdepth are returned.

    Args:
        clip: Input clip.
        expected_depth: Expected bitdepth. Default: 16.

    Returns:
        Tuple containing the clip dithered to the expected depth and the original bitdepth.
    """
    bits = get_depth(clip)

    if bits != expected_depth:
        clip = depth(clip, expected_depth, **kwargs)

    return clip, bits

flatten_vnodes

flatten_vnodes(
    *clips: VideoNodeIterable, split_planes: bool = False
) -> Sequence[VideoNode]

Flatten an array of VideoNodes.

Parameters:

  • *clips

    (VideoNodeIterable, default: () ) –

    An array of clips to flatten into a list.

  • split_planes

    (bool, default: False ) –

    Optionally split the VideoNodes into their individual planes as well. Default: False.

Returns:

  • Sequence[VideoNode]

    Flattened list of VideoNodes.

Source code in vstools/functions/utils.py
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
def flatten_vnodes(*clips: VideoNodeIterable, split_planes: bool = False) -> Sequence[vs.VideoNode]:
    """
    Flatten an array of VideoNodes.

    Args:
        *clips: An array of clips to flatten into a list.
        split_planes: Optionally split the VideoNodes into their individual planes as well. Default: False.

    Returns:
        Flattened list of VideoNodes.
    """
    nodes = list[vs.VideoNode](flatten(clips))

    if not split_planes:
        return nodes

    return reduce(operator.iadd, map(split, nodes), [])

frame2clip

frame2clip(frame: VideoFrame) -> VideoNode

Convert a VideoFrame to a VideoNode.

Parameters:

  • frame

    (VideoFrame) –

    Input frame.

Returns:

  • VideoNode

    1-frame long VideoNode of the input frame.

Source code in vstools/functions/utils.py
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
def frame2clip(frame: vs.VideoFrame) -> vs.VideoNode:
    """
    Convert a VideoFrame to a VideoNode.

    Args:
        frame: Input frame.

    Returns:
        1-frame long VideoNode of the input frame.
    """

    key = hash((frame.width, frame.height, frame.format.id))

    if _f2c_cache.get(key, None) is None:
        _f2c_cache[key] = blank_clip = vs.core.std.BlankClip(
            None, frame.width, frame.height, frame.format.id, 1, 1, 1, [0] * frame.format.num_planes, True
        )
    else:
        blank_clip = _f2c_cache[key]

    frame_cp = frame.copy()

    return vs.core.std.ModifyFrame(blank_clip, blank_clip, lambda n, f: frame_cp)

get_b

get_b(clip: VideoNode) -> VideoNode

Extract the blue plane of the given clip.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

Returns:

  • VideoNode

    B plane of the input clip.

Raises:

Source code in vstools/functions/utils.py
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
def get_b(clip: vs.VideoNode, /) -> vs.VideoNode:
    """
    Extract the blue plane of the given clip.

    Args:
        clip: Input clip.

    Returns:
        B plane of the input clip.

    Raises:
        UnsupportedColorFamilyError: Clip is not RGB.
    """

    UnsupportedColorFamilyError.check(clip, vs.RGB, get_b)

    return plane(clip, 2)

get_g

get_g(clip: VideoNode) -> VideoNode

Extract the green plane of the given clip.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

Returns:

  • VideoNode

    G plane of the input clip.

Raises:

Source code in vstools/functions/utils.py
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
def get_g(clip: vs.VideoNode, /) -> vs.VideoNode:
    """
    Extract the green plane of the given clip.

    Args:
        clip: Input clip.

    Returns:
        G plane of the input clip.

    Raises:
        UnsupportedColorFamilyError: Clip is not RGB.
    """

    UnsupportedColorFamilyError.check(clip, vs.RGB, get_g)

    return plane(clip, 1)

get_r

get_r(clip: VideoNode) -> VideoNode

Extract the red plane of the given clip.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

Returns:

  • VideoNode

    R plane of the input clip.

Raises:

Source code in vstools/functions/utils.py
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
def get_r(clip: vs.VideoNode, /) -> vs.VideoNode:
    """
    Extract the red plane of the given clip.

