technology enables low power and low cost waveletbased compression
Supports both 9/7 and 5/3 wavelet transforms with up to
6 levels of transform
Programmable tile/image size with widths up to 2048 pixels
in 3-component 4:2:2 interleaved mode, and up to
4096 pixels in single-component mode
Maximum tile/image height: 4096 pixels
Video interface directly supporting ITU.R-BT656,
SMPTE125M PAL/ NTSC, SMPTE274M, SMPTE293M (525p),
ITU.R-BT1358 (625p) or any video format with a maximum
input rate of 65 MSPS for irreversible mode or 40 MSPS for
reversible mode
Two or more ADV202s can be combined to support full-
frame SMPTE274M HDTV (1080i) or SMPTE296M (720p)
Interlaces temporally coherent frame-based SD video
sources for improved performance
Flexible asynchronous SRAM-style host interface allows
glueless connection to most 16-/32-bit microcontrollers
and ASICs
2.5 V to 3.3 V I/O and 1.5 V core supply
12 mm × 12 mm 121-lead CSPBGA, speed grade 115 MHz, or
13 mm × 13 mm 144-lead CSPBGA, speed grade 135 MHz, or
13 mm × 13 mm 144-lead CSPBGA, speed grade 150 MHz
ADV202
APPLICATIONS
Networked video and image distribution systems
Wireless video and image distribution
Image archival/retrieval
Digital CCTV and surveillance systems
Digital cinema systems
Professional video editing and recording
Digital still cameras
Digital camcorders
GENERAL DESCRIPTION
The ADV202 is a single-chip JPEG2000 codec targeted for
video and high bandwidth image compression applications that
can benefit from the enhanced quality and feature set provided
by the JPEG2000 (J2K)—ISO/IEC15444-1 image compression
standard. The part implements the computationally intensive
operations of the JPEG2000 image compression standard as
well as providing fully compliant code-stream generation for
most applications.
The ADV202’s dedicated video port provides glueless
connection to common digital video standards such as ITU.RBT656, SMPTE125M, SMPTE293M (525p), ITU.R-BT1358
(625p), SMPTE274M(1080i), or SMPTE296M(720p). A variety
of other high speed synchronous pixel and video formats can
also be supported using the programmable framing and
validation signals.
FUNCTIONAL BLOCK DIAGRAM
ADV202
PIXEL I/F
HOST I/F
Rev. B
Information furnished by Analog Devices is believed to be accurate and reliable.
However, no responsibility is assumed by Analog Devices for its use, nor for any
infringements of patents or other rights of third parties that may result from its use.
Specifications subject to change without notice. No license is granted by implication
or otherwise under any patent or patent rights of Analog Devices. Trademarks and
registered trademarks are the property of their respective owners.
Changes to Features ..........................................................................1
Changes to Table 2.............................................................................4
Changes to Table 16 ....................................................................... 24
Changes to Table 23 ....................................................................... 32
7/04—Revision 0: Initial Version
Page 3
ADV202
GENERAL DESCRIPTION
(continued from Page 1)
The ADV202 can process images at a rate of 40 MSPS in
reversible mode and at higher rates when used in irreversible
mode. The ADV202 contains a dedicated wavelet transform
engine, three entropy codecs, an on-board memory system, and
an embedded RISC processor that can provide a complete
JPEG2000 compression/decompression solution.
The wavelet processor supports the 9/7 irreversible wavelet
transform and the 5/3 wavelet transform in reversible and
irreversible modes. The entropy codecs support all features in
the JPEG2000 Part 1 specification, except Maxshift ROI.
The ADV202 operates on a rectangular array of pixel samples
called a tile. A tile can contain a complete image, up to the
maximum supported size, or some portion of an image. The
maximum horizontal tile size supported depends on the wavelet
transform selected and the number of samples in the tile.
Images larger than the ADV202’s maximum tile size can be
broken into individual tiles and then sent sequentially to the
chip while still maintaining a single, fully compliant JPEG2000
code stream for the entire image.
JPEG2000 FEATURE SUPPORT
The ADV202 supports a broad set of features that are included
in Part 1 of the JPEG2000 standard (ISO/IEC 15444). See
Getting Started with ADV202 for information on the JPEG2000
features that the ADV202 currently supports.
Depending on the particular application requirements, the
ADV202 can provide varying levels of JPEG2000 compression
support. It can provide raw code-block and attribute data
output, which allows the host software to have complete control
over the generation of the JPEG2000 code stream and other
aspects of the compression process such as bit-rate control.
Otherwise, the ADV202 can create a complete, fully compliant
JPEG2000 code stream (.j2c) and enhanced file formats such as
.jp2, .jpx, and .mj2 (Motion JPEG2000). See Getting Started with ADV202 for information on the formats that the ADV202
currently supports.
Rev. B | Page 3 of 40
Page 4
ADV202
SPECIFICATONS
SUPPLY VOLTAGES AND CURRENT
Table 1.
Parameter Description Min Typ Max Unit
VDD DC Supply Voltage, Core 1.425 1.5 1.575 V
IOVDD DC Supply Voltage, I/O 2.375 3.3 3.63 V
PLLVDD DC Supply Voltage, PLL 1.425 1.5 1.575 V
V
Input Range −0.3 V
Input
Temp Operating Ambient Temperature Range in Free Air −40 +25 +85 °C
IDD Static Current1 300 mA
Dynamic Current, Core (JCLK Frequency = 150 MHz)
2
570 mA
Dynamic Current, Core (JCLK Frequency = 108 MHz) 420 mA
Dynamic Current, Core (JCLK Frequency = 81 MHz) 325 mA
Dynamic Current, I/O 20 mA
Dynamic Current, PLL 2.6 mA
1
No clock or I/O activity.
