# US 2001/0055025 A1 · Graphics system configured to filter samples using a variable support filter

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2001/0055025 A1
- Document ID: US20010055025A1
- Status: Published application
- Kind: A1
- Application number: 09/755,643
- Filed: 2001-01-04
- Published: 2001-12-27
- Earliest claimed: 1998-02-17
- Assignees: Sun Microsystems, Inc.
- Inventors: Michael F. Deering; N. David Naegle; Scott R. Nelson
- Web page: https://patentgrep.com/patent/US20010055025A1

## CPC

- G06T 5/20: using local operators
- G06T 5/70: Denoising; Smoothing
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A computer graphics system may comprise a graphics processor, a sample buffer, and a sample-to-pixel calculation unit. The graphics processor renders samples into the sample buffer in response to received graphics data. The sample-to-pixel calculation unit generates a plurality of output pixels by filtering the rendered samples based on a filter function. The pixels may be computed by generating a weighted sum of sample values (e.g. red sample values) for samples falling within the filter support. The coefficients used in the weighted sum may be added to form a normalization factor. One weighted sum of sample values may be computed per pixel attribute such as red, green, blue and alpha. The normalization factor may be computed in parallel with one or more of the weighted sums. Normalized pixel values may be obtained by dividing the weighted-sums by the normalization factor.

## Claims

### Claim 1

A method for generating pixels for a display device, the method comprising:
receiving graphics data;
rendering a plurality of samples into a sample buffer in response to said graphics data; and
filtering said plurality of samples to generate one or more pixels, wherein said filtering comprises:
(a) reading said plurality of samples from the sample buffer;
(b) determining filter coefficients for a subset of said plurality of samples;
(c) multiplying a first sample attribute of each sample in said subset by a corresponding one of said filter coefficients to generate first weighted sample attributes;
(d) adding the first weighted sample attributes to generate a first weighted attribute sum;
(e) computing a normalization factor by adding the filter coefficients which correspond to said subset of samples;
(f) dividing the first weighted attribute sum by said normalization factor to generate a first pixel attribute for a first of said one or more pixels.

### Claim 2 (depends on 1)

The method of claim 1, wherein said first sample attribute of each sample in said subset is a sample color value.

### Claim 3 (depends on 1)

The method of claim 1, wherein said determining said filter coefficients comprises evaluating a filter function for each of said samples in said subset, wherein the filter function is defined over a filter support, wherein the filter support overlaps two or more regions having different sample densities.

### Claim 4 (depends on 3)

The method of claim 3, wherein the filter support has a first area in a first region of said two or more regions, and a second area in a second region of said two or more regions, wherein the first area is proportional to the inverse of a first sample density of said first region, wherein the second area is proportional to the inverse of a second sample density of said second region.

### Claim 5 (depends on 1)

The method of claim 1, wherein said normalization factor is computed for each of said one or more pixels.

### Claim 6 (depends on 1)

The method of claim 1, wherein said plurality of samples are positioned in a two-dimensional sample space according to an irregular grid.

### Claim 7 (depends on 1)

The method of claim 1, wherein (d) and (e) are performed in parallel.

### Claim 8 (depends on 1)

The method of claim 1, wherein said determining said filter coefficients comprises evaluating a filter function for each of said samples in said subset, wherein the filter function is defined over a filter support, the method further comprising:
determining if the filter support intersects a first region having a first sample density and a second region having a second sample density;
multiplying the filter coefficients corresponding to samples in said second region by a factor equal to the ratio of said first density to said second density in response to the filter support area intersecting the first region and the second region.

### Claim 9 (depends on 1)

The method of claim 1, wherein (b) comprises:
evaluating a square distance of a first sample of said subset from a center of a first pixel of said one or more pixels;
accessing a look-up table using the square distance, wherein the lookup table stores values of a filter function.

