Simple custom node pack for nodes I use in my workflows. Includes Dilate Mask Linear for animating masks.

Akatz Custom Nodes


    Adjusts the brightness of each frame in a depth map video batch to match the brightness of the first frame.
    - depthmap_batch: Batch of depth maps, shape [B, H, W, C].
    

AK_AdjustDepthmapBrightness


    # AK Adjust List Size
    Adjust the size of a list of floats to match the specified batch size.
    - float_list: The input list of float values.
    - batch_size: The desired batch size for the output list.
    

AK_AdjustListSize


    # Animated Dilate Mask Linear
    - mask: Input mask or mask batch
    - shape: "circle" or "square", "circle" is most accurate to mask shape, "square" is fast to compute for testing purposes
    - step: how much should the mask be dilated per frame
    - delay: delay in frames before starting dilation
    

AK_AnimatedDilationMaskLinear


    This node syncs audio to frames by calculating the loudness of the audio at each frame.
    

AK_AudioFramesyncSchedule


    # Audioreactive Dilate Mask Infinite
    - mask: Input mask or mask batch
    - normalized_amp: The normalized amplitude values
    - mask_colors: Colors for the dilation masks in the format "(r, g, b), (r, g, b), ..."
    - threshold: The threshold of the dilation
    - dilation_speed: Speed of dilation in pixels per frame
    - quality_factor: Quality factor for dilation
    - should_composite_subject: Boolean to composite the subject mask over the final result
    - subject_mask_color: Color for the subject mask in the format "R, G, B"
    - initial_background_color: Color for the initial background in the format "R, G, B"
    - start_frame: Start frame for the dilation
    - end_frame: End frame for the dilation (0 for infinite)
    

AK_AudioreactiveDilateMaskInfinite


    # Dilate Mask with Amplitude
    - masks: The mask to dilate
    - norm_amps: The normalized amplitude values
    - fps: Frames per second for time-based calculations
    - shape: The shape of the dilation
    - max_radius: The maximum radius of the dilation
    - min_radius: The minimum radius of the dilation
    - threshold: The threshold of the dilation
    - attack: The attack duration in seconds
    - decay: The decay duration in seconds
    - attack_function: The attack easing function
    - decay_function: The decay easing function
    

AK_AudioreactiveDilationMask


    # Dilate Mask dynamically based on Amplitude
    - masks: The mask to dilate
    - norm_amps: The normalized amplitude values
    - shape: The shape of the dilation
    - max_radius: The maximum radius of the dilation
    - min_radius: The minimum radius of the dilation
    - quality_factor: The quality factor of the dilation
    

AK_AudioreactiveDynamicDilationMask


    # AK Binary Amplitude Gate
    Apply a binary amplitude gate to a list of float values.
    - float_list: The input list of float values.
    - min_value: The value to set if the float is below the threshold.
    - max_value: The value to set if the float is equal to or above the threshold.
    - threshold: The threshold to determine the gating.
    

AK_BinaryAmplitudeGate


    # AK Blob Track (SimpleBlobDetector)
    
    1. Uses frame differencing (with 'cache_frames') plus a threshold to isolate movement.
    2. Feeds the threshold image to OpenCV's SimpleBlobDetector.
    3. For each detected blob, draws:
       - An alpha-blended square outline in the composite.
       - A fully filled white rectangle for the blob in the mask.
    4. Also draws connecting lines between blob centers in composite.
       - If line_alpha > 0, the lines are drawn in the mask; otherwise, they are skipped.
    

AK_BlobTrack


    # Brightness to Float List
    - image: Input image (tensor object)
    - This node calculates the average pixel brightness for each frame and normalizes it between 0.0 (black) and 1.0 (white).
    

AK_BrightnessToFloatList


    # Converts a PyTorch tensor representing audio data into raw audio bytes in WAV format.

    Parameters:
    - audio: LazyAudioMap-like object containing the waveform and sample rate

    Returns:
    - audio_bytes: Raw audio bytes in WAV format
    

AK_ConvertAudioToSaltAudio


    # AK Convert to Float Array
    Convert any input list type (NumPy array, Python integer list, Python float list) into a Python float array.
    - input_list: The input list or array to be converted to a Python float array.
    

AK_ConvertListToFloatList


    # Converts raw audio bytes (in WAV format) into a LazyAudioMap-like format.

    Parameters:
    - audio_bytes: Raw audio bytes in WAV format

    Returns:
    - lazy_audio_map: A LazyAudioMap-like object containing the waveform and sample rate
    

AK_ConvertSaltAudioToAudio


    # Dilate Mask Linear Infinite
    - mask: Input mask or mask batch
    - dilation_schedule: Schedule for mask dilations in the format:
      (start_frame, dilation_speed, (r, g, b)),...
    - quality_factor: Quality factor for dilation (recommend between 0.25 and 0.15 for good balance between quality and speed)
    - use_percentage: Boolean to specify if the start_frame is in percentage of the total frames
    - should_composite_subject: Boolean to composite the subject mask over the final result
    - subject_mask_color: Color for the subject mask in the format "R, G, B"
    - initial_background_color: Color for the initial background in the format "R, G, B"
    

AK_DilateMaskLinearInfinite


    # AK Fade Between Batches
    This node takes two image batches and blends them together by transitioning 
    with overlapping frames. The output is a single image batch where:
    - image1 starts,
    - a number of overlap frames fade between image1 and image2,
    - image2 finishes.
    