2
ADV202-150 only.
INPUT/OUTPUT SPECIFICATIONS
Table 2.
Parameter Description Test Conditions Min Typ Max Unit
V
High Level Input Voltage VDD = max 2.2 V
IH (3.3 V)
V
IH (2.5 V)
V
IL (3.3 V, 2.5 V)
V
OH (3.3 V)
V
High Level Output Voltage VDD = min, IOH = −0.5 mA 2.0 V
OH (2.5 V)
V
OL (3.3 V, 2.5 V)
IIH High Level Input Current VDD = max, VIN = VDD 1.0 µA
IIL Low Level Input Current VDD = max, VIN = 0V 1 µA
I
High Level Three-State Leakage Current VDD = max, VIN = VDD 1.0 µA
HOLDSU HOLD Setup to Rising MCLK 3 ns
HOLDHD HOLD Hold from Rising MCLK 3 ns
JDATASU JDATA Setup to Rising MCLK 3 ns
JDATAHD JDATA Hold from Rising MCLK 3 ns
VDATATD VCLK to VDATA Valid Delay (VDATA Output) 12 ns
VDATASU VDATA Setup to Rising VCLK (VDATA Input) 4 ns
VDATAHD VDATA Hold from Rising VCLK (VDATA Input) 4 ns
HSYNCSU HSYNC Setup to Rising VCLK 3 ns
HSYNC
HD
HSYNC
TD
VSYNC
SU
VSYNCHD VSYNC Hold from Rising VCLK 4 ns
VSYNCTD VCLK to VSYNC Valid Delay 12 ns
FIELD
SU
HSYNC Hold from Rising VCLK 4 ns
VCLK to HSYNC Valid Delay 12 ns
VSYNC Setup to Rising VCLK 3 ns
FIELD Setup to Rising VCLK 4 ns
Rev. B | Page 15 of 40
Page 16
ADV202
V
)
V
)
V
)
Parameter Description Min Typ Max Unit
FIELDHD FIELD Hold from Rising VCLK 3 ns
FIELD
TD
SYNC DELAY Decode Data Sync Delay for HD Input with EAV/SAV Codes 7 VCLK cycles
Decode Data Sync Delay for SD Input with EAV/SAV Codes 9 VCLK cycles
Decode Data Sync Delay for DUAL_LANE (Extended) Input 7 VCLK cycles
VCLK to FIELD Valid 12
Decode Data Sync Delay for HVF Input (from First Rising VCLK after
10 VCLK cycles
HSYNC Low to First Data Sample)
VCLK
VDATA(IN)
VCLK
DATA(OUT
VCLK
DATA(OUT
VCLK
DATA(OUT
HSYNC
VSYNC
VCLK
VDATA
HD
VDATA
VDATA
TD
VDATA
SU
TD
SYNC DELAY
TD
SYNC DELAY
CbYCrY
HSYNCHD*
VSYNCHD*
CrYCbYFFEAVFFSAVCbYCr
ENCODE CCIR-656 LINE
CrYCbYFFEAVFFSAVCbYCr
DECODE MASTER CCIR-656 LINE
VDATA
CrYYCb YFFEAVFFSAVCbY
DECODE SLAVE CCIR-656 LINE
CbYCrYCbCbY
DECODE SLAVE HVF MODE
VDATA(IN)
HSYNC
VSYNC
ENCODE HVF MODE
*HSYNC AND VSYNC DO NOT HAVE TO BE APPLIED SIMULTANEOUSLY
CrYYCbYCrYCbYYCrYCbCb
HSYNC
VSYNC
SU
SU
HSYNC
VSYNC
HD
HD
04723-030
Figure 21. Video Mode Timing
Rev. B | Page 16 of 40
Page 17
ADV202
V
RAW PIXEL MODE TIMING
Table 12.
Parameter Description Min Typ Max Unit
VDATATD VCLK to PIXELDATA Valid Delay (PIXELDATA Output) 12 ns
VDATASU PIXELDATA Setup to Rising VCLK (PIXELDATA Input) 4 ns
VDATAHD PIXELDATA Hold from Rising VCLK (PIXELDATA Input) 4 ns
VRDY
TD
VFRMSU VFRM Setup to Rising VCLK (VFRAME Input) 3 ns
VFRMHD VFRM Hold from Rising VCLK (VFRAME Input) 4 ns
VFRM
TD
VSTRB
SU
VSTRBHD VSTRB Hold from Rising VCLK 3 ns
VCLK to VRDY Valid Delay 12 ns
VCLK to VFRM Valid Delay (VFRAME Output) 12 ns
VSTRB Setup to Rising VCLK 4 ns
VCLK
VDATA
HD
VDATA
PIXEL
DATA(IN)
VFRM(IN)
N–1N012
VFRM
SU
VFRM
HD
SU
VRDY
VSTRB
VCLK
PIXEL
DATA
FRM(OUT)
VRDY
TD
VSTRB
HD
VSTRB
SU
VDATA
TD
NN012
VRFM
TD
Figure 22. Raw Pixel Mode Timing
04723-031
Rev. B | Page 17 of 40
Page 18
ADV202
SPI PORT TIMING
Table 13.
