Objective specification
File under Integrated Circuits, IC02
1997 Jan 29
Page 2
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
FEATURES
General
• Uses single external Synchronous DRAM (SDRAM)
organized as 1M × 16 interfacing at 81 MHz; compatible
with the SDRAM ‘lite’ or ‘PC’
• Fast external CPU interface; 16-bit data + 8-bit address
• Dedicated input for audio and video data in PES or ESformat; data input rate: ≤9 Mbytes/s in byte mode;
≤20 Mbit/s in bit serial mode; audio and/or video data
can also serve as input via CPU interface
• Single 27 MHz external clock for time base reference
and internal processing; all required decoding and
presentation clocks are generated internally
• Internal system time base at 90 kHz can be
synchronized via CPU port
• Flexible memory allocation under control of the external
CPU enables optimized partitioning of memory for
different tasks
• Boundary scan (JTAG) plus external SDRAM self test
implemented
• Supply voltage 3.3 V
• Package 160 QFP.
CPU relation
• 16-bit data, 8-bit address, or 16-bit multiplexed bus;
Motorola and Intel mode supported
• Support for fast DMA transfer to either internal registers
or external SDRAM
• Maximum sustained rate to the external SDRAM is
9 Mbytes/s.
MPEG2 audio
• Decoding of 2 channel, layer I and II MPEG audio;
support for mono, stereo, intensity stereo and dual
channel mode
• Constant and variable bit rates up to 448 kbit/s
• Audio sampling frequencies: 48, 44.1, 32, 24, 22.05 and
16 kHz
• CRC error detection
• Selectable output channel in dual channel mode
• Independent volume control for both channels and
programmable inter-channel crosstalk control through a
baseband audio processing unit
• Storage ancillary data up to 54 bytes
• Dynamic range control at output
• Muting possibility via external controller; automatic
muting in case of errors
• Generation of ‘beeps’ with programmable tone height,
duration and amplitude
• Serial two channel digital audio output with 16, 18, 20 or
22 bits per sample, compatible with either I2S or
Japanese formats
• Serial SPDIF audio output
• Clock output 256 or 384 × fs for external D/A converter
• Audio input buffer in external SDRAM with
programmable size (default is 64 kbit)
• Programmable processing delay compensation
• Software controlled stop, pause, restricted skip, and
restart functions.
MPEG2 video
MPEG2 system
• Parsing of MPEG2 PES and MPEG1 packet streams
• Double System Time Clock (STC) counters for
discontinuity handling
• Time stamps or CPU controlled audio/video
synchronization
• Support for seamless time base change (edition)
• Processing of errors flagged by channel decoding or
demux section
• Support for retrieval of PES header and PES private
data.
1997 Jan 292
• Decoding of MPEG2 video up to main level, main profile
• Nominal video input buffer size equals 2.6 Mbit for Video
Main Profile and Main Level (MP@ML)
• Output picture format: CCIR-601 4 : 2 : 2 interlaced
pictures; picture format 720 × 576 at 50 Hz or 720 × 480
at 60 Hz
• 3 : 2 pull-down supported with 24 and 30 Hz sequences
• Support of constant and variable bit rates up to 15 Mbit/s
• Output interface at 8-bit wide, 27 MHz UYVY
multiplexed bus
• Horizontal and vertical pan and scan allows the
extraction of a window from the coded picture
Page 3
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
• Flexible horizontal continuous scaling from 0.5 up to 4
allows easy aspect ratio conversion including support
for 2.21 : 1 aspect ratio movies
• Vertical scaling with fixed factors 0.5, 1 or 2 to support
picture scaling and up-sampling
• Scaling of incoming pictures to 25% of their original size
with anti-aliasing filtering to free screen space for
graphics applications like electronic program guides
• Non-full screen MPEG pictures will be displayed in a box
of which position and background colour are adjustable
by the external CPU
• Video output may be slaved to internally (master)
generated or externally (slave) supplied HV
synchronization signals; the position of active video is
programmable; MPEG timebase changes do not
affected the display phase
• Video output direct connectable to SAA718X encoder
family
• Various trick modes under control of external CPU:
– Freeze I or P pictures; restart on I picture
– Freeze on B pictures; restart at any moment
– Scanning and decoding of I or I and P pictures
– Single step mode
– Repeat/Skip field for time base correction.
Graphics
• Graphics is region based and presented in boxes
independent of video format
• Screen arrangement of boxes is determined by display
list mechanism which allows for multiple boxes,
background loading, fast switching, scrolling and fading
of regions
• Support of 2, 4, 8 bits/pixel bit-maps in fixed bit-maps or
coded in accordance to the DVB variable/run length
standard for region bases graphics
• Optimized memory control in MPEG video decoding
allows for storage of graphical bit-maps up to 1.2 Mbit in
50 Hz and 2.0 Mbit in 60 Hz systems
• VL/RL encoding enables full screen graphics at
8 bit/pixel in 50 Hz
• Fast CPU access enables full bit-map updates within a
display field period
• Display colours are obtained via colour look-up tables;
CLUT output is YUVT at 8-bit for each signal component
thus enabling 16M different colours and 6-bit for T
(transparency) which gives 64 mixing levels with video
• Bit-map table mechanism to specify a sub-set of entries
if the CLUT is larger than required by the coded bit
pattern; supported bit-map tables are 16 to 256,
4 to 256 and 4 to 16
• Graphics boxes may not overlap vertically; if 256 entry
CLUT has to be down loaded, a vertical separation of
1 field line is mandatory
• Internal support for fast block moves in the external
SDRAM during MPEG decoding
• Graphics mechanism can be used for signal generation
in the vertical blanking interval; useful for teletext, wide
screen signalling, closed caption etc.
• Support for a single down-loadable cursor of 1 kpixel
with programmable shape; supported shapes are
8 × 128, 16 × 64, 32 × 32, 64 × 16 and 128 × 8
• Cursor colours are determined via a 4-entry CLUT with
YUVT at 6, 4, 4 respectively 2 bits; mixing of cursor with
video + graphics in 4 levels
• Cursor can be moved freely across the screen without
overlapping restrictions.