    Args:
        clip: Input clip.

    Returns:
        R plane of the input clip.

    Raises:
        UnsupportedColorFamilyError: Clip is not RGB.
    """

    UnsupportedColorFamilyError.check(clip, vs.RGB, get_r)

    return plane(clip, 0)

get_u

get_u(clip: VideoNode) -> VideoNode

Extract the first chroma (U) plane of the given clip.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

Returns:

  • VideoNode

    Y plane of the input clip.

Raises:

Source code in vstools/functions/utils.py
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
def get_u(clip: vs.VideoNode, /) -> vs.VideoNode:
    """
    Extract the first chroma (U) plane of the given clip.

    Args:
        clip: Input clip.

    Returns:
        Y plane of the input clip.

    Raises:
        UnsupportedColorFamilyError: Clip is not YUV.
    """

    UnsupportedColorFamilyError.check(clip, vs.YUV, get_u)

    return plane(clip, 1)

get_v

get_v(clip: VideoNode) -> VideoNode

Extract the second chroma (V) plane of the given clip.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

Returns:

  • VideoNode

    V plane of the input clip.

Raises:

Source code in vstools/functions/utils.py
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
def get_v(clip: vs.VideoNode, /) -> vs.VideoNode:
    """
    Extract the second chroma (V) plane of the given clip.

    Args:
        clip: Input clip.

    Returns:
        V plane of the input clip.

    Raises:
        UnsupportedColorFamilyError: Clip is not YUV.
    """

    UnsupportedColorFamilyError.check(clip, vs.YUV, get_v)

    return plane(clip, 2)

get_y

get_y(clip: VideoNode) -> VideoNode

Extract the luma (Y) plane of the given clip.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

Returns:

  • VideoNode

    Y plane of the input clip.

Raises:

Source code in vstools/functions/utils.py
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
def get_y(clip: vs.VideoNode, /) -> vs.VideoNode:
    """
    Extract the luma (Y) plane of the given clip.

    Args:
        clip: Input clip.

    Returns:
        Y plane of the input clip.

    Raises:
        UnsupportedColorFamilyError: Clip is not GRAY or YUV.
    """

    UnsupportedColorFamilyError.check(clip, [vs.YUV, vs.GRAY], get_y)

    return plane(clip, 0)

insert_clip

insert_clip(
    clip: VideoNode, /, insert: VideoNode, start_frame: int, strict: bool = True
) -> VideoNode

Replace frames of a longer clip with those of a shorter one.

The insert clip may NOT exceed the final frame of the input clip. This limitation can be circumvented by setting strict=False.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

  • insert

    (VideoNode) –

    Clip to insert into the input clip.

  • start_frame

    (int) –

    Frame to start inserting from.

  • strict

    (bool, default: True ) –

    Throw an error if the inserted clip exceeds the final frame of the input clip. If False, truncate the inserted clip instead. Default: True.

Returns:

  • VideoNode

    Clip with frames replaced by the insert clip.

Raises:

  • CustomValueError

    Insert clip is too long, strict=False, and exceeds the final frame of the input clip.

Source code in vstools/functions/utils.py
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
def insert_clip(clip: vs.VideoNode, /, insert: vs.VideoNode, start_frame: int, strict: bool = True) -> vs.VideoNode:
    """
    Replace frames of a longer clip with those of a shorter one.

    The insert clip may NOT exceed the final frame of the input clip.
    This limitation can be circumvented by setting `strict=False`.

    Args:
        clip: Input clip.
        insert: Clip to insert into the input clip.
        start_frame: Frame to start inserting from.
        strict: Throw an error if the inserted clip exceeds the final frame of the input clip. If False, truncate the
            inserted clip instead. Default: True.

    Returns:
        Clip with frames replaced by the insert clip.