Parameter Description Min Typ Max Unit
SCLK
FALL
SCLK
RIS
SCLK_hi SCLK high time 75 ns
SCLK_lo SCLK Low Time 75 ns
Data_su Data Setup Time 6.5 ns
Data_hd Data Hold Time 6.5 ns
CSEL_SU Active Setup Time 135 ns
CSEL_HD Active Hold Time 155 ns
DV_SCLK SCLK to Output Data Valid 2 ns
DV_CS
SCLK SCLK Period 150 ns
S_CLK Fall Time 5 ns
S_CLK Rise Time 5 ns
CS to Output Data Valid
36 ns
S_CLK
S_MO
S_MI
S_CSEL
SCLK_HI
CSEL
SCLK_LO
MSBLSB
DV_SCLK
MSBLSB
SU
DC_CS
SCLK
RISE
SCLK
FALL
DATA
SU
DATA
HD
CSEL
HD
04723-032
Figure 23. SPI Port—Input Timing
Rev. B | Page 18 of 40
Page 19
ADV202
PIN BGA ASSIGNMENTS AND FUNCTION DESCRIPTIONS
PIN BGA ASSIGNMENTS
Table 14. Pin BGA Assignments for 121-Lead Package
1 J7 H10 I Write Enable Used with the Host Interface.
Read Enable when Fly-By DMA Is Enabled.
1 H9 G12 I Read Enable Used with the Host Interface.
Write Enable when Fly-By DMA Is Enabled.
1 H8 G11 O
1 G10 G10 O
1 F8 F12 O
O
O
1 F9 F11 I
121-Pin
Package
D4–D1, C5–
C3, B5, B4, C2,
B3–B1, A2,
A6–A5
H11, K8, H10,
J9
144-Pin
Package
F4, E1–E3,
D1–D3, C1–
C3, B1–B3, A2,
A3, A4
J12, J11, J10,
H12
I/O Description
System Input Clock. For details, see the PLL section. Maximum input
frequency on MCLK is 74.25 MHz.
Reset. Causes the ADV202 to immediately reset.
DACK1, DREQ0, and DREQ1 must be held high when a RESET is
applied.
I/O
Host Data Bus. With HDATA<23:16>, <27:24>, <31:28>, these pins
make up the 32-bit wide host data bus. The async host interface is
interfaced together with ADDR<3:0>,
Unused HDATA pins should be pulled down via a 10 kΩ resistor.
Address Bus for the Host Interface.
I
Chip Select.This signal is used to qualify addressed read and write
access to the ADV202 using the host interface.
Note: Simultaneous assertion of
HDATA bus, even if the DMA channels are disabled.
Note: Simultaneous assertion of
HDATA bus, even if the DMA channels are disabled.
Acknowledge. Used for direct register accesses. This signal indicates
that the last register access was successful.
Note: Due to synchronization issues, control and status register
accesses might incur an additional delay, so the host software should
wait for acknowledgment from the ADV202.
Accesses to the FIFOs (external DMA modes), on the other hand, are
guaranteed to occur immediately, provided that space is available,
and should not wait for
are observed.
If
ACK is shared with more than one device, ACK should be connected
to a pull-up resistor (10 kΩ) and the PLL_HI register, Bit 4, must be set
to 1.
Interrupt. This pin indicates that the ADV202 requires the attention of
the host processor. This pin can be programmed to indicate the status
of the internal interrupt conditions within the ADV202. The interrupt
sources are enabled via bits in register EIRQIE.
Data Request for external DMA Interface. Indicates that the ADV202
is ready to send/receive data to/from the FIFO assigned to DMA
Channel 0.
Used in DCS-DMA Mode. Service request from the FIFO assigned to
Channel 0 (asynchronous mode).
Valid Indication for JDATA Input/Output Stream. Polarity of this pin is
programmable in the EDMOD0 register.
Boot Mode Configuration. This pin is read on reset to determine the
boot configuration of the on-board processor. The pin should be tied
to IOVDD or DGND through a 10 kΩ resistor.
Data Acknowledge for External DMA Interface. Signal from the host
CPU, which indicates that the data transfer request (
acknowledged and data transfer can proceed. This pin must be held
high at all times, if the DMA interface is not used, even if the DMA
channels are disabled.
CS, RD, WE, DACK0,
CS, WE, RD, and ACK.
WE and DACK low activates the
RD and DACK low activates the
ACK, provided that the timing constraints
VALID is always an output.
DREQ0) has been
Rev. B | Page 22 of 40
Page 23
ADV202
Pins
Mnemonic
HOLD
FCS0
DREQ1
FSRQ1
CFG<2> I
DACK1
FCS1
HDATA<31:28> 4 J2–J4, H1 K3, J1–J3 I/O Host Expansion Bus.
JDATA<7:4> I/O JDATA Bus (JDATA Mode).
HDATA<27:24> 4 H2–H4, G4 J4, H1–H3 I/O Host Expansion Bus.
JDATA<3:0> I/O JDATA Bus (JDATA Mode).
VDATA<23:20> I/O Video Data Expansion Bus.
HDATA<23:16> 8
VDATA<19:12> I/O
SCOMM<7> 8 L2 M2 I/O When not used, this pin should be tied low.
SCOMM<6> L3 M3 I/O When not used, this pin should be tied low.
SCOMM<5> L4 M4 I/O
External Hold Indication for JDATA Input/Output Stream. Polarity is
programmable in the EDMOD0 register. This pin is always an input.