1997 Jan 293
Page 4
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
GENERAL DESCRIPTION
The SAA7201 is an MPEG2 decoder which combines
audio decoding and video decoding. Additionally to these
basic MPEG functions it also provides means for
Due to an optimized architecture for audio and video
decoding, maximum capacity in the external memory and
processing power from the external CPU is available for
the support for graphics.
enhanced graphics and/or on-screen display.
QUICK REFERENCE DATA
SYMBOLPARAMETERMIN.TYP.MAX.UNIT
V
DD
V
CC
I
DD(tot)
f
CLK
∆f
CLK
functional supply voltage3.03.33.6V
pad supply voltage3.03.33.6V
total supply current at VDD= 3.3 V−tbf−mA
clock frequency−27.0−MHz
frequency deviation−30 × 10−6−+30 × 10
SDRAM_ADDR886memory address3.3O
SDRAM_ADDR987memory address3.3O
SDRAM_ADDR11 88memory address3.3O
V
SS10
SDRAM_ADDR790memory address3.3O
SDRAM_ADDR10 91memory address3.3O
SDRAM_ADDR692memory address3.3O
V
DD10
SDRAM_ADDR094memory address3.3O
SDRAM_ADDR595memory address3.3O
SDRAM_ADDR196memory address3.3O
V
SS11
SDRAM_ADDR498memory address3.3O
SDRAM_ADDR299memory address3.3O
SDRAM_ADDR3100memory address3.3O
V
SSCO3
V
DDCO3
V
DD11
TEST8104IC test interface3.3I/O
TEST7105IC test interface3.3I/O
HS106horizontal synchronization input and output3.3I/O
VS107vertical synchronization input and output3.3I/O
V
SS12
YUV0109YUV video output at 27 MHz3.3O/Z
YUV1110YUV video output at 27 MHz3.3O/Z
YUV2111YUV video output at 27 MHz3.3O/Z
YUV3112YUV video output at 27 MHz3.3O/Z
V
DD12
YUV4114YUV video output at 27 MHz3.3O/Z
YUV5115YUV video output at 27 MHz3.3O/Z
YUV6116YUV video output at 27 MHz3.3O/Z
YUV7117YUV video output at 27 MHz3.3O/Z
TEST6118IC test interface3.3I/O
GRPH119indicator for graphics information output3.3O
TEST5120IC test interface3.3I/O
81read command output3.3O
82ground for pad ring−−
85supply for pad ring3.3−
89ground for pad ring−−
93supply for pad ring3.3−
97ground for pad ring−−
101ground for core logic−−
102supply for core logic3.3−
103supply for pad ring3.3−
108ground for pad ring−−
113supply for pad ring3.3−
1997 Jan 298
Page 9
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
SYMBOLPINDESCRIPTIONVI/O
V
DDA
V
SSA
V
SS13
CLK12427 MHz clock input3.3I
V
SS14
TCLK126boundary scan test clock input3.3I
TRST127boundary scan test reset input3.3I
TMS128boundary scan test mode select input3.3I
TD
O
TD
I
V
DD13
TEST4132IC test interface3.3I/O
TEST3133IC test interface3.3I/O
TEST2134IC test interface3.3I/O
TEST1135IC test interface3.3I/O
TEST0136IC test interface3.3I/O
V
DD14
RESET138hard reset input (active LOW)3.3I
FSCLK139256 or 384 fs (audio sampling) output3.3O/Z
V
DDCO4
V
SSCO4
SCLK142serial audio clock output3.3O/Z
SD143serial audio data output3.3O/Z
V
SS15
WS145word select output3.3O/Z
SPDIF146digital audio output3.3O/Z
ERROR147flag for bitstream error input5.0I
V_STROBE148video strobe input5.0I
V
DD15
AV_DATA0150MPEG input port for PES data5.0I
AV_DATA1151MPEG input port for PES data5.0I
AV_DATA2152MPEG input port for PES data5.0I
AV_DATA3153MPEG input port for PES data5.0I
V
SS16
AV_DATA4155MPEG input port for PES data5.0I
AV_DATA5156MPEG input port for PES data5.0I
AV_DATA6157MPEG input port for PES data5.0I
AV_DATA7158MPEG input port for PES data5.0I
A_STROBE159audio strobe input5.0I
V
DD16
121supply for analogue blocks3.3−
122ground for analogue blocks−−
123ground for pad ring−−
125ground for pad ring−−
129boundary scan test data output3.3O
130boundary scan test data input3.3I
131supply for pad ring3.3−
137supply for pad ring3.3−
140supply for core logic3.3−
141ground for core logic−−
The SAA7201 is an MPEG2 decoder which combines
audio decoding, video decoding and enhanced region
based graphics. The decoder operates with a single
16 Mbit external synchronous dynamic random access
memory (SDRAM) and runs from a single external 27 MHz
clock. Due to the optimized memory control for MPEG2
decoding, more than 1 Mbit is available for graphics in
50 Hz systems.
MPEG2 data can be accepted up to 9 Mbytes/s through a
dedicated byte wide interface. The data on this interface
can be either in PES (Packetized Elementary Stream),
MPEG1 packet or ES (Elementary Stream) format as
described in Chapter “References”. Two additional strobe
signals distinguish between audio and video data.
The internal video decoder is capable of decoding all
MPEG compliant streams up to main level main profile as
specified in Chapter “References”. The audio decoder
implements 2 channel audio decoding according to the
standards in Chapter “References”.
All real time audio/video decoding and synchronization
tasks are performed autonomously, so the external
microcontroller only needs to perform high-level tasks like
initialization, status monitoring and trick mode control.
The main support task of the external microcontroller
concerns the control of the graphical unit. This unit should
be supplied with bit-maps, determining the contents of the
graphical regions and by a simple set of instructions
determining the appearance of the graphical data on the
screen. Most graphical information should be stored in the
external memory which implies multiple data transfers
between CPU and the external memory. By performing
these data transfers on a direct memory access (DMA)
basis, full bit-maps can be transferred within one video
frame period.
The video output, containing a mix of MPEG video and
graphical data, is at a YUV multiplexed format which can
be directly connected to an external composite video
encoder. The audio output, containing a mix of MPEG
audio and programmable ‘beeps’, is in a serial, I2S or
Japanese format which can be directly supplied to most
commercially available up-sampling audio DA converters.
A functional block diagram of the decoder is given in Fig.1.