AK_FadeBetweenBatches


    AK_FlexFeatureToFloatList:
    This node converts a FEATURE type input into a list of float values, one per frame.
    
    - feature: A custom Feature object with attributes:
      - feature.get_value_at_frame(i): Method that returns a float value for frame `i`.
      - feature.frame_count: The total number of frames in the feature.
    

AK_FlexFeatureToFloatList


    # Float List to Dilate Mask Schedule
    - float_list: Input float list
    - mask_colors: Colors for the dilation masks in the format "(r, g, b), (r, g, b), ..."
    - threshold: The threshold of the dilation
    - dilation_speed: Speed of dilation in pixels per frame
    - This node transforms the input float list and parameters into a dilation mask schedule string.
    

AK_FloatListToDilateMaskSchedule


    AK_FloatListToFlexFeature:
    This node converts a list of float values into a FEATURE type input.
    
    - original_feature: A custom Feature object with attributes:
      - original_feature.get_value_at_frame(i): Method that returns a float value for frame `i`.
      - original_feature.frame_count: The total number of frames in the feature.
    

AK_FloatListToFlexFeature


    Used to provide custom timings for crossfading multiple images using just two IPAdapters.
    Text input should be a string of weights in the format:
    (<weights (float or float tuple)>, <frame_index (int or float)>, <transition_frames (int or float, optional)>, <easing_function (string, optional)>),...
    E.g. Frame: "((0.0, 0.0), 0), (1.0, 5, 20, linear), (0.5, 25, 10, ease_in), ((0.0, 0.0), 48, 24, ease_out)"
    Percentage: "((0.0, 0.0), 0), (1.0, 0.1, 0.1, linear), (0.5, 0.25, 0.1, ease_in), ((0.0, 0.0), 0.5, 0.25, ease_out)"
    A default timing of "ease_in_out" is used if no interpolation function is specified.
    - weights: The weights and timings to be applied to the images
    - default_weights: The default starting weights (weights, weights_invert) to be used for transitions
    - default_easing: The default easing function to be used for transitions
    - timing_mode: The timing mode to be used for transitions (Frame or Percent)
    - frames: The number of frames in the output
    - image: The image batch to be crossfaded by the weights
    

AK_IPAdapterCustomWeights

AK_KeyframeScheduler


    # AK Lag Chop
    Apply a lag effect to a list of float values.
    - float_list: The input list of float values.
    - lag_factor: The factor by which the output lags behind the input (0 < lag_factor <= 1).
    

AK_LagChop


    # AK List to Numpy Float Array
    Convert a list of float values to a NumPy float array.
    - float_list: The input list of float values to be converted to a NumPy array.
    

AK_ListToNumpyFloatArray


    # AK Make Depthmap Seamless
    Adjusts depth maps in the batch to become seamless by fitting and removing a plane using least-squares.
    - depthmap_batch: The input depth maps to be adjusted to become seamless.
    

AK_MakeDepthmapSeamless


    # AK Normalize Image Color
    Sets the non-black pixels of input images to one specified color.
    - image: The input video tensor with shape (num_frames, H, W, C).
    - threshold: The threshold value for determining black pixels.
    - r: Red channel value (0-255) for the non-black pixels.
    - g: Green channel value (0-255) for the non-black pixels.
    - b: Blue channel value (0-255) for the non-black pixels.
    

AK_NormalizeMaskImage


    # AK Rescale Float List
    Rescale a list of float values from an original range to a new range.
    - float_list: The input list of float values to be rescaled.
    - orig_min: The minimum value of the original range.
    - orig_max: The maximum value of the original range.
    - new_min: The minimum value of the new range.
    - new_max: The maximum value of the new range.
    

AK_RescaleFloatList

AK_ScaleMask

AK_ScheduledBinaryComparison

AK_ShrinkNumSequence


    # AK Split Image Batch
    This node splits an input image batch into two at a given index, and returns one of the partitions.
    - split_index: The index at which to split the image batch.
    - split_batch_index: 0 to return the first partition, 1 to return the second partition.
    

AK_SplitImageBatch


    # AK Speed Adjust
    Adjust the speed of the video dynamically based on the speed schedule.
    - image_batch: The input batch of images with shape (B, H, W, C).
    - speed_schedule: A list of floats where each value represents the speed at that frame.
    - fps: The frames per second of the original video.
    

ComfyUI Extension: Akatz Custom Nodes

Custom Nodes (30)

README

AKatz Nodes