Used in DCS-DMA Mode. Chip select for the FIFO assigned to
Channel 0 (asynchronous mode).
Data Request for External DMA Interface. Indicates that the ADV202
is ready to send/receive data to/from the FIFO assigned to DMA
Channel 1.
Used in DCS-DMA Mode. Service request from the FIFO assigned to
Channel 1 (asynchronous mode).
Boot Mode Configuration. This pin is read on reset to determine the
boot configuration of the on-board processor. The pin should be tied
to IOVDD or DGND through a 10 kΩ resistor.
Data Acknowledge for External DMA Interface. Signal from the host
CPU, which indicates that the data transfer request (
acknowledged and data transfer can proceed. This pin must be held
high at all times unless a DMA or JDATA access is occurring. This pin
must be held high at all times, if the DMA interface is not used, even if
the DMA channels are disabled.
Used in DCS-DMA Mode. Chip select for the FIFO assigned to
Channel 1 (asynchronous mode).
H4, G1–G4,
F1–F3
D10–D12,
C10–C12,
B10–B12,
A9–A11
I/O Host Expansion Bus.
Video Data Expansion Bus. Extended pixel interface mode. Used for
video formats that use Y and CrCb on separate buses.
This pin must be used in multiple chip mode to align the outputs of
two or more ADV202s. For details, see the Applications section and
the ADV202 Multichip Application application note. When not used,
this pin should be tied low.
LCODE Output in Encode Mode. When LCODE is enabled, the output
on this pin indicates on a high transition that the last data-word for a
field has been read from the FIFO. For an 8-bit interface, such as
JDATA, LCODE is asserted for four consecutive bytes and is enabled
by default.
SPI interface: S_CSEL. When not used, this pin should be tied low.
Used only with boot mode 6.
SPI interface: S_MO. When not used, this pin should be tied low.
Used only with boot mode 6.
SPI interface: S_MI. When not used, this pin should be tied low.
Used only with boot mode 6.
SPI interface: S_CLK. When not used, this pin should be tied low.
Used only with boot mode 6.
Video Data Clock. Must be supplied, if video data is input/output on
the VDATA bus.
I/O Video Data. Unused pins should be pulled down via a 10 kΩ resistor.
DREQ1) has been
Rev. B | Page 23 of 40
Page 24
ADV202
Pins
Mnemonic
VSYNC 1 D8 E10 I/O Vertical Sync for Video Mode.
VFRM
HSYNC 1 D9 E11 I/O Horizontal Sync for Video Mode.
VRDY O Raw Pixel Mode Ready Signal.
FIELD 1 E10 E9 I/O Field Sync for Video Mode.
VSTRB I Raw Pixel Mode Transfer Strobe.
TEST1 1 J6 K12 I This pin should be connected to ground via a pull-down resistor.
TEST2 1 K9 K11 I This pin should be connected to ground via a pull-down resistor.
TEST3 1 J10 K10 I This pin should be connected to ground via a pull-down resistor.
TEST4 1 L6 M9 I This pin should be connected to ground via a pull-down resistor.
TEST5 1 K10 L10 O No connect.
VDD A3, A8, D7, H7
The input video or pixel data is passed on to the ADV202’s pixel
interface, where samples are de-interleaved and passed on to the
wavelet engine, where each tile or frame is decomposed into
subbands using the 5/3 or 9/7 filters. The resultant wavelet
coefficients are then written to internal memory. The entropy
codecs then code the image data so that it conforms to the
JPEG2000 standard. An internal DMA provides high bandwidth
memory-to-memory transfers, as well as high performance
transfers between functional blocks and memory.
WAVELET ENGINE
The ADV202 provides a dedicated wavelet transform processor
based on the Analog Devices proven and patented SURF™
technology. This processor can perform up to six wavelet
decomposition levels on a tile. In encode mode, the wavelet
transform processor takes in uncompressed samples, performs
the wavelet transform and quantization, and writes the wavelet
coefficients in all frequency subbands to internal memory. Each
of these subbands is then further broken down into code blocks.
The code-block dimensions can be user-defined, and are used
by the wavelet transform processor to organize the wavelet
coefficients into code blocks when writing to internal memory.
Each completed code block is then entropy coded by one of the
entropy codecs.
In decode mode, wavelet coefficients are read from internal
memory and recomposed into uncompressed samples.
ENTROPY CODECS
The entropy codec block performs context modeling and
arithmetic coding on a code block of the wavelet coefficients.
Additionally, this block also performs the distortion metric
calculations during compression that are required for optimal
rate and distortion performance. Because the entropy coding
process is the most computationally intensive operation in the
JPEG2000 compression process, three dedicated hardware
entropy codecs are provided on the ADV202.
EMBEDDED PROCESSOR SYSTEM
The ADV202 incorporates an embedded 32-bit RISC processor.
This processor is used for configuration, control, and management of the dedicated hardware functions, as well as for parsing
and generation of the JPEG2000 code stream. The processor
system includes ROM and RAM for both program and data
memory, an interrupt controller, standard bus interfaces, and
other hardware functions such as timers and counters.
MEMORY SYSTEM
The memory system’s main function is to manage wavelet
coefficient data, interim code-block attribute data, and
temporary work space for creating, parsing, and storing the
JPEG2000 code stream. The memory system can also be used
for program and data memory for the embedded processor.
INTERNAL DMA ENGINE
The internal DMA engine provides high bandwidth memoryto-memory transfers, as well as high performance transfers
between memory and functional blocks. This function is critical
for high speed generation and parsing of the code stream.
