Method and apparatus for correcting errors in pixel characteristics when interpolating polygons into a pixel grid

• US 5,625,768 A
• Filed: 09/01/1994
• Issued: 04/29/1997
• Est. Priority Date: 05/23/1994
• Status: Expired due to Term

• Current Assignee: NVIDIA Corporation
• Original Assignee: Cirrus Logic Incorporated
• Inventors: Dye, Thomas A.
• Primary Examiner(s): Jankus, Almis R.
• Application Number:
  US08/299,739
• Time in Patent Office: 971 Days
• Field of Search: 395/141, 395/133, 395/142, 395/143, 395/150, 395/151, 395/119, 395/120, 395/126-132, 382/266-275
• US Class Current: 345/441
• CPC Class Codes: G06F 9/383 Operand prefetching cache prefetching G06F12/0862
• CPC Class Codes: G06F 9/3867 using instruction pipelines
• CPC Class Codes: G06F 9/3877 using a slave processor, e.g. coprocessor peripheral processor G06F13/12; vector processor G06F15/8053
• CPC Class Codes: G06F 9/3879 for non-native instruction execution, e.g. executing a command; for Java instruction set
• CPC Class Codes: G06T 1/20 Processor architectures; Processor configuration, e.g. pipelining
• CPC Class Codes: G06T 15/80 Shading

Abstract

An incremental orthogonal error correction process for 3D graphics for correcting errors or otherwise reducing abrupt changes in the characteristics of adjacent pixels while interpolating a polygon into a pixel grid. Error values are calculated for each desired characteristic to be corrected, where the error values are calculated based on the slope of the main slope and the orthogonal slope of the corresponding characteristic value. Thus, the error values adjust the characteristics of the pixels in the scan line to adjust for the slant of the main slope of the polygon. An interpolator for each corrected characteristic accumulates the error value in the opposite direction as the corresponding orthogonal slope. When the fractional component of the x parameter overflows, the orthogonal slope values are subtracted from the corresponding accumulated error values to thereby reduce the magnitude of the error values. Thus, the error values accumulate until an overflow causes a pixel shift of the scan line, and then the magnitude of the errors is reduced. The resulting values of the characteristics along the surface of the polygon appear smoother and monotonic. Furthermore, intersection lines between polygons drawn in 3D space appear straight rather than jagged.

Claims

1. A method of correcting errors of at least one variable characteristic of a polygon while interpolating the polygon into a pixel grid, wherein the pixel grid includes reference and orthogonal dimensions and corresponding coordinates defining pixel boundaries and wherein the polygon is interpolated with orthogonal span lines of pixels each beginning at an initial pixel along a main slope of the polygon, comprising the steps of:

calculating an error value for a variable characteristic based on an orthogonal component of the main slope and an orthogonal slope value of the variable characteristic;

incorporating the calculated error value into the variable characteristic of each initial pixel of the orthogonal scan lines in the numerically opposite direction as the characteristic orthogonal slope value of the variable characteristic;

indicating when the main slope crosses a pixel grid boundary while interpolating to a next orthogonal scan line due to an overflow of a fractional portion of the orthogonal component of the main slope; and

adding the orthogonal slope value of the variable characteristic to the variable characteristic of an initial pixel of the orthogonal scan line when fractional overflow is indicated.

2. The method of claim 1, wherein a variable characteristic of the polygon is pixel intensity.

3. The method of claim 1, wherein a variable characteristic of the polygon is depth, wherein depth values are included to simulate a depth dimension.

4. The method of claim 1, wherein a variable characteristic of the polygon is transparency in the form of alpha blending values.

5. The method of claim 1, wherein said step of calculating an error value comprises the steps of:

receiving vertices of the polygon;

determining a main slope spanning the reference dimension of the polygon between two of the polygon vertices;

calculating an orthogonal component of the main slope;

calculating an orthogonal slope value of the variable characteristic; and

multiplying a fractional portion of the orthogonal component of the main slope by the orthogonal slope value of the variable characteristic.

6. The method of claim 1, wherein said step of incorporating the calculated error value comprises the steps of:

accumulating the error value resulting in an accumulated error value having a numeric sign being the negative of the numeric sign of the orthogonal slope value of the variable characteristic; and

adding the accumulated error value to the variable characteristic of initial pixels of each orthogonal scan line.

7. The method of claim 6, wherein said step of adding the orthogonal slope value of the variable characteristic to the variable characteristic reduces the magnitude of the accumulated error value.

8. The method of claim 1, wherein said step of indicating when the main slope crosses a pixel grid boundary comprises the steps of:

interpolating the orthogonal component by adding an integer orthogonal slope value to an integer component and adding a fractional orthogonal slope value to a fractional component;

providing an overflow value from the fractional portion to the integer portion; and

indicating when the overflow value is provided.