    Raises:
        CustomValueError: Insert clip is too long, ``strict=False``, and exceeds the final frame of the input clip.
    """

    if start_frame == 0:
        return vs.core.std.Splice([insert, clip[insert.num_frames :]])

    pre = clip[:start_frame]
    insert_diff = (start_frame + insert.num_frames) - clip.num_frames

    if insert_diff == 0:
        return vs.core.std.Splice([pre, insert])

    if insert_diff < 0:
        return vs.core.std.Splice([pre, insert, clip[insert_diff:]])

    if strict:
        raise ClipLengthError(
            "Inserted clip is too long and exceeds the final frame of the input clip.",
            insert_clip,
            {"clip": clip.num_frames, "diff": insert_diff},
        )

    return vs.core.std.Splice([pre, insert[:-insert_diff]])

join

join(
    luma: VideoNode,
    chroma: VideoNode,
    /,
    *,
    prop_src: VideoNode | SupportsIndex | None = ...,
) -> VideoNode
join(
    plane0: VideoNode,
    plane1: VideoNode,
    plane2: VideoNode,
    alpha: VideoNode | None = None,
    /,
    *,
    family: ColorFamily = YUV,
    prop_src: VideoNode | SupportsIndex | None = ...,
) -> VideoNode
join(
    planes: Sequence[VideoNode],
    family: ColorFamily = YUV,
    /,
    *,
    prop_src: VideoNode | SupportsIndex | None = ...,
) -> VideoNode
join(
    clips: Mapping[Planes, VideoNode | None],
    family: ColorFamily = YUV,
    /,
    *,
    prop_src: VideoNode | SupportsIndex | None = ...,
) -> VideoNode
join(
    clips: VideoNode | Sequence[VideoNode] | Mapping[Planes, VideoNode | None],
    plane1_or_family: VideoNode | ColorFamily | None = None,
    plane2: VideoNode | None = None,
    alpha: VideoNode | None = None,
    /,
    *,
    family: ColorFamily = YUV,
    prop_src: VideoNode | SupportsIndex | None = None,
) -> VideoNode

General interface to combine clips or planes into a single clip.

Parameters:

  • clips

    (VideoNode | Sequence[VideoNode] | Mapping[Planes, VideoNode | None]) –

    First plane, sequence of single-plane clips or mapping of planes to clips:

  • plane1_or_family

    (VideoNode | ColorFamily | None, default: None ) –

    Chroma clip, second plane or color family, depending on usage.

  • plane2

    (VideoNode | None, default: None ) –

    Third plane when combining three planes.

  • alpha

    (VideoNode | None, default: None ) –

    Optional alpha clip.

  • family

    (ColorFamily, default: YUV ) –

    Output color family. Defaults to YUV.

  • prop_src

    (VideoNode | SupportsIndex | None, default: None ) –

    Optional clip or index to copy frame properties from.

Raises:

  • CustomIndexError

    Invalid plane index.

  • CustomTypeError

    Invalid input type.

Returns:

  • VideoNode

    Clip with combined planes.

Source code in vstools/functions/utils.py
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
def join(
    clips: vs.VideoNode | Sequence[vs.VideoNode] | Mapping[Planes, vs.VideoNode | None],
    plane1_or_family: vs.VideoNode | vs.ColorFamily | None = None,
    plane2: vs.VideoNode | None = None,
    alpha: vs.VideoNode | None = None,
    /,
    *,
    family: vs.ColorFamily = vs.ColorFamily.YUV,
    prop_src: vs.VideoNode | SupportsIndex | None = None,
) -> vs.VideoNode:
    """
    General interface to combine clips or planes into a single clip.

    Args:
        clips: First plane, sequence of single-plane clips or mapping of planes to clips:
        plane1_or_family: Chroma clip, second plane or color family, depending on usage.
        plane2: Third plane when combining three planes.
        alpha: Optional alpha clip.
        family: Output color family. Defaults to YUV.
        prop_src: Optional clip or index to copy frame properties from.

    Raises:
        CustomIndexError: Invalid plane index.
        CustomTypeError: Invalid input type.