Rev. B | Page 25 of 40
Page 26
ADV202
ADV202 INTERFACE
There are several possible modes to interface to the ADV202
using the VDATA bus and the HDATA bus or the HDATA bus
alone.
VIDEO INTERFACE (VDATA BUS)
The video interface can be used in applications in which
uncompressed pixel data is on a separate bus from compressed
data. For example, it is possible to use the VDATA bus to input
uncompressed video while using the HDATA bus to output the
compressed data. This interface is ideal for applications
requiring very high throughput such as live video capture.
Optionally, the ADV202 interlaces ITU.R-BT656 resolution
video on the fly prior to wavelet processing, which yields
significantly better compression performance for temporally
coherent frame-based video sources. Additionally, high
definition digital video such as SMPTE274M (1080i) is
supported using two or more ADV202 devices.
The video interface can support video data or still image data
input/output, 8-, 10-, and 12-bit single or multiplexed
components, and dual-lane 8-, 10-, and 12-bit components. The
VDATA interface supports digital video in YCbCr format in
single input mode or Y and CbCr in dual-lane input mode.
YCbCr data must be in 4:2:2 format.
Video data can be input/output in several different modes on
the VDATA bus, as described in Table 17. In all these modes, the
pixel clock must be input on the VCLK pin.
Table 17. Video Input/Output Modes
Mode Description
EAV/SAV
HVF
Extended
Raw video
HDTV
HOST INTERFACE (HDATA BUS)
The ADV202 can connect directly to a wide variety of host
processors and ASICs using an asynchronous SRAM-style
interface, DMA accesses, or streaming mode (JDATA) interface.
The ADV202 supports 16- and 32-bit buses for control and
8-, 16-, and 32-bit buses for data transfer.
Accepts video with embedded EAV/SAV codes,
where the YCbCr data is interleaved onto a single
bus.
Accepts video data accompanied with separate H,
V, and F signals where YCbCr data is interleaved
onto a single bus.
Y and CrCb are on separate buses accompanied by
EAV/SAV codes.
Used for still picture data and nonstandard video.
VFRM, VSTRB, and VRDY are used to program the
dimensions of the image.
For applications in which video data is clocked into
the part at higher rates than 27 MHz.
The control and data channel bus widths can be specified
independently, which allows the ADV202 to support
applications that require control and data buses of different
widths.
The host interface is used for configuration, control, and status
functions, as well as for transferring compressed data streams. It
can be used for uncompressed data transfers in certain modes.
The host interface can be shared by as many as four concurrent
data streams in addition to control and status communications.
The data streams are
• Uncompressed tile data (for example, still image data)
The ADV202 uses big endian byte alignment for 16- and 32-bit
transfers. All data is left-justified (MSB).
Pixel Input on the Host Interface
Pixel input on the host interface supports 8-, 10-, 12-, 14-, and
16-bit raw pixel data formats. It can be used for pixel (still
image) input/output or compressed video output. Because there
are no timing codes or sync signals associated with the input
data on the host interface, dimension registers and internal
counters are used and must be programmed to indicate the start
and end of the frame. See the ADV202 in HIPI Mode technical
note for details on how to use the ADV202 in this mode.
Host Bus Configuration
For maximum flexibility, the host interface provides several
configurations to meet particular system requirements. The
default bus mode uses the same pins to transfer control, status,
and data to and from the ADV202. In this mode, the ADV202
can support 16- and 32-bit control transfers and 8-, 16-, and
32-bit data transfers. The size of these buses can be selected
independently, allowing, for example, a 16-bit microcontroller
to configure and control the ADV202 while still providing
32-bit data transfers to an ASIC or external memory system.
DIRECT AND INDIRECT REGISTERS
To minimize pin count and cost, the number of address pins has
been limited to four, which yields a total direct address space of
16 locations. These locations are most commonly used by the
external controller and are, therefore, accessible directly. All
other registers in the ADV202 can be accessed indirectly
through the IADDR and IDATA registers.
Rev. B | Page 26 of 40
Page 27
ADV202
CONTROL ACCESS REGISTERS
With the exception of the indirect address and data registers
(IADDR and IDATA), all control/status registers in the ADV202
are 16 bits wide and are half-word (16-bit) addressable only.
When 32-bit host mode is enabled, the upper 16 bits of the
HDATA bus are ignored on writes and return all zeros on reads
of 16-bit registers.
PIN CONFIGURATION AND BUS SIZES/MODES
The ADV202 provides a wide variety of control and data
configurations, which allows it to be used in many applications
with little or no glue logic. The following modes are configured
using the BUSMODE register. In the following descriptions, host
CS/RD
refers to normal addressed accesses (
data refers to external DMA accesses (
32-Bit Host/32-Bit Data
In this mode, the HDATA<31:0> pins provide full 32-bit wide
data access to PIXEL, CODE, ATTR, and ANCL FIFOs. The
expanded video interface (VDATA) is not available in this
mode.
16-Bit Host/32-Bit Data
This mode allows a 16-bit host to configure and communicate
with the ADV202 while still allowing 32-bit accesses to the
PIXEL, CODE, ATTR, and ANCL FIFOs u sing the external
DMA capability.
All addressed host accesses are 16 bits and, therefore, use only
the HDATA<15:0> pins. The HDATA<31:16> pins provide the
additional 16 bits necessary to support the 32-bit external DMA
transfers to and from the FIFOs only. The expanded video
interface (VDATA) is not available in this mode.