9. The method of claim 8, wherein the fractional overflow is based on the overflow value and the numeric sign of the integer portion.

10. A method of correcting errors in pixel characteristics when interpolating a coplanar polygon into a pixel grid having a reference dimension and an orthogonal dimension and corresponding coordinates, the polygon being interpolated with orthogonal scan lines of pixels from a main slope of the polygon, comprising the steps of:

receiving an initial characteristic value of a base point at one end of the main slope, a main slope characteristic slope value, a characteristic orthogonal slope value and a characteristic error value;

accumulating an orthogonal coordinate beginning with an initial orthogonal coordinate of the base point and adding an orthogonal component of the main slope;

accumulating a main slope characteristic value by starting with the initial characteristic value and adding the main slope characteristic slope value for each pixel along the main slope;

accumulating an error value by adding characteristic error values for each pixel along the main slope;

adding the accumulated error value to the main slope characteristic value for each pixel along the main slope;

indicating a fractional overflow when a fractional portion of the orthogonal coordinate overflows to an integer portion; and

adding the characteristic orthogonal slope value to reduce the magnitude of the accumulated error value when fractional overflow is indicated.

11. The method of claim 10, wherein the pixel characteristic is pixel intensity.

12. The method of claim 10, wherein the pixel characteristic is pixel transparency.

13. The method of claim 10, wherein the pixel grid further includes a depth dimension and corresponding depth values, wherein the pixel characteristic is depth.

14. The method of claim 10, before said receiving step, further comprising the steps of:

calculating the main slope characteristic slope value based on the change of the characteristic along the main slope;

calculating the characteristic orthogonal slope value based on the change of the characteristic per pixel in the orthogonal direction; and

calculating the characteristic error value based on an orthogonal component of the main slope and the characteristic orthogonal slope value.

15. The method of claim 10, wherein during said step of accumulating an error value, the accumulated error value is accumulated in the opposite direction as the characteristic orthogonal slope value.

16. The method of claim 10, wherein said step of accumulating an orthogonal coordinate comprises the steps of:

adding an integer orthogonal slope value to the integer portion; and

adding a fractional orthogonal slope value to the fractional portion.

17. The method of claim 16, further comprising the steps of providing an overflow value from the fractional portion to the integer portion, wherein said step of indicating a fractional overflow occurs when the overflow value is provided.

18. An apparatus for incrementally correcting errors of at least one variable characteristic of a polygon while interpolating the polygon into a pixel grid, wherein the pixel grid includes reference and orthogonal dimensions and corresponding coordinates and wherein the polygon is interpolated with orthogonal span lines of pixels each beginning at an initial pixel along a main slope of the polygon, comprising:

means for calculating an initial orthogonal value, an orthogonal component of the main slope, an initial characteristic value of a characteristic, a characteristic orthogonal slope value of said characteristic and a characteristic error value based on said orthogonal component of the main slope and said characteristic orthogonal slope value;

an interpolator receiving said initial orthogonal value and said orthogonal component of the main slope for providing orthogonal coordinates of the initial pixels along the main slope;

a comparator circuit coupled to said interpolator for indicating a fractional overflow of an orthogonal coordinate; and

an accumulator circuit coupled to said comparator circuit and receiving said characteristic error value and said characteristic orthogonal slope value for incorporating said characteristic error value into characteristic values of the initial pixels along the main slope in a numerically opposite direction as said characteristic orthogonal slope value, said accumulator circuit further adding said characteristic orthogonal slope value to one of said characteristic values when said comparator circuit indicates fractional overflow.

19. The apparatus of claim 18, wherein said calculating means comprises a central processing unit executing a software driver routine.

20. The apparatus of claim 18, wherein said accumulator circuit comprises:

memory for storing said characteristic error value and said characteristic orthogonal slope value;

a register; and

an adder circuit coupled to said memory, said register and said comparator circuit, wherein said adder circuit adds the magnitude of said characteristic error value to the contents of said register for each orthogonal scan line while said fractional overflow is not indicated so that said register holds and accumulated error value having a numeric sign opposite the numeric sign of said characteristic orthogonal slope value, and wherein said adder circuit adds said characteristic orthogonal slope value to said accumulated error value when said fractional overflow is indicated.

21. The apparatus of claim 20, wherein said adder circuit comprises:

an arithmetic logic unit having a first data input coupled to the output of said register, a second data input, an output provided to said register and an input for receiving a function select signal, wherein said arithmetic logic unit either adds data provided to said second data input to said accumulated error value, or otherwise subtracts said data from said accumulated error value depending upon said function select signal; and

a select circuit coupled to said memory, said arithmetic logic unit and said comparator circuit, for providing either said characteristic orthogonal slope value when said fractional overflow is indicated, or else providing said characteristic error value to said second input of said arithmetic logic unit when said fractional overflow is not indicated, and for asserting said function select signal to indicate an add function when said fractional overflow is indicated or if the numeric signs of said characteristic orthogonal slope value and said characteristic error value are opposite, but otherwise providing said function select signal to indicate a subtract function.

