2d fir filter ip core, Calculation precision, 2d fir filter ip core -1 – Altera Video and Image Processing Suite User Manual
Page 106: Calculation precision -1

2D FIR Filter IP Core
5
2015.05.04
UG-VIPSUITE
The 2D FIR Filter IP core performs 2D convolution using matrices of 3×3, 5×5, or 7×7 coefficients.
The 2D FIR Filter IP core retains full precision throughout the calculation while making efficient use of
FPGA resources. With suitable coefficients, the IP core performs operations such as sharpening,
smoothing, and edge detection. You can configure the 2D FIR Filter to change coefficient values at run
time with an Avalon-MM slave interface.
The 2D FIR Filter IP core calculates an output pixel from the multiplication of input pixels in a filter size
grid (kernel) by their corresponding coefficient in the filter. These values are summed together. To
produce the output, the IP core:
1. Scales the input.
2. Removes the fractional bits.
3. Converts the input to the desired output data type.
4. Finally, constrains the output to a specified range.
The position of the output pixel corresponds to the mid-point of the kernel. If the kernel runs over the
edge of an image, the IP core uses zeros for the out of range pixels.
The 2D FIR Filter IP core fully defines its input, output and coefficient data types.
• Input and output: constraints 4 to 20 bits per pixel per color plane.
• Coefficients: constraints up to 35 bits per pixel per color plane.
The 2D FIR Filter IP core supports symmetric coefficients—reduces the number of multipliers, resulting
in smaller hardware. You can set the Coefficients can be set at compile time, or changed at run time using
an Avalon-MM slave interface.
Calculation Precision
The 2D FIR Filter IP core does not lose calculation precision during the FIR calculation.
The calculation and result data types are derived from the range of input values (as specified by the input
data type, or input guard bands if provided), the coefficient fixed point type and the coefficient values. If
scaling is selected, then the result data type is scaled up appropriately such that precision is not lost.
©
2015 Altera Corporation. All rights reserved. ALTERA, ARRIA, CYCLONE, ENPIRION, MAX, MEGACORE, NIOS, QUARTUS and STRATIX words and logos are
trademarks of Altera Corporation and registered in the U.S. Patent and Trademark Office and in other countries. All other words and logos identified as
trademarks or service marks are the property of their respective holders as described at
www.altera.com/common/legal.html
. Altera warrants performance
of its semiconductor products to current specifications in accordance with Altera's standard warranty, but reserves the right to make changes to any
products and services at any time without notice. Altera assumes no responsibility or liability arising out of the application or use of any information,
product, or service described herein except as expressly agreed to in writing by Altera. Altera customers are advised to obtain the latest version of device
specifications before relying on any published information and before placing orders for products or services.
www.altera.com
101 Innovation Drive, San Jose, CA 95134