Its application environment is depicted in Fig.24. In the
following sections, a brief description of the individual
internal blocks of the MPEG2 decoder will be given.
Audio/video interface
In a basic set-top box application the SAA7201 receives
audio and video PES data in a byte wide format at rates up
to 9 Mbytes/s. A timing diagram is shown in Fig.3. Next to
the 8-bit wide data bus an audio and video strobe is
expected at the input. Erroneous data may be flagged via
the error indicator.
1997 Jan 2911
Page 12
Philips SemiconductorsObjective specification
Fig.4 Timing diagram of serial input mode.
handbook, full pagewidth
V_STROBE
A_STROBE
AV_DATA0
≥25 ns≥25 ns
≥50 ns
video bit (n + 6)video bit (n + 7)audio bit (m + 0)
first bit of a byte
MGD324
Integrated MPEG2 AVG decoderSAA7201
Alternatively data can be received in a 1-bit serial format at
rates up to 20 Mbit/s. In this mode, data is input at the LSB
input of the AV_DATA bus. Audio and video data must be
input in multiples of 8 bits. The first bit after switching from
audio to video (or the other way around) must be the first
bit of a byte since this transition will be used for the internal
bit-to-byte conversion.
Audio/video data can also be received via the CPU
interface in 8 or 16-bit mode. The peak rate is 27 Mbytes/s
in bursts of ≤128 bytes with a sustained rate up to
9 Mbytes/s. However, the MPEG bit rate is still limited to
15 Mbit/s for video and 448 kbit/s for audio.
Independent of the input mode all audio and video input
data are stored sequentially in the audio or video input
buffer area of the external memory. The audio and video
data can be either in MPEG2 PES, MPEG1 packet or ES
format.
Memory interface unit
The memory interface takes care of addressing and
control of the 16-Mbit external SDRAM. The SDRAM
should be either JEDEC compliant either the ‘lite/PC’
version.
Due to memory communication requirements this interface
runs at 81 MHz. The SDRAM types used with the
SAA7201 should be organized as 1M × 16, split internally
in two banks, each having 2048 pages of 256 words of
16 bits.
The target SDRAM type is NECµPD 4516161G5-A12-7FJ
(83 MHz JEDEC version) or NEC
µPD 4516421G5-A83-7FJ-PC (83 Mhz PC version).
Clock generation
The clock generation unit generates all the internal
processing clocks, the clock for the system time base
counter and the audio oversampling clock for the audio
DAC. For this purpose a non-integer divider plus a PLL is
implemented. In order to get reliable audio and video
decoding the 27 MHz input clock should be locked
externally to the MPEG time base.
Host interface system
The host interface system handles the communication
between on one side the SAA7201 plus SDRAM and on
the other side the external CPU. The interface consists of
a 16-bit wide data bus plus 8 address lines. It is compatible
with both Motorola’s 68xxx and Intel’s x86 family.
An optimized interface with the SAA7208 is also
supported. Via this interface a fast direct access to a large
number of internal status and data registers can be
achieved. Moreover, the external SDRAM can be
accessed via a specific register in combination with an
internally implemented auto increment counter.
The access to the external SDRAM is guaranteed up to a
sustained data rate of 9 Mbyte per second. However, in
practice the achievable data rate can be much higher.
Next to the data and address lines, 4 interrupt lines are
part of the host interface bus. Each interrupt line can
monitor up to 32 internal events which all can be masked
individually. Examples of internal events are audio/video
bit stream information, decoder status, internal error
conditions and input buffer occupation. The latter may be
very useful in interactive applications to serve as input data
request line.
1997 Jan 2912
Page 13
Philips SemiconductorsObjective specification
Fig.5 Audio decoding unit.
handbook, full pagewidth
BUFFER AND
SYNC UNIT
AUDIO
CLOCK
GENERATOR
AUDIO
BEEP
DRAM-bus
81 MHz
MPEG
AUDIO
DECODER
OUTPUT
INTERFACE
Sony or
I
2
S-bus
SPDIF
MGD325
Audio decoding unit
+
Integrated MPEG2 AVG decoderSAA7201
System time base unit
The system time base unit serves as a timing master for all
internal processes. It consists of two 24-bit wide System
Time Clock (STC) counters, running at 90 kHz. The STCs
will be used as internal synchronization reference for audio
and video.The contents of the STC can be loaded by the
external CPU which should insure that the phase of the
SAA7201 internal STC is identical to the main system time
clock in the system demultiplexer. The CPU should correct
for possible latency problems.
Because two counters are implemented, the previous time
base reference which might still be required as reference
for some time in case of time base discontinuity, can be
maintained. Thus all information for audio/video
synchronization is available in the decoder chip and only
minor support of the external controller is required.
The synchronization of graphics for e.g. subtitling, should
be controlled by the external CPU.
Video input buffer and synchronization control
The size and position of the video input buffer in the
external SDRAM is programmable. By default 2.6 Mbit/s
are reserved for the video input buffer but in principle any
other value can be programmed. The current fullness of
the video input buffer can be monitored by the CPU and an
internal interrupt will be generated is case of either near
over- or near underflow.
Data retrieval from the input buffer can be controlled by
DTS time stamps parsed from the PES or MPEG1 packet
stream. For those frames where no DTS time stamp is
present in the video bitstream a DTS is emulated by the
SAA7201.
Obviously this emulation mode can also be used when the
input stream is a video elementary stream (ES). The latter
case should be handled by start and stop decode
commands from the CPU.
The external CPU can select to retrieve the video PES
header and/or video PES private data for further software
processing.
Audio input buffer and synchronization control
The audio input buffer and synchronization control
basically behaves identical to its video counter part.
The default buffer size is 64 kbit in this case.
Synchronization will be controlled by PTS time stamps in
the audio Packetized Elementary Stream. Also in this case
an PTS emulation or a free running start/stop controlled
mode are supported.
Audio decoder
A functional block diagram for the audio decoding part is
depicted in Fig.5.
1997 Jan 2913
Page 14
Philips SemiconductorsObjective specification
Fig.6 I2S-bus format with 18-bit sample precision.
handbook, full pagewidth
b17
MSB aligned
MGD326
left sample nleft sample n + 1right sample n
0
31 3263
SCLK
WS
SD
b0
b17b0b17b0
Integrated MPEG2 AVG decoderSAA7201
Audio decoding will be performed at a clock locked to the
video decoding clock and only the output interface is
running on the audio oversampling clock.