22. The apparatus of claim 20, wherein said accumulator circuit further comprises an adder coupled to an output of said register for adding said accumulated error value to said characteristic values of the initial pixels along the main slope.

23. The apparatus of claim 18, wherein said interpolator comprises:

a first register for receiving said orthogonal component of the main slope;

a second register for receiving said initial orthogonal value; and

an adder circuit coupled to said first and second registers for adding the contents of said first and second registers and providing the sum back into said second register.

24. The apparatus of claim 23, wherein said first and second registers each have fractional and integer portions, wherein said adder circuit comprises:

a first adder for adding said fractional portions of said first and second registers and providing a carry output signal; and

a second adder receiving said carry input signal for adding said integer portions of said first and second registers and the value of said carry output signal.

25. The apparatus of claim 24, wherein said comparator circuit compares a sign bit of said integer portion of said first register to said carry output signal, and provides a fractional overflow signal indicative thereof.

26. The apparatus of claim 18, wherein said variable characteristic is color intensity.

27. The apparatus of claim 18, wherein said variable characteristic is depth, wherein depth values are included to simulate a third dimension.

28. The apparatus of claim 18, wherein said variable characteristic is transparency in the form of alpha blending values.

29. The apparatus of claim 18, wherein the polygon has at least two variable characteristics including depth and color intensity, further comprising:

said calculating means calculating an initial depth value and an initial intensity value, a depth slope value and an intensity slope value, a depth error value based on said orthogonal component of the main slope and said depth orthogonal slope value and an intensity error value based on said orthogonal component of the main slope and said intensity orthogonal slope value;

wherein said accumulator circuit receives said depth error value and said depth orthogonal slope value for incorporating said depth error value into depth values of the initial pixels along the main slope in a numerically opposite direction as said depth orthogonal slope value, said accumulator circuit further adding said depth orthogonal slope value to one of said depth values when said comparator circuit indicates fractional overflow; and

a second accumulator circuit receiving said intensity error value and said intensity orthogonal slope value for incorporating said intensity error value into intensity values of the initial pixels along the main slope in a numerically opposite direction as said intensity orthogonal slope value, said second accumulator circuit further adding said intensity orthogonal slope value to one of said intensity values when said comparator circuit indicates factional overflow.

30. An apparatus for interpolating a coplanar polygon into a pixel grid, the pixel grid having reference and orthogonal coordinates for locating each pixel in the pixel grid and further including memory for storing characteristic values for at least one characteristic of the polygon, and wherein the polygon is interpolated with orthogonal rows of pixels from a main side of the polygon, comprising:

means for calculating initial and incremental values for the reference and orthogonal coordinates and at least one characteristic of the polygon and for calculating a count value indicative of the number of orthogonal rows of pixels for the polygon, said incremental values including an orthogonal coordinate slope value representing the slope of the main side of the polygon in the orthogonal direction and an orthogonal characteristic slope value representing the change of the characteristic between orthogonally adjacent pixels, said calculating means further for calculating a characteristic error value based on said orthogonal coordinate slope value and said orthogonal characteristic slope value;

a reference interpolator for providing a reference coordinate for each of the orthogonal row of pixels based on initial and incremental reference coordinate values;

an orthogonal interpolator for providing an orthogonal coordinate for an initial pixel of each orthogonal row of pixels based on said orthogonal coordinate slope value, said orthogonal interpolator including integer and fractional portions;

an overflow circuit coupled to said orthogonal interpolator for determining when said fractional portion of said orthogonal interpolator overflows into said integer portion, and for providing an overflow signal indicative thereof;

a width interpolator for providing a width value based on initial and incremental orthogonal width values;

a characteristic interpolator for providing a characteristic value for each said initial pixel of the orthogonal rows of pixels along the main side of the polygon based on an initial characteristic value and a main side characteristic slope value; and

an error interpolator coupled to said characteristic interpolator, said overflow circuit and receiving said characteristic error value for accumulating said characteristic error value and for adding said accumulated error value to said characteristic value of each said initial pixel, and wherein said error interpolator further receives and adds said orthogonal characteristic slope value to reduce the magnitude of said accumulated error value when said overflow signal is provided.

31. The apparatus of claim 30, wherein said calculating means comprises a central processing unit executing a software driver routine.