16-Bit Host/16-Bit Data
This mode uses 16-bit transfers, if used for host or external
DMA data transfers. This mode allows for the use of the
extended pixel interface modes.
16-Bit Host/8-Bit Data (JDATA Bus Mode)
This mode provides separate data input/output and host control
interface pins. Host control accesses are 16 bits and use
HDATA<15:0>, while the dedicated data bus uses JDATA<7:0>.
/WR/ADDR) and
DREQ/DACK
).
has been provided to allow 16-bit hosts to access these registers
and memory locations using the stage register (STAGE). STAGE
is accessed as a 16-bit register using HDATA[15:0]. Prior to
writing to the desired register, the stage register must be written
with the upper (most significant) half-word.
When the host subsequently writes the lower half-word to the
desired control register, HDATA is combined with the
previously staged value to create the required 32-bit value that is
written. When a register is read, the upper (most significant)
half-word is returned immediately on HDATA and the lower
half-word can be retrieved by reading the stage register on a
subsequent access. For details on using the stage register, see the
ADV202 User’s Guide.
Note: The stage register does not apply to the four data channels
(PIXEL, CODE, ATTR , or ANC L). These channels are always
accessed at the specified data width and do not require the use
of the stage register.
JDATA MODE
JDATA mode is typically used only when the dedicated video
interface (VDATA) is also enabled. This mode allows code
stream data (compressed data compliant with JPEG2000) to be
input or output on a single dedicated 8-bit bus (JDATA<7:0>).
The bus is always an output during compression operations, and
is an input during decompression.
A 2-pin handshake is used to transfer data over this
synchronous interface. VALID is used to indicate that the
ADV202 is ready to provide or accept data and is always an
output. HOLD is always an input and is asserted by the host if it
cannot accept/provide data. For example, JDATA mode allows
real-time applications, in which pixel data is input over the
VDATA bus while the compressed data stream is output over
the JDATA bus.
EXTERNAL DMA ENGINE
The external DMA interface is provided to enable high
bandwidth data I/O between an external DMA controller and
the ADV202 data FIFOs. Two independent DMA channels can
each be assigned to any one of the four data stream FIFOs
(PIXEL, CODE, ATTR , or ANC L).
JDATA uses a valid/hold synchronous transfer protocol. The
direction of the JDATA bus is determined by the mode of the
ADV202. If the ADV202 is encoding (compression), then
JDATA<7:0> is an output. If the ADV202 is decoding
(decompression), then JDATA<7:0> is an input. Host control
accesses remain asynchronous. See also JDATA section below.
STAGE REGISTER
Because the ADV202 contains both 16-bit and 32-bit registers
and its internal memory is mapped as 32-bit data, a mechanism
Rev. B | Page 27 of 40
The controller supports asynchronous DMA using a
Data-Request/Data-Acknowledge (
either single or burst access modes. Additional functionality is
provided for single address compatibility (fly-by) and dedicated
chip select (DCS) modes.
DREQ/DACK
) protocol in
SPI PORT
The SPI port provides serial communication to and from the
ADV202. The ADV202 is always the SPI master.
Page 28
ADV202
INTERNAL REGISTERS
This section describes the internal registers of the ADV202.
DIRECT REGISTERS
The ADV202 has 16 direct registers, as listed in Table 18. The
direct registers are accessed over the ADDR [3–0],
HDATA[31…0],
The host must first initialize the direct registers before any
application-specific operation can be implemented.
For additional information on accessing and configuring these
registers, see the ADV202 User’s Guide.
Rev. B | Page 28 of 40
Page 29
ADV202
INDIRECT REGISTERS
The indirect registers, listed in Table 19, are accessed by both
the host system and the internal 32-bit embedded processor, via
the ESF or the firmware.
Both 32-bit and 16-bit hosts can access the indirect registers.
32-bit hosts use the IADDR and IDATA registers, while the
16 bit hosts use IADDR, IDATA, and the stage register.
In certain modes, such as custom-specific input format or HIPI
mode, indirect registers must be accessed by the user through
the use of the IADDR and IDATA registers. The indirect
For additional information on accessing and configuring these
registers, see the ADV202 User’s Guide.
register address space starts at Internal Address 0xFFFF0000.