The audio decoder unit performs the decoding of the
selected MPEG audio stream in a range from 8 up to
448 kbit/s in a fixed or variable bit rate format. Decoding is
restricted to 2 channel, layer I, II MPEG audio at sampling
frequency of 48.0, 44.1, 32.0, 24.0, 22.05 or 16.0 kHz.
The audio decoder support the stop, mute, and skip
function to support insertion
Apart from basic MPEG processing the audio decoder
core contain also:
• Support for: stop, mute and skip function.
• Fully parameterized dynamic range compression unit to
decrease the dynamic range of the output signal on
audio frame basis. Depending on the power level a
programmable amplification and offset may be applied.
• Fully programmable base band audio processing unit to
control the gain in both output channels independently
and/or to mix both channels.
• MPEG de-emphasis filtering on the output data, thus
avoiding the need of external analog de-emphasis filter
circuitry.
• Storage buffer for the last 54 bytes of each audio frame.
The CPU can retrieve eventual ancillary data from this
buffer.
The output of the audio decoder unit can be mixed with
square waveform audio signals which are generated by a
beep generator. Programmable parameters for the beep
generator are amplitude, frequency and duration.
The audio output interface module produces stereo base
band output samples on two different outputs at the same
time:
• Serial digital audio in I2S-bus or in Japanese format in
16, 18, 20 or 22-bit
• SPDIF (Sony/Philips Digital Interface).
Any of the two outputs may be enabled or set to high
impedance mode. The I2S-bus format with 18-bit sample
precision is shown in Fig.6.
The difference between I2S-bus and the Japanese format
is that I2S-bus is MSB aligned whereas the Japanese
format is LSB aligned.
The 1-bit serial interface SPDIF contains 64-bit per audio
sample period. Complete frames must be transmitted at
the audio sample rate. Not only left/right information but
also validity flags, channel status, user data and parity
information is contained in an SPDIF frame (see Chapter
“References”).
1997 Jan 2914
Page 15
Philips SemiconductorsObjective specification
Fig.7 Video decoding unit.
handbook, full pagewidth
VLD
FLD
IZZIQIDCT
MC
INTERP
FBM
to display
unit
to external
memory
from reference
memory
from
input
buffer
MGD327
Integrated MPEG2 AVG decoderSAA7201
Video decoder
The video decoding unit provides all actions required for
compliant decoding of MPEG2 main level, main profile
coded video bit streams. The decoding process consists of
fixed and variable length decoding, run length decoding,
inverse quantization, inverse discrete cosine
transformation, motion compensation and interpolation.
In general the arithmetic decoding result is stored as
reference picture in the external memory.
Decoded B-frames are only stored for the conversion from
the frame coded macro block (MB) to the scanning line
format. In many cases a field storage is sufficient for this
conversion but in some cases the user might decide to use
a full frame storage to enable chroma frame up-conversion
or full performance 3 : 2 pull-down in 60 Hz systems.
Obviously when using less memory for the video decoding
process more memory is available for non-video decoding
tasks.
The Frame Buffer Management unit (FBM) manages the
allocation of frame buffers in external SDRAM for both
video decoding and display unit and can be programmed
to use less memory in not fully MP@ML bitstreams:
smaller pictures (e.g. 544 × 576), simple profile, etc.
Apart from decoding compliant MPEG video streams the
decoder deals with some trick modes. Supported are field
or frame freeze at I or P pictures or freeze field on
B-pictures. In the latter case decoding will continue as a
background process and the output can be restarted at
any moment. When receiving non-compliant MPEG
streams the decoder can be switched to a scanning mode
in which only I or I + P frames are decoded while skipping
all other pictures. In the single step mode, the decoder
decodes just one frame and awaits a next step command.
The functional diagram of the video decoding unit is shown
in Fig.7.
1997 Jan 2915
Page 16
Philips SemiconductorsObjective specification
Fig.8 Graphics unit.
handbook, full pagewidth
REGION-1 (VBI-SIGNALS)
REGION-2
active
video
REGION-3 (256 ENTRY CLUT)
REGION-4
MGD328
active
video
H-size (2)H-start (2)
V-start (2)
V-size (2)
Fig.9 Display list.
handbook, full pagewidth
DESCRIPTOR 1
MAP
CLUT
bottom-field
top-field
table
data
PIXEL
DATA
PIXEL
DATA
128 bit
gfx
anchor
address
eodl
position, size, format
pointers to locations in SDRAM
DESCRIPTOR 2
DESCRIPTOR 3
DESCRIPTOR 4
MGD329
Integrated MPEG2 AVG decoderSAA7201
Graphics unit
The SAA7201 incorporates the display support for pixel
based graphics. Possible applications are the user
interface, logos and subtitling. Graphical data should be
grouped logically in regions and will be displayed in boxes
at the screen.
The definition of each region in the decoder consists of
four parts being a region descriptor, a top-field descriptor,
a bottom-field descriptor and a table-data descriptor:
• The region descriptor contains information relevant for
the full region like format, size, position and pointers to
the other descriptors.
• The top-field and bottom-field descriptor contain a pixel
based bit-map for the contents of that region for both
fields independently. The bit-maps can be stored in
either straight forward or in a compressed bit-map
format.
• The table-data descriptor defines the tables to be used
for the transformation of bit-maps to display colours.
All descriptors should be loaded under control of the
external CPU in the external memory.
The appearance of graphical data at the display is
determined by the assembly of region descriptors in a so
called display list. An example of such a display list for the
4 regions example is shown in Fig.9.
1997 Jan 2916
Page 17
Philips SemiconductorsObjective specification
Fig.10 Bit-map to pixel conversion.
handbook, full pagewidth
8
4 LSBs
2 LSBs
2 LSBs
4 LSBs
8 LSBs
8
CLUT
8
8
MGD330
MAP
TABLE 4 to 8
MAP
TABLE 2 to 8
8
Y
U
V
T
8
8
8
Integrated MPEG2 AVG decoderSAA7201
Basically there is no restriction on the number of different
regions but because regions may not vertically overlap the
practical limit will be the number of lines within a field.