    Returns:
        Clip with combined planes.
    """
    if isinstance(clips, Mapping):
        clips_map = dict[int, vs.VideoNode]()

        for p_key, node in clips.items():
            if node is None:
                continue

            if p_key is None:
                clips_map.update(enumerate(flatten_vnodes(node, split_planes=True)))
            else:
                clips_map.update((i, plane(node, i)) for i in to_arr(p_key))

        return join([clips_map[i] for i in sorted(clips_map)], family, prop_src=prop_src)

    if not isinstance(clips, vs.VideoNode):
        if isinstance(plane1_or_family, vs.ColorFamily):
            family = plane1_or_family

        if (n_clips := len(clips)) == 1:
            return clips[0]
        if n_clips in (2, 3, 4):
            return join(*clips, family=family, prop_src=prop_src)
        else:
            raise CustomIndexError("Too many or not enough clips/planes passed!", join, n_clips)

    if not isinstance(plane1_or_family, vs.VideoNode):
        raise CustomTypeError(func=join)

    if not plane2:
        UnsupportedColorFamilyError.check(
            (family, plane1_or_family),
            vs.YUV,
            join,
            "When combining two clips, color family and chroma clip must be {correct}, not {wrong}.",
        )

        clips = [clips, plane1_or_family]
        planes = [0, 1, 2]

    else:
        clips = [clips, plane1_or_family, plane2]
        planes = [0, 0, 0]

    if not isinstance(prop_src, (vs.VideoNode, NoneType)):
        prop_src = clips[prop_src.__index__()]

    joined = vs.core.std.ShufflePlanes(clips, planes, family, prop_src)

    if alpha:
        joined = joined.std.ClipToProp(alpha, "_Alpha")

    return joined

limiter

limiter(
    clip: VideoNode,
    /,
    min_val: float | Sequence[float] | None = None,
    max_val: float | Sequence[float] | None = None,
    *,
    tv_range: bool = False,
    mask: bool = False,
    planes: Planes = None,
    func: FuncExcept | None = None,
) -> VideoNode
limiter(
    _func: Callable[P, VideoNode],
    /,
    min_val: float | Sequence[float] | None = None,
    max_val: float | Sequence[float] | None = None,
    *,
    tv_range: bool = False,
    mask: bool = False,
    planes: Planes = None,
    func: FuncExcept | None = None,
) -> Callable[P, VideoNode]
limiter(
    *,
    min_val: float | Sequence[float] | None = None,
    max_val: float | Sequence[float] | None = None,
    tv_range: bool = False,
    mask: bool = False,
    planes: Planes = None,
    func: FuncExcept | None = None
) -> Callable[[Callable[P, VideoNode]], Callable[P, VideoNode]]
limiter(
    clip_or_func: VideoNode | Callable[P, VideoNode] | None = None,
    /,
    min_val: float | Sequence[float] | None = None,
    max_val: float | Sequence[float] | None = None,
    *,
    tv_range: bool = False,
    mask: bool = False,
    planes: Planes = None,
    func: FuncExcept | None = None,
) -> Union[
    VideoNode,
    Callable[P, VideoNode],
    Callable[[Callable[P, VideoNode]], Callable[P, VideoNode]],
]

Wraps vszip.Limiter but only processes if clip format is not integer, a min/max val is specified or tv_range is True.

Parameters:

  • clip_or_func

    (VideoNode | Callable[P, VideoNode] | None, default: None ) –

    Clip to process or function that returns a VideoNode to be processed.

  • min_val

    (float | Sequence[float] | None, default: None ) –

    Lower bound. Defaults to the lowest allowed value for the input. Can be specified for each plane individually.

  • max_val

    (float | Sequence[float] | None, default: None ) –

    Upper bound. Defaults to the highest allowed value for the input. Can be specified for each plane individually.

  • tv_range

    (bool, default: False ) –

    Changes min/max defaults values to LIMITED.

  • mask

    (bool, default: False ) –

    Float chroma range from -0.5/0.5 to 0.0/1.0.

  • planes

    (Planes, default: None ) –

    Which planes to process.

  • func

    (FuncExcept | None, default: None ) –

    Function returned for custom error handling. This should only be set by VS package developers.