Table 19. Indirect Registers
Address Name Description
0xFFFF0400 PMODE1 Pixel/Video Format
0xFFFF0404 COMP_CNT_STATUS Horizontal Count
0xFFFF0408 LINE_CNT_STATUS Vertical Count
0xFFFF040C XTOT Total Samples per Line
0xFFFF0410 YTOT Total Lines per Frame
0xFFFF0414 F0_START Start Line of Field 0 [F0]
0xFFFF0418 F1_START Start Line of Field 1 [F1]
0xFFFF041C V0_START Start of Active Video Field 0 [F0]
0xFFFF0420 V1_START Start of Active Video Field 1 [F1]
0xFFFF0424 V0_END End of Active Video Field 0 [F0]
0xFFFF0428 V1_END End of Active Video Field 1 [F1]
0xFFFF042C PIXEL_START Horizontal Start of Active Video
0xFFFF0430 PIXEL_END Horizontal End of Active Video
0xFFFF0440 MS_CNT_DEL Master/Slave Delay
0xFFFF0444 LINE_CNT_INTERRUPT Line Count Interrupt
0xFFFF0448 PMODE2 Pixel Mode 2
0xFFFF044C VMODE Video Mode
0xFFFF1408 EDMOD0 External DMA Mode Register 0
0xFFFF140C EDMOD1 External DMA Mode Register 1
0xFFFF1410 FFTHRP FIFO Threshold for Pixel FIFO
0xFFFF1414 FFCNTP FIFO Full/Empty Count for Pixel FIFO
0xFFFF1418 FFMODE FIFO Mode Register
0xFFFF141C FFTHRC FIFO Threshold for Code FIFO
0xFFFF1420 FFTHRA FIFO Threshold for ATTR FIFO
0xFFFF1424 FFTHRN FIFO Threshold for ANCL FIFO
0xFFFF1428 FFCNTC FIFO Full/ Empty Count for CODE FIFO
0xFFFF142C FFCNTA FIFO Full/Empty Count for ATTR FIFO
0xFFFF1430 FFCNTN FIFO Full/Empty Count for ANCL FIFO
0xFFFF1434 to 0xFFFF14FC Reserved Reserved
Rev. B | Page 29 of 40
Page 30
ADV202
PLL
The ADV202 uses the PLL_HI and PLL_LO direct registers to
configure the PLL. Any time the PLL_LO register is modified,
the host must wait at least 20 µs before reading or writing any
other register. If this delay is not implemented, erratic behavior
might result.
• The maximum burst frequency for external DMA modes is
≤ 0.36 JCLK.
• For MCLK frequencies greater than 50 MHz, the input clock
divider must be enabled, that is, IPD set to 1.
The PLL can be programmed to have any possible final
multiplier value as long as
• JCLK > 50 MHz and < 150 MHz (144-pin version).
• IPD cannot be enabled for MCLK frequencies below 20 MHz.
To achieve the lowest power consumption, an MCLK frequency
of 27 MHz is recommended for a standard definition CCIR656
input. The PLL circuit is recommended to have a multiplier of 3.
• JCLK > 50 MHz and < 135 MHz (144-pin version).
• JCLK > 50 MHz and < 115 MHz (121-pin version).
• HCLK < 115 MHz.
This sets JCLK and HCLK to 81 MHz.
MCLK
IPD
÷
2
PHASE
DETECT
BYPASS
LPF
VCO
JCLK
• JCLK ≥ 2 × VCLK for single-component input.
HCLK
• JCLK ≥ 2 × VCLK for YCrCb [4:2:2] input.
• In JDATA mode (JDATA), JCLK must be 4 × MCLK or
higher.
÷
÷
LFB
Figure 24. PLL Architecture and Control Functions
PLLMULT
2
÷2
HCLKD
04723-009
Table 20. Recommended PLL Register Settings
IPD LFB PLLMULT HCLKD HCLK JCLK
0 0 N 0 N × MCLK N × MCLK
0 0 N 1 N × MCLK/2 N × MCLK
0 1 N 0 2 × N × MCLK 2 × N × MCLK
0 1 N 1 N × MCLK 2 × N × MCLK
1 0 N 0 N × MCLK/2 N × MCLK/2
1 0 N 1 N × MCLK/4 N × MCLK/2
1 1 N 0 N × MCLK N × MCLK
1 1 N 1 N × MCLK/2 N × MCLK
Table 21. Recommended Values for PLL_HI and PLL_LO Registers
Video Standard CLKIN Frequency on MCLK PLL_HI PLL_LO
The boot mode can be configured via hardware using the CFG
pins or via software (see the ADV202 User’s Guide). The first
boot mode after power-up is set by the CFG pins.
Table 22. Hardware Boot Modes
Boot Mode Settings Description
Hardware Boot
Mode 2
Hardware Boot
Mode 4
Hardware Boot
Mode 6
CFG<1> tied high,
CFG<2> tied low
CFG<1> tied low,
CFG<2> tied high
CFG<1> and <2>
tied high
No-Boot Host Mode. ADV202 does not boot, but all internal registers and memory are accessible
through normal host I/O operations.
For details, see the ADV202 User’s Guide and the Getting Started with the ADV202 application note.
SoC boot mode. The embedded software framework (ESF) takes control and establishes
communications with the host.
SPI boot mode. Boot firmware over SPI from external flash memory.
Only boot modes 2, 4, and 6, described in Table 22, are available
via hardware.
Rev. B | Page 31 of 40
Page 32
ADV202
VIDEO INPUT FORMATS
The ADV202 supports a wide variety of formats for
uncompressed video and still image data. The actual interface
and bus modes selected for transferring uncompressed data
dictates the allowed size of the input data and the number of
samples transferred with each access.