32. The apparatus of claim 30, wherein said reference interpolator comprises:

a first register for receiving an incremental reference value;

a second register for receiving an initial reference coordinate value; and

an adder circuit coupled to said first and second registers for adding the contents of said first and second registers and providing the sum back into said second register, wherein a portion of said second register provides said reference coordinate for each of the orthogonal row of pixels.

33. The apparatus of claim 30, wherein said orthogonal interpolator comprises:

a first register for receiving said orthogonal coordinate slope value;

a second register for receiving an initial orthogonal value; and

an adder circuit coupled to said first and second registers for adding the contents of said first and second registers and providing the sum back into said second register.

34. The apparatus of claim 33, wherein said first and second registers each have fractional and integer portions, wherein said adder circuit comprises:

a first adder for adding said fractional portions of said first and second registers and providing a carry output signal; and

a second adder receiving said carry input signal for adding said integer portions of said first and second registers and the value of said carry output signal.

35. The apparatus of claim 34, wherein said overflow circuit compares a sign bit of said integer portion of said first register to said carry output signal, and provides a fractional overflow signal indicative thereof.

36. The apparatus of claim 30, wherein said width interpolator comprises:

a first register for receiving an incremental width value;

a second register for receiving an initial orthogonal width value; and

an adder circuit coupled to said first and second registers for adding the contents of said first and second registers and providing the sum back into said second register, wherein a portion of said second register provides said width values for each of the orthogonal row of pixels.

37. The apparatus of claim 30, wherein said characteristic interpolator comprises:

a first register for receiving said main side characteristic slope value;

a second register for receiving said initial characteristic value; and

an adder circuit coupled to said first and second registers for adding the contents of said first and second registers and providing the sum back into said second register, wherein a portion of said second register provides said characteristic values for each initial pixel of the orthogonal rows of pixels.

38. The apparatus of claim 30, wherein said error interpolator comprises:

memory for storing said characteristic error value and said orthogonal characteristic slope value;

a register; and

an adder circuit coupled to said memory, said register and said overflow circuit, wherein said adder circuit adds the magnitude of said characteristic error value to the contents of said register for each orthogonal row of pixels while said overflow signal does not indicate fractional overflow, so that said register holds and accumulated error value having a numeric sign opposite the numeric sign of said orthogonal characteristic slope value, and wherein said adder circuit adds said orthogonal characteristic slope value to said accumulated error value when said fractional overflow is indicated.

39. The apparatus of claim 38, wherein said adder circuit comprises:

an arithmetic logic unit having a first data input coupled to the output of said register, a second data input, an output provided to said register and an input for receiving a function select signal, wherein said arithmetic logic unit either adds data provided to said second data input to, or subtracts said data from, said accumulated error value depending upon said function select signal; and

a select circuit coupled to said memory, said arithmetic logic unit and said comparator circuit, for providing either said orthogonal characteristic slope value when said fractional overflow is indicated, or else providing said characteristic error value to said second input of said arithmetic logic unit when said fractional overflow is not indicated, and for asserting said function select signal to indicate an add function when said fractional overflow is indicated or if the numeric signs of said orthogonal characteristic slope value and said characteristic error value are opposite, but otherwise providing said function select signal to indicate a subtract function.

40. The apparatus of claim 38, wherein said error interpolator further comprises an adder coupled to an output of said register for adding said accumulated error value to said characteristic values of said initial pixels of each orthogonal row of pixels.

41. The apparatus of claim 30, further comprising a controller receiving said count value and coupled to said reference interpolator, said orthogonal interpolator, said width interpolator, said characteristic interpolator and said error interpolator for asserting signals to control the interpolation process.

42. The apparatus of claim 30, wherein said coplanar polygon includes a first and a second side opposite the main side, further comprising:

wherein said count value includes first and second count values corresponding to the first and second opposite sides, respectively, wherein said first count value represents the number of orthogonal rows of pixels to complete the polygon between the main side and the first opposite side, and wherein said second count value represents the number of orthogonal rows of pixels to complete the polygon between the main side and the second opposite side;

wherein said width interpolator comprises;

a first register for receiving a first incremental width value corresponding to the first side;

a second register for receiving a second incremental width value corresponding to the second side;

a third register for receiving a first and a second initial orthogonal width value;

a select circuit having two inputs coupled to said first and second registers, respectively, and an output, said select circuit further receiving a select signal for selecting between said first and second registers and providing the contents of the selected register to said output depending upon said select signal; and

an adder circuit coupled to said select circuit and said third register for adding the contents of said register to the output of said select circuit and providing the sum back into said third register, wherein a portion of said third register provides said width values for each of the orthogonal row of pixels; and

a controller receiving said first and second count values and providing said select signal to said width interpolator.

43. The apparatus of claim 30, wherein one characteristic of the polygon is depth.

44. The apparatus of claim 30, wherein one characteristic of the polygon is color intensity.