However, one should realize that each region requires its
own 128-bit region descriptor.
The display list will be scanned twice per frame, once for
each display field. The region descriptors should be
ordered properly in the external SDRAM, starting from the
graphics anchor address. The last descriptor in the list
must have the end of display list indicator set.
Multiple pixel bit-maps, CLUTs and map tables may be
stored in the external memory but per region only two
bit-maps (one for each fields) and two tables
(CLUT + map table) may be used. Obviously bit-maps and
tables may be shared by multiple regions.
Pixel data bit-maps can be described in 2, 4 or 8 bit/pixel
in either a direct bit-map or coded in a one-dimensional (H)
variable and run length encoded format according the
pixel-data-sub-block syntax as specified in Chapter
“References” and illustrated in Chapter “Appendix”.
The actual coding format is specified in the region
descriptor for each region thus allowing different coding
schemes within a picture.
During display the 2, 4 and 8 bit/pixel bit-maps will be
transformed, eventually with run length decoding, via a
table look-up mechanism into a 4, 16 or 256 different YUV
colours with 8-bit resolution for each component plus a
factor T for mixing of graphics and MPEG video.
In order to obtain maximum flexibility two cascaded tables
are active in this bit-map to pixel conversion as indicated
in Fig.10.
The tables are retrieved from the external memory just
before the region is going to be displayed. One table per
region can be updated and for small tables this occurs
during the horizontal blanking interval. However, updating
a 256 entry CLUT may take about one line period which
means that a spatial separation of one line with the
previous region is mandatory in this case. If the required
tables for a certain region are already stored in the local
memory, the table down load action can be skipped.
Additionally some special bits can be set in the region
descriptor.
• Transparency shift: this parameter overrules the pixel
based transparency in order to support fading of the
entire graphical region.
• Zoom: this parameter initiates horizontal pixel repetition.
It should be noted that a copy of pixels in vertical
direction can be achieved by pointing to a single bit-map
for both fields.
Regions can also be defined in the vertical blanking
interval. In combination with 8 bit/pixel coding, arbitrary
test signals on 13.5 MHz grid can be programmed.
Possible application areas are teletext, closed caption,
wide screen signalling bits, Video Programming Signals
(VPS) and Vertical Interval Test Signals (VITS).
As indicated above multiple regions can be specified in a
display list which will be scanned sequentially every frame.
In case of stationary graphics no updates of the display list
are required, but the external CPU can update it
dynamically to achieve scrolling and/or fading of one or
more graphical boxes. The display list mechanism also
allows for non real time transfer of large bit-maps by
keeping that region out of the display list during loading.
1997 Jan 2917
Page 18
Philips SemiconductorsObjective specification
Fig.11 Display unit.
handbook, full pagewidth
WINDOW
EXTRACTION
SCALING
64 STEP
MIXER
4 STEP
MIXER
GRAPHICSCURSOR
OUTPUT
FORMAT
to digital
CVBS
encoder
SHIFTINGCLIPPING
CHROMA
UP-CONVERSION
MPEG
decoded
picture
MGD331
Integrated MPEG2 AVG decoderSAA7201
CURSOR PROCESSING
Additionally to the above defined graphics boxes one
cursor can be activated on the screen. Since the cursor
data is fully stored locally, no overlapping restriction apply
to this box so the cursor can moved over the entire screen.
The cursor can be as large as 1 kpixel with a 2 bits/pixel
colour depth. Obviously data transfer can be done on DMA
basis and need only be performed when a cursor is
required or when its contents must be modified.
The cursor XY dimensions (where the Y dimension refers
to frame lines) can be selected between 8 × 128, 16 × 64,
32 × 32, 64 × 16 and 128 × 8. On top of these shapes, a
zoom with a factor 2 can be applied in both directions
independently.
The cursor pixels will be translated via a 4-entry CLUT to
YUV colours and a transparency factor T. The resolution
of the YUV parameters is 6, 4, 4 bits respectively.
The T parameter is coded in 2 bits to enable the mixing
with video and graphics in 4 steps being 100% (cursor
only), 50%, 25% and 0% (fully transparent cursor).
Display unit
Before feeding the MPEG decoded and graphical data to
the output, a display unit re-formats the MPEG specific
4 : 2 : 0 format to CCIR-601 4 : 2 : 2 format and performs
a mixing between video and graphics where required.
The output picture can be up to 720 × 576 pixels at 50 Hz
or 720 × 480 pixels at 60 Hz.
A schematic representation of this unit is shown in Fig.11.
• In a first step a selected window can be retrieved from
the decoded MPEG data. This might be useful for e.g.
pan and scan operations for aspect ratio conversion.
• In case the resulting number of pixels per line does not
match the 720 pixels/line output format a horizontal
scaler can be activated. This scaling unit can transform
any number of bits below 720 to the required output
format. Internally a poly-phase filter is used which
performs a 64 phases interpolation. Not only
up-conversion but also down-conversion is supported
up to a scaling by a factor 2. Thus horizontal scaling can
be performed in a range from 0.5 up to 64. In practice
the maximum up-conversion factor will be less or equal
to 4.
• In vertical direction the picture can be expanded or
scaled down, in both cases by a factor 2. Expansion with
a factor 2 might be relevant for the up-conversion of SIF
resolution pictures to full screen. The factor 2 scaling, if
combined with the appropriate horizontal scaling, results
is1⁄4picture thus freeing-up a large screen area for
graphics. This might be very useful for electronic
program guide applications. It should be noted that in
case of picture compression an anti-aliasing filter can be
activated.
• Shifting: when the resulting MPEG picture is smaller
than the 720 × 576 (480) display format, this picture can
be located anywhere on the display screen.
Moreover, the non-covered area can be given any
background colour.
• Clipping: the amplitude of the MPEG decoded and
re-scaled video signal is kept within the range 16 to 235
for luminance and 16 to 240 for both chrominance
components.
1997 Jan 2918
Page 19
Philips SemiconductorsObjective specification
Fig.12 Timing diagram of graphics information output.
handbook, full pagewidth
CLK
YUV
HREF
CREF
(example)
GRPH
'128''16'
U
0
Y
0
V
0
Y
2
MGD332
Integrated MPEG2 AVG decoderSAA7201
• The chroma up-conversion unit converts the MPEG
4 : 2 : 0 format into the at the output required
4 : 2 : 2 format. This vertical up-conversion is performed
by a simple 8-phase interpolation between two adjacent
lines.