Returns:

Source code in vstools/functions/utils.py
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
def limiter[**P](
    clip_or_func: vs.VideoNode | Callable[P, vs.VideoNode] | None = None,
    /,
    min_val: float | Sequence[float] | None = None,
    max_val: float | Sequence[float] | None = None,
    *,
    tv_range: bool = False,
    mask: bool = False,
    planes: Planes = None,
    func: FuncExcept | None = None,
) -> Union[
    vs.VideoNode,
    Callable[P, vs.VideoNode],
    Callable[[Callable[P, vs.VideoNode]], Callable[P, vs.VideoNode]],
]:
    """
    Wraps [vszip.Limiter](https://github.com/dnjulek/vapoursynth-zip/wiki/Limiter)
    but only processes if clip format is not integer, a min/max val is specified or tv_range is True.

    Args:
        clip_or_func: Clip to process or function that returns a VideoNode to be processed.
        min_val: Lower bound. Defaults to the lowest allowed value for the input. Can be specified for each plane
            individually.
        max_val: Upper bound. Defaults to the highest allowed value for the input. Can be specified for each plane
            individually.
        tv_range: Changes min/max defaults values to LIMITED.
        mask: Float chroma range from -0.5/0.5 to 0.0/1.0.
        planes: Which planes to process.
        func: Function returned for custom error handling. This should only be set by VS package developers.

    Returns:
        Clamped clip.
    """
    if callable(clip_or_func):
        _func = clip_or_func

        @wraps(_func)
        def _wrapper(*args: P.args, **kwargs: P.kwargs) -> vs.VideoNode:
            return limiter(
                _func(*args, **kwargs),
                min_val,
                max_val,
                tv_range=tv_range,
                mask=mask,
                planes=planes,
                func=func or _func,
            )

        return _wrapper

    func = func or limiter
    clip = clip_or_func

    if clip is None:
        return partial(limiter, min_val=min_val, max_val=max_val, tv_range=tv_range, planes=planes, func=func)

    if all([clip.format.sample_type == vs.INTEGER, min_val is None, max_val is None, tv_range is False]):
        return clip

    if not (min_val == max_val is None):
        from ..utils import get_lowest_values, get_peak_values

        min_val = normalize_seq(min_val or get_lowest_values(clip, clip), clip.format.num_planes)
        max_val = normalize_seq(max_val or get_peak_values(clip, clip), clip.format.num_planes)

    return clip.vszip.Limiter(min_val, max_val, tv_range, mask, planes)

plane

plane(
    clip: VideoNode, index: SupportsIndex, /, strict: bool = True
) -> VideoNode

Extract a plane from the given clip.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

  • index

    (SupportsIndex) –

    Index of the plane to extract.

  • strict

    (bool, default: True ) –

    If False, removes _Matrix property when the input clip is RGB.

Returns:

  • VideoNode

    Grayscale clip of the clip's plane.

Source code in vstools/functions/utils.py
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
def plane(clip: vs.VideoNode, index: SupportsIndex, /, strict: bool = True) -> vs.VideoNode:
    """
    Extract a plane from the given clip.

    Args:
        clip: Input clip.
        index: Index of the plane to extract.
        strict: If False, removes `_Matrix` property when the input clip is RGB.

    Returns:
        Grayscale clip of the clip's plane.
    """
    if clip.format.num_planes == 1 and index.__index__() == 0:
        return clip

    if not strict and clip.format.color_family is vs.RGB:
        clip = vs.core.std.RemoveFrameProps(clip, "_Matrix")

    return vs.core.std.ShufflePlanes(clip, index.__index__(), vs.GRAY)

split

split(clip: VideoNode, /, strict: bool = True) -> list[VideoNode]

Split a clip into a list of individual planes.

Parameters:

  • clip

    (VideoNode) –

    Input clip.

  • strict

    (bool, default: True ) –

    If False, removes _Matrix property when the input clip is RGB.

Returns:

  • list[VideoNode]

    List of individual planes.

Source code in vstools/functions/utils.py
763
764
765
766
767
768
769
770
771
772
773
774
def split(clip: vs.VideoNode, /, strict: bool = True) -> list[vs.VideoNode]:
    """
    Split a clip into a list of individual planes.