### Claim 10 (depends on 1)

The method of claim 1, wherein the filter coefficients correspond to a first filter, wherein (b) comprises:
evaluating a square distance of a first sample of said subset from a center of said first filter;
accessing a look-up table using the square distance, wherein the lookup table stores values of the first filter.

### Claim 11 (depends on 1)

The method of claim 1 further comprising transmitting the first pixel attribute for said first pixel to a display device without an intervening frame buffer.

### Claim 12 (depends on 1)

The method of claim 1, wherein said rendering said plurality of samples into said sample buffer includes double buffering at least a portion of said plurality of samples.

### Claim 13 (depends on 1)

The method of claim 1, wherein said filtering comprises selecting and filtering said subset of samples to form output pixels on one of: a real time basis or an on-the-fly basis.

### Claim 14

A computer software program embodied on a computer-readable medium, said computer software program comprising a plurality of instructions, wherein said instructions are configured to:
receive graphics data;
render a plurality of samples into a sample buffer in response to said graphics data; and
filter said plurality of samples to generate one or more pixels, wherein said filtering comprises:
(a) reading said plurality of samples from the sample buffer;
(b) determining filter coefficients for a subset of said plurality of samples;
(c) multiplying a first sample attribute of each sample in said subset by a corresponding one of said filter coefficients to generate first weighted sample attributes;
(d) adding the first weighted sample attributes to generate a first weighted attribute sum;
(e) computing a normalization factor by adding the filter coefficients which correspond to said subset of samples;
(f) dividing the first weighted attribute sum by said normalization factor to generate a first pixel attribute for a first of said one or more pixels.

### Claim 15 (depends on 14)

The computer software program of claim 14, wherein said first sample attribute of each sample in said subset is a sample color value.

### Claim 16 (depends on 14)

The computer software program of claim 14, wherein said determining said filter coefficients comprises evaluating a filter function for each of said samples in said subset, wherein the filter function is defined over a filter support, wherein the filter support overlaps two or more regions having different sample densities.

### Claim 17 (depends on 16)

The computer software program of claim 16, wherein the filter support has a first area in a first region of said two or more regions, and a second area in a second region of said two or more regions, wherein the first area is proportional to the inverse of a first sample density of said first region, wherein the second area is proportional to the inverse of a second sample density of said second region.

### Claim 18 (depends on 14)

The computer software program of claim 14, wherein said plurality of samples are positioned in a two-dimensional sample space according to an irregular grid.

### Claim 19 (depends on 14)

The computer software program of claim 14, wherein said determining said filter coefficients comprises evaluating a filter function for each of said samples in said subset, wherein the filter function is defined over a filter support, wherein the instructions are further configured to:
determine if the filter support intersects a first region having a first sample density and a second region having a second sample density; and
multiply the filter coefficients corresponding to samples in said second region by a factor equal to the ratio of said first density to said second density in response to the filter support area intersecting the first region and the second region.

### Claim 20 (depends on 14)

The computer software program of claim 14, wherein (b) comprises:
evaluating a square distance of a first sample of said subset from a center of a first pixel of said one or more pixels;
accessing a look-up table using the square distance, wherein the lookup table stores values of a filter function.

### Claim 21 (depends on 14)

The computer software program of claim 14, wherein the filter coefficients correspond to a first filter, wherein (b) comprises:
evaluating a square distance of a first sample of said subset from a center of said first filter;
accessing a look-up table using the square distance, wherein the lookup table stores values of the first filter.

### Claim 22 (depends on 14)

The computer software program of claim 14, wherein the instructions are further configured to transmit the first pixel attribute for said first pixel to a display device without an intervening frame buffer.

### Claim 23 (depends on 14)

The computer software program of claim 14, wherein said rendering said plurality of samples into said sample buffer includes double buffering at least a portion of said plurality of samples.