• The mixer units combine MPEG video with graphics and
cursor information in two steps. In a first step the MPEG
decoded information is mixed with graphical information.
Mixing can be done at pixel basis in 64 steps and is
controlled by the internally implemented colour look up
table. In a second step, the video plus graphics can be
mixed in 4 steps with the internally generated cursor.
• The output formatting unit performs two main tasks, i.e.
synchronization and formatting. Synchronization is
characterized by three signals being horizontal (H),
vertical (V) and field parity (FP), all having
programmable length and polarity.
Since the decoder can operate in master or slave mode,
the synchronization signals can be generated by the
decoder or should be delivered by an external device.
In both cases the length and polarity should be
programmed internally.
The video output samples are supplied in a multiplexed
YUV format to the output. Next to this byte wide YUV
stream, which can directly be supplied to most
commercially available composite video encoder ICs,
three additional signals are delivered at the output.
HREF indicates all active samples; CREF can flag any
combination of pixels: U, V, Y
and/or Y
odd
even
; GRPH
flags all the pixels inside a graphical box.
Additionally the full YUV bus can be set to a HIGH
impedance state under control of the signal YUV_ENA.
This might be useful for multiplexing the MPEG decoder
output with any other signal source on static basis.
1997 Jan 2919
Page 20
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
JTAG
The SAA7201 supports the standard Boundary Scan test
instructions: bypass, extest, sample, intest, runbist,
idcode.
Memory requirements
As indicated before the MPEG source decoder operates
with 16 Mbit of external memory. Several processes
require access to the external memory, mostly being the
video decoding process. In normal main level, main profile
video applications about 1.2 Mbit of memory space is free
for non-video processes. In practice most of this capacity
will be used for graphics. In combination with the internal
variable length decoding, full screen graphics at 8-bit per
pixel is feasible. Moreover, by having introduced a flexible
memory allocation procedure the available memory
capacity for graphics may be enlarged when decoding
lower resolution MPEG pictures or when the input bit rate
is less than 15 Mbit/s.
Table 1 Required memory capacity for some user defined modes
System50 Hz60 Hz
Obviously for graphics-only applications all 16 Mbit can be
used for the storage of bit-maps and look-up tables.
In Table 1 an overview is given of the required memory
capacity for some user defined modes.
In 50 Hz systems memory capacity can be saved by
restricting the chroma vertical interpolation to field
interpolation. This mode would only bring some extra
chroma resolution in case the input stream contains
progressive coded pictures.
In 60 Hz systems the reduction of storage capacity for
B-frames to field capacity has not only consequences for
the chroma vertical interpolation but also for the
3 : 2 pull-down operation mode. The operation
repeat-first-field is not possible in all cases and a modified
3 : 2 pull-down is performed under the control of the
SAA7201. The user may decide to use this modified
3 : 2 pull-down mode in order to have more memory
available for OSD or graphics.
HIGH level output voltage2.4−−V
LOW level output voltage−−0.4V
cycle time37.03637.03737.038ns
rise time2−4ns
fall time2−4ns
set-up time8−−ns
hold time0−−ns
address/CS set-up time20−−ns
address/CS hold time75−−ns
data write set-up time20−−ns
data read set-up time20−−ns
data release time0−10ns
control signal hold time0−−ns
acknowledge pulse width25−−ns
acknowledge release time0−10ns
delay time for acknowledge read96−125ns
delay time for acknowledge write48−75ns
write/read pulse width25−−ns
hold time3−t
delay timet
h
−20ns
d
ns
load capacitance10−30pF
1997 Jan 2922
Page 23
Philips SemiconductorsObjective specification
handbook, full pagewidth
ADDRESS
DATA
MGD333
ADDRESS
(8 to 1)
CS
DS
DTACK
DATA
(15 to 0)
R/W
t
su(A-CS)
t
h(A-CS)
t
su(D-W)
t
h(D)
t
W(ACK)
t
rel
t
rel(ACK)
t
d(ACK-W)
t
W(RW)
Fig.13 Motorola write timing (non-multiplexed).
handbook, full pagewidth
ADDRESS
DATA
MGD334
ADDRESS
(8 to 1)
CS
DS
DTACK
DATA
(15 to 0)
R/W
t
su(A-CS)
t
h(A-CS)
t
su(D-R)
t
AR
t
rel(D)
t
W(ACK)
t
rel
t
rel(ACK)
t
d(ACK-R)
t
W(RW)
Fig.14 Motorola read timing (non-multiplexed).
Integrated MPEG2 AVG decoderSAA7201
1997 Jan 2923
Page 24
Philips SemiconductorsObjective specification
Fig.15 Intel write timing (non-multiplexed).
handbook, full pagewidth
ADDRESS
DATA
MGD335
ADDRESS
(8 to 1)
CS
RDY
DATA
(15 to 0)
R/W
t
su(A-CS)
t
h(A-CS)
t
su(D-W)
t
h(D)
t
rel
t
rel(ACK)
t
d(ACK-W)
t
W(RW)
Fig.16 Intel read timing (non-multiplexed).
handbook, full pagewidth
ADDRESS
DATA
MGD336
ADDRESS
(8 to 1)
CS
RDY
DATA
(15 to 0)
R/W
t
su(A-CS)
t
h(A-CS)
t
rel(D)
t
AR
t
su(D-R)
t
rel
t
rel(ACK)
t
d(ACK-R)
t
W(RW)
Integrated MPEG2 AVG decoderSAA7201
1997 Jan 2924
Page 25
Philips SemiconductorsObjective specification
Fig.17 DMA read access.
handbook, full pagewidth
DATA
MGD337
CS
DMA_ACK
DMA_RDY
DATA
(15 to 0)
DMA_REQ
t
su(A-CS)
t
h(A-CS)
t
su(D-W)
t
h(D)
t
rel
t
rel(ACK)
t
d(ACK-W)
t
W(RW)
Fig.18 DMA write access.