    Args:
        clip: Input clip.
        strict: If False, removes `_Matrix` property when the input clip is RGB.

    Returns:
        List of individual planes.
    """
    return [clip] if clip.format.num_planes == 1 else [plane(clip, i, strict) for i in range(clip.format.num_planes)]

stack_clips

stack_clips(clips: Iterable[VideoNodeIterable]) -> VideoNode

Recursively stack clips in alternating directions: horizontal → vertical → horizontal → ...

This function takes a nested sequence of clips (or lists of clips) and stacks them alternately along the horizontal and vertical axes at each nesting level.

Examples:

  • Stack a list of clips horizontally:

    from vstools import core, split, vs
    
    clip = core.std.BlankClip(format=vs.RGB24)
    clips = split(clip)
    
    stacked = stack_clips(clips)
    

  • Stack a list of clips vertically (wrap in another list):

    from vstools import core, split, vs
    
    clip = core.std.BlankClip(format=vs.RGB24)
    clips = split(clip)
    
    stacked = stack_clips([clips])
    

  • Stack the Y plane horizontally with the U and V planes stacked vertically:

    from vstools import core, split, vs
    
    clip = core.std.BlankClip(format=vs.YUV420P8)
    y, u, v = split(clip)
    
    stacked = stack_clips([y, [u, v]])
    

  • Stack multiple YUV clips, with Y planes horizontally and UV planes vertically:

    from vstools import core, split, vs
    
    yuv_clips = [...]  # all must share format and height
    
    clips = []
    for yuv_clip in yuv_clips:
        y, *uv = split(yuv_clip)
        clips.extend([y, uv])
    
    stacked = stack_clips(clips)
    

  • Using append instead of extend (and wrapping the sequence, e.g. stack_clips([clips])) changes the stacking layout, since it alters the nesting depth.

Parameters:

  • clips

    (Iterable[VideoNodeIterable]) –

    A (possibly nested) sequence of clips to be stacked.

Returns:

  • VideoNode

    Stacked clips.

Source code in vstools/functions/utils.py
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
def stack_clips(clips: Iterable[VideoNodeIterable]) -> vs.VideoNode:
    """
    Recursively stack clips in alternating directions: horizontal → vertical → horizontal → ...

    This function takes a nested sequence of clips (or lists of clips) and stacks them
    alternately along the horizontal and vertical axes at each nesting level.

    Examples:
        - Stack a list of clips horizontally:
          ```py
          from vstools import core, split, vs

          clip = core.std.BlankClip(format=vs.RGB24)
          clips = split(clip)

          stacked = stack_clips(clips)
          ```

        - Stack a list of clips vertically (wrap in another list):
          ```py
          from vstools import core, split, vs

          clip = core.std.BlankClip(format=vs.RGB24)
          clips = split(clip)

          stacked = stack_clips([clips])
          ```

        - Stack the Y plane horizontally with the U and V planes stacked vertically:
          ```py
          from vstools import core, split, vs

          clip = core.std.BlankClip(format=vs.YUV420P8)
          y, u, v = split(clip)

          stacked = stack_clips([y, [u, v]])
          ```

        - Stack multiple YUV clips, with Y planes horizontally and UV planes vertically:
          ```py
          from vstools import core, split, vs

          yuv_clips = [...]  # all must share format and height

          clips = []
          for yuv_clip in yuv_clips:
              y, *uv = split(yuv_clip)
              clips.extend([y, uv])

          stacked = stack_clips(clips)
          ```

        - Using ``append`` instead of ``extend`` (and wrapping the sequence, e.g. ``stack_clips([clips])``)
          changes the stacking layout, since it alters the nesting depth.

    Args:
        clips: A (possibly nested) sequence of clips to be stacked.

    Returns:
        Stacked clips.
    """
    return vs.core.std.StackHorizontal(
        [
            inner_clips
            if isinstance(inner_clips, vs.VideoNode)
            else (
                vs.core.std.StackVertical(
                    [clipa if isinstance(clipa, vs.VideoNode) else stack_clips(clipa) for clipa in inner_clips]
                )
            )
            for inner_clips in clips
        ]
    )