### Claim 24

A graphics system comprising:
a rendering unit operable to receive graphics data and to render a plurality of samples in response to the graphics data;
a sample buffer configured to store the plurality of samples;
a sample-to-pixel calculation unit coupled to the sample buffer, and configured to filter said plurality of samples to generate one or more pixels by:
(a) reading said plurality of samples from the sample buffer;
(b) determining filter coefficients for a subset of said plurality of samples;
(c) multiplying a first sample attribute of each sample in said subset by a corresponding one of said filter coefficients to generate first weighted sample attributes;
(d) adding the first weighted sample attributes to generate a first weighted attribute sum;
(e) computing a normalization factor by adding the filter coefficients which correspond to said subset of samples;
(f) dividing the first weighted attribute sum by said normalization factor to generate a first pixel attribute for a first of said one or more pixels.

### Claim 25 (depends on 24)

The graphics system of claim 24, wherein said first sample attribute of each sample in said subset is a sample color value.

### Claim 26 (depends on 24)

The graphics system of claim 24, wherein said sample-to-pixel calculation unit is configured to determine said filter coefficients based on a filter function, wherein the filter function is defined over a filter support, wherein the filter support overlaps two or more regions having different sample densities.

### Claim 27 (depends on 25)

The graphics system of claim 25, wherein the filter support has a first area in a first region of said two or more regions, and a second area in a second region of said two or more regions, wherein the first area is proportional to the inverse of a first sample density of said first region, wherein the second area is proportional to the inverse of a second sample density of said second region.

### Claim 28 (depends on 24)

The graphics system of claim 24, wherein said rendering unit is configured to position the plurality of samples in a two-dimensional sample space according to an irregular grid.

### Claim 29 (depends on 24)

The graphics system of claim 24, wherein said sample-to-pixel calculation unit is configured to determine said filter coefficients comprises by evaluating a filter function for each of said samples in said subset, wherein the filter function is defined over a filter support, wherein the sample-to-pixel calculation unit is further configured to:
determine if the filter support intersects a first region having a first sample density and a second region having a second sample density; and
multiply the filter coefficients corresponding to samples in said second region by a factor equal to the ratio of said first density to said second density in response to the filter support area intersecting the first region and the second region.

### Claim 30 (depends on 24)

The graphics system of claim 24, wherein the sample-to-pixel calculation unit is further configured to transmit the first pixel attribute for said first pixel to a display device without an intervening frame buffer.

### Claim 31

A method for generating pixels for a display device, the method comprising:
receiving graphics data;
rendering one or more samples in response to said graphics data; and
applying a filter to at least a subset of the plurality of samples to generate one or more output pixels, wherein each sample comprises one or more sample attributes, wherein each output pixel comprises one or more pixel attributes, and wherein said applying the filter comprises:
applying weighting factors to the sample attributes;
summing the weighted sample attributes to form pixel attributes;
computing one or more normalization factors based on the weighting factors; and
normalizing the pixel attributes using the one or more normalization factors.

### Claim 32 (depends on 31)

The method of claim 31, wherein said normalizing comprises dividing the pixel attributes by said one or more normalization factor.

### Claim 33 (depends on 31)

The method of claim 31, wherein said applying comprises multiplying said weighting factors to the sample attributes.

### Claim 34 (depends on 31)

The method of claim 31, further comprising double buffering the rendered samples.

### Claim 35 (depends on 31)

The method of claim 31, wherein said applying is performed in real-time or on-the-fly.

### Claim 36 (depends on 31)

The method of claim 31, further comprising looking up the weighting factors for each particular sample based on the distance of the particular sample from a predetermined position in the filter.

### Claim 37 (depends on 36)

The method of claim 36, wherein the predetermined position corresponds to the center of the filter.

### Claim 38 (depends on 36)

The method of claim 36, wherein the predetermined position corresponds to the center of the output pixel.

### Claim 39 (depends on 31)

The method of claim 31, further comprising determining which samples fall within the support of the filter.

### Claim 40 (depends on 31)

The method of claim 31, wherein the samples are rendered into the sample buffer in bins.