handbook, full pagewidth
DATA
MGD338
CS
DMA_ACK
DMA_RDY
DATA
(15 to 0)
DMA_REQ
t
su(A-CS)
t
h(A-CS)
t
AR
t
su(D-R)
t
rel
t
rel(ACK)
t
rel(D)
t
d(ACK-R)
t
W(RW)
Integrated MPEG2 AVG decoderSAA7201
1997 Jan 2925
Page 26
Philips SemiconductorsObjective specification
Fig.19 DMA_DONE timing.
handbook, full pagewidth
MGD339
DMA_ACK
DMA_DONE
DMA_REQ
t
su(A-CS)th(A-CS)
Fig.20 SAA7208 mode; read timing.
handbook, full pagewidth
MGD919
ADDRESS
ADDRESS
(8 to 0)
DATA
DATA
(15 to 0)
RDY
ALE
t
h(CS)
t
h(CS-ALE)
t
W(ALE)
t
W(SDR)
t
rel(D)
t
su(SDR-D)
t
su(ALE-SDR)
t
su(A-CS)
t
h(A-CS)
CS
SYSRD
Integrated MPEG2 AVG decoderSAA7201
1997 Jan 2926
Page 27
Philips SemiconductorsObjective specification
Fig.21 SAA7208 mode; write timing.
handbook, full pagewidth
MGD918
DATA
(15 to 0)
ADDRESS
(8 to 0)
CS
SYSRD
RDY
DATA
ADDRESS
ALE
t
h(CS)
t
h(A-CS)
t
su(A-CS)
t
h(CS-ALE)
t
W(SDR)
t
W(ALE)
t
W(D-SDR)
t
h(D-SDR)
Integrated MPEG2 AVG decoderSAA7201
1997 Jan 2927
Page 28
Philips SemiconductorsObjective specification
Fig.22 SAA7208 mode; burst write timing.
handbook, full pagewidth
MGD920
DATA
(15 to 0)
ADDRESS
(8 to 0)
CS
SYSRD
RDY
DATADATA
ALE
t
W(ALE)
t
su(D-SDR)
t
h(D-SDR)
t
su(D-SDR)
t
h(D-SDR)
t
W(SDR)
t
W(SDR)
t
rel(SDR)
ADDRESSADDRESS
t
h(CS)
t
h(CS-ALE)
t
su(A-SDR)
t
h(A-SDR)
t
su(A-SDR)
t
h(A-SDR)
Integrated MPEG2 AVG decoderSAA7201
1997 Jan 2928
Page 29
Philips SemiconductorsObjective specification
Fig.23 Intel read/write timing multiplexed.
handbook, full pagewidth
ADDRESS
ADDRESS
DATA
DATA
MGD341
AD(read)
AD(write)
CS
RD
RDY
ALE
t
W(ACK)
t
rel(D)
t
rel(ACK)
t
W(A-CS)
t
su(D-R)
t
AR
t
h(D)
t
su(D-W)
t
h(A-CS)
t
su(A-S)
t
rel
t
W(RW)
WR
Integrated MPEG2 AVG decoderSAA7201
REFERENCES
1.
2.
3.
4.
5.
6.
7.
MPEG ISO/IEC 11172-1 International standard; MPEG-1 systems.
MPEG ISO/IEC 13818-1 International standard; MPEG-2 systems.
MPEG ISO/IEC 11172-2 International standard; MPEG-1 Video.
MPEG ISO/IEC 13818-2 International standard; MPEG-2 Video.
MPEG ISO/IEC 11172-3 International standard; MPEG-1 Audio.
MPEG ISO/IEC 13818-3 International standard; MPEG-2 Audio.
DVB subtitling system; working draft 2.0; TM 1398 rev 2
.
1997 Jan 2929
Page 30
Philips SemiconductorsObjective specification
Fig.24 Application diagram.
handbook, full pagewidth
SAA7208
(DEMUX/MIPS)
4 Mbit
EPROM
4 Mbit
DRAM
SAA7201
SAA7183
(EURO-DENC)
16 Mbit
SDRAM
AUDIO
DA
8+3
12
address datacontrol
8
16
16
INT
H,V validYUV
CVBS
MGD342
RGB
Y/C
L
R
strobe
27.0
MHz
27 MHz
I
2
C-bus
I
2
C-bus
H, V, FP
TTX/TTXRQ
high
speed
data
Integrated MPEG2 AVG decoderSAA7201
APPLICATION INFORMATION
1997 Jan 2930
Page 31
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
APPENDIX
Syntax of pixel-data-sub-block
data typepixel-code-string of N-coded wordsend of string
2 bit/pixel011 pixel in colour 1
101 pixel in colour 2
0001 0000111 pixel in colour 300 00 00
00 011 pixel in colour 0
00 00 012 pixels in colour 0
00 1L LL CCL pixels (3 to 10) in colour C
00 00 10 LL LL CCL pixels (12 to 27) in colour C
00 00 11 LL LL LL LL CCL pixels (29 to 284) in colour C
4 bit/pixel00011 pixel in colour 1
↓↓↓↓
0001 000111111 pixel in colour 150000 0000
0000 11001 pixel in colour 0
0000 11012 pixels in colour 0
0000 0LLL (L>0)L pixels (3 to 9) in colour 0
0000 10LL CCCCL pixels (4 to 7) in colour C
0000 1110 LLLL CCCCL pixels (9 to 24) in colour C
0000 1111 LLLL LL11 CCCCL pixels (25 to 280) in colour C
8 bit/pixel000000011 pixel in colour 100000000-- --00000000
↓↓↓↓
0001 0010111111111 pixel in colour 255
00000000 0LLLLLLLL pixels (1 to 127) in colour 0
00000000 1LLLLLLL CCCCCCCCL pixels (3 to 127) in colour C
1997 Jan 2931
Page 32
Philips SemiconductorsObjective specification
UNITA1A2A3b
p
cE
(1)(1)(1)
eHELLpQZywvθ
REFERENCES
OUTLINE
VERSION
EUROPEAN
PROJECTION
ISSUE DATE
IEC JEDEC EIAJ
mm
0.40
0.25
3.70
3.15
0.25
0.40
0.25
0.23
0.13
28.1
27.9
0.650.31.95
32.2
31.6
1.70
1.55
1.5
1.1
8
0
o
o
0.15 0.1
DIMENSIONS (mm are the original dimensions)
Note
1. Plastic or metal protrusions of 0.25 mm maximum per side are not included.
1.1
0.7
SOT322-1 MO112DD1
93-08-25
95-02-04
D
(1)
28.1
27.9
H
D
32.2
31.6
E
Z
1.5
1.1
D
pin 1 index
b
p
e
θ
E
A
1
A
L
p
Q
detail X
L
(A )
3
B
40
c
D
H
b
p
E
H
A
2
v M
B
D
Z
D
A
Z
E
e
v M
A
X
1
160
121
120
81
80
41
y
w M
w M
0510 mm
scale
SOT322-1
160 leads (lead length 1.95 mm); body 28 x 28 x 3.4 mm; high stand-off height
QFP160: plastic quad flat package;
A
max.