The host interface can support 8-, 10-, 12-, 14-, and 16-bit data
formats. The video interface can support video data or still
image data input/output. Supported formats are 8-, 10-, 12-, or
16-bit single or 2 × 8-bit, 2 × 10-bit, 2 × 12-bit multiplexed
Table 23. Maximum Pixel Data Input Rates
Input Rate Limit
Interface
Compression
Mode
Input Format
Active Resolution
1
(MSPS)
144-PIN PACKAGE
HDATA Irreversible 8-bit data 45 [40] 130 200
Irreversible 10-bit data 45 [40] 130 200
Irreversible 12-bit data 45 [40] 130 200
Irreversible 16-bit data 45 [40] 130 200
Reversible 8-bit data 40 [36] 130 200
Reversible 10-bit data 32 [28] 130 200
Reversible 12-bit data 27 [24] 130 200
Reversible 14-bit data 23 [20] 130 200
VDATA Irreversible 8-bit data 65 [55] 130 200
Irreversible 10-bit data 65 [55] 130 200
Irreversible 12-bit data 65 [55] 130 200
Reversible 8-bit data 40 [34] 130 200
Reversible 10-bit data 32 [28] 130 200
Reversible 12-bit data 27 [23] 130 200
121-PIN PACKAGE
HDATA Irreversible 8-bit data 34 98 150
Irreversible 10-bit data 34 98 150
Irreversible 12-bit data 34 98 150
Irreversible 16-bit data 34 98 150
Reversible 8-bit data 30 98 150
Reversible 10-bit data 24 98 150
Reversible 12-bit data 20 98 150
Reversible 14-bit data 17 98 150
VDATA Irreversible 8-bit data 48 98 150
Irreversible 10-bit data 48 98 150
Irreversible 12-bit data 48 98 150
Reversible 8-bit data 30 98 150
Reversible 10-bit data 24 98 150
Reversible 12-bit data 20 98 150
1
Input rate limits for HDATA might be less for certain applications depending on input picture size and content, host interface settings, and DMA transfer settings.
Values in brackets refer to the 135 MHz speed grade version of the ADV202.
2
Minimum peak output rate or guaranteed sustained output rate.
3
Maximum output rate, or output rate above this value is not possible.
formats. See the ADV202 User’s Guide for details. All formats
can support less precision than provided by specifying the
actual data width/precision in the PMODE register.
The maximum allowable data input rate is limited by using
irreversible or reversible compression modes and the data width
(or precision) of the input samples. Use Table 23 and Table 24 to
determine the maximum data input rate.
Approx Min Peak Output
Rate, Compressed Data
(Mbps)
Approx Max Output Rate,
2
Compressed Data
(Mbps)
3
Rev. B | Page 32 of 40
Page 33
ADV202
Table 24. Maximum Supported Tile Width for Data Input on HDATA and VDATA Buses
Compression Mode Input Format Tile/Precinct Maximum Width
This section describes typical video applications for the
ADV202 JPEG2000 video processor.
ENCODE—MULTICHIP MODE
Due to the data input rate limitation (see Table 23), an 1080i
application requires at least two ADV202s to encode or decode
full-resolution 1080i video. In encode mode, the ADV202
accepts Y and CbCr data on separate buses. The input data must
be in EAV/SAV format. An encode example is shown in
Figure 25.
In decode mode, a master/slave configuration (as shown in
Figure 26) or a slave/slave configuration can be used to
synchronize the outputs of the two ADV202s. See the ADV202 Multichip Application application note for details on how to
configure the ADV202s in a multichip application.
Applications that have two separate VDATA outputs sent to an
FPGA or buffer before they are sent to an encoder do not
require synchronization at the ADV202 outputs.
32-BIT HOST CPU
DATA[31:0]HDATA[31:0]
ADDR[3:0]ADDR[3:0]
CSCS
RDRD
WRWE
ACKACK
IRQ
DREQDREQ
DACKDACK
G I/OSCOMM[5]
CS
RD
WR
ACK
IRQ
DREQ
DACK
ADV202
_1_SLAVE
VCLK
IRQ
ADV202
_2_SLAVE
HDATA[31:0]
ADDR[3:0]
CS
RD
WE
ACK
IRQ
DREQ
DACK
SCOMM[5]
MCLK
VDATA[11:2]
FIELD
VSYNC
HSYNC
MCLK
HSYNC
VSYNC
FIELD
VDATA[11:2]
Y
CbCr
VCLK
CbCr
Figure 25. Encode—Multichip Application
ADV7402
10-BIT SD/HD
VIDEO
DECODER
LLC
Y[9:0]
C[9:0]
1080i
VIDEO OUT
04723-002
Rev. B | Page 34 of 40
Page 35
ADV202
DECODE—MULTICHIP MASTER/SLAVE
In a master/slave configuration, it is expected that the master
HVF outputs are connected to the slave HVF inputs and that
each SCOMM[5] pin is connected to the same GPIO on the
host.
In a slave/slave configuration, the common HVF for both
ADV202s is generated by an external house sync and each
SCOMM[5] is connected to the same GPIO output on the host.
SWIRQ1, Software Interrupt 1 in the EIRQIE register, must be
unmasked on both devices to enable multichip mode.
ADV202BBC-115 –40°C to +85°C 115 MHz 1.5 V internal, 2.5 V or 3.3 V I/O 121-Lead CSPBGA BC-121
ADV202BBCZ-1151–40°C to +85°C 115 MHz 1.5 V internal, 2.5 V or 3.3 V I/O 121-Lead CSPBGA BC-121
ADV202BBC-135 –40°C to +85°C 135 MHz 1.5 V internal, 2.5 V or 3.3 V I/O 144-Lead CSPBGA BC-144-3
ADV202BBCZ-1351 –40°C to +85°C 135 MHz 1.5 V internal, 2.5 V or 3.3 V I/O 144-Lead CSPBGA BC-144-3
ADV202BBC-150 –40°C to +85°C 150 MHz 1.5 V internal, 2.5 V or 3.3 V I/O 144-Lead CSPBGA BC-144-3
ADV202BBCZ-1501 –40°C to +85°C 150 MHz 1.5 V internal, 2.5 V or 3.3 V I/O 144-Lead CSPBGA BC-144-3
ADV202-HD-EB High Definition Evaluation Board
ADV202-SD-EB