3.95
Integrated MPEG2 AVG decoderSAA7201
PACKAGE OUTLINE
1997 Jan 2932
Page 33
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
SOLDERING
Introduction
There is no soldering method that is ideal for all IC
packages. Wave soldering is often preferred when
through-hole and surface mounted components are mixed
on one printed-circuit board. However, wave soldering is
not always suitable for surface mounted ICs, or for
printed-circuits with high population densities. In these
situations reflow soldering is often used.
This text gives a very brief insight to a complex technology.
A more in-depth account of soldering ICs can be found in
our
“IC Package Databook”
Reflow soldering
Reflow soldering techniques are suitable for all QFP
packages.
The choice of heating method may be influenced by larger
plastic QFP packages (44 leads, or more). If infrared or
vapour phase heating is used and the large packages are
not absolutely dry (less than 0.1% moisture content by
weight), vaporization of the small amount of moisture in
them can cause cracking of the plastic body. For more
information, refer to the Drypack chapter in our
Reference Handbook”
Reflow soldering requires solder paste (a suspension of
fine solder particles, flux and binding agent) to be applied
to the printed-circuit board by screen printing, stencilling or
pressure-syringe dispensing before package placement.
Several techniques exist for reflowing; for example,
thermal conduction by heated belt. Dwell times vary
between 50 and 300 seconds depending on heating
method. Typical reflow temperatures range from
215 to 250 °C.
Preheating is necessary to dry the paste and evaporate
the binding agent. Preheating duration: 45 minutes at
45 °C.
(order code 9398 652 90011).
“Quality
(order code 9397 750 00192).
Wave soldering
Wave soldering is not recommended for QFP packages.
This is because of the likelihood of solder bridging due to
closely-spaced leads and the possibility of incomplete
solder penetration in multi-lead devices.
If wave soldering cannot be avoided, the following
conditions must be observed:
• A double-wave (a turbulent wave with high upward
pressure followed by a smooth laminar wave)
soldering technique should be used.
• The footprint must be at an angle of 45° to the board
direction and must incorporate solder thieves
downstream and at the side corners.
Even with these conditions, do not consider wave
soldering the following packages: QFP52 (SOT379-1),
QFP100 (SOT317-1), QFP100 (SOT317-2),
QFP100 (SOT382-1) or QFP160 (SOT322-1).
During placement and before soldering, the package must
be fixed with a droplet of adhesive. The adhesive can be
applied by screen printing, pin transfer or syringe
dispensing. The package can be soldered after the
adhesive is cured.
Maximum permissible solder temperature is 260 °C, and
maximum duration of package immersion in solder is
10 seconds, if cooled to less than 150 °C within
6 seconds. Typical dwell time is 4 seconds at 250 °C.
A mildly-activated flux will eliminate the need for removal
of corrosive residues in most applications.
Repairing soldered joints
Fix the component by first soldering two diagonallyopposite end leads. Use only a low voltage soldering iron
(less than 24 V) applied to the flat part of the lead. Contact
time must be limited to 10 seconds at up to 300 °C. When
using a dedicated tool, all other leads can be soldered in
one operation within 2 to 5 seconds between
270 and 320 °C.
1997 Jan 2933
Page 34
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
DEFINITIONS
Data sheet status
Objective specificationThis data sheet contains target or goal specifications for product development.
Preliminary specificationThis data sheet contains preliminary data; supplementary data may be published later.
Product specificationThis data sheet contains final product specifications.
Limiting values
Limiting values given are in accordance with the Absolute Maximum Rating System (IEC 134). Stress above one or
more of the limiting values may cause permanent damage to the device. These are stress ratings only and operation
of the device at these or at any other conditions above those given in the Characteristics sections of the specification
is not implied. Exposure to limiting values for extended periods may affect device reliability.
Application information
Where application information is given, it is advisory and does not form part of the specification.
LIFE SUPPORT APPLICATIONS
These products are not designed for use in life support appliances, devices, or systems where malfunction of these
products can reasonably be expected to result in personal injury. Philips customers using or selling these products for
use in such applications do so at their own risk and agree to fully indemnify Philips for any damages resulting from such
improper use or sale.
1997 Jan 2934
Page 35
Philips SemiconductorsObjective specification
Integrated MPEG2 AVG decoderSAA7201
NOTES
1997 Jan 2935
Page 36
Philips Semiconductors – a worldwide company
Argentina: see South America
Australia: 34 Waterloo Road, NORTH RYDE, NSW 2113,
United States: 811 East Arques Avenue, SUNNYVALE, CA 94088-3409,
Tel. +1 800 234 7381
Uruguay: see South America
Vietnam: see Singapore
Yugoslavia: PHILIPS, Trg N. Pasica 5/v, 11000 BEOGRAD,
Tel. +381 11 625 344, Fax.+381 11 635 777
For all other countries apply to: Philips Semiconductors, Marketing & Sales Communications,
Building BE-p, P.O. Box 218, 5600 MD EINDHOVEN, The Netherlands, Fax. +31 40 27 24825
The information presented in this document does not form part of any quotation or contract, is believed to be accurate and reliable and may be changed
without notice. No liability will be accepted by the publisher for any consequence of its use. Publication thereof does not convey nor imply any license
under patent- or other industrial or intellectual property rights.
Internet: http://www.semiconductors.philips.com
Printed in The Netherlands547047/1200/01/pp36 Date of release: 1997 Jan 29Document order number: 9397 750 00989
Loading...
+ hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.