cellar M. Niedermayer
Internet-Draft
Intended status: Standards Track D. Rice
Expires: March 29, 2019
J. Martinez
September 25, 2018
FFV1 Video Coding Format Version 4
draft-ietf-cellar-ffv1-v4-02
Abstract
This document defines FFV1, a lossless intra-frame video encoding
format. FFV1 is designed to efficiently compress video data in a
variety of pixel formats. Compared to uncompressed video, FFV1
offers storage compression, frame fixity, and self-description, which
makes FFV1 useful as a preservation or intermediate video format.
Status of This Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
This Internet-Draft will expire on March 29, 2019.
Copyright Notice
Copyright (c) 2018 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(https://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
Niedermayer, et al. Expires March 29, 2019 [Page 1]
Internet-Draft FFV1 September 2018
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4
2. Notation and Conventions . . . . . . . . . . . . . . . . . . 4
2.1. Definitions . . . . . . . . . . . . . . . . . . . . . . . 4
2.2. Conventions . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.1. Pseudo-code . . . . . . . . . . . . . . . . . . . . . 5
2.2.2. Arithmetic Operators . . . . . . . . . . . . . . . . 5
2.2.3. Assignment Operators . . . . . . . . . . . . . . . . 6
2.2.4. Comparison Operators . . . . . . . . . . . . . . . . 6
2.2.5. Mathematical Functions . . . . . . . . . . . . . . . 7
2.2.6. Order of Operation Precedence . . . . . . . . . . . . 7
2.2.7. Range . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.8. NumBytes . . . . . . . . . . . . . . . . . . . . . . 8
2.2.9. Bitstream Functions . . . . . . . . . . . . . . . . . 8
3. General Description . . . . . . . . . . . . . . . . . . . . . 8
3.1. Border . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2. Samples . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.3. Median Predictor . . . . . . . . . . . . . . . . . . . . 9
3.4. Context . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.5. Quantization Table Sets . . . . . . . . . . . . . . . . . 11
3.6. Quantization Table Set Indexes . . . . . . . . . . . . . 11
3.7. Color spaces . . . . . . . . . . . . . . . . . . . . . . 11
3.7.1. YCbCr . . . . . . . . . . . . . . . . . . . . . . . . 11
3.7.2. RGB . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.8. Coding of the Sample Difference . . . . . . . . . . . . . 13
3.8.1. Range Coding Mode . . . . . . . . . . . . . . . . . . 13
3.8.2. Golomb Rice Mode . . . . . . . . . . . . . . . . . . 17
4. Bitstream . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1. Parameters . . . . . . . . . . . . . . . . . . . . . . . 20
4.1.1. version . . . . . . . . . . . . . . . . . . . . . . . 21
4.1.2. micro_version . . . . . . . . . . . . . . . . . . . . 22
4.1.3. coder_type . . . . . . . . . . . . . . . . . . . . . 23
4.1.4. state_transition_delta . . . . . . . . . . . . . . . 23
4.1.5. colorspace_type . . . . . . . . . . . . . . . . . . . 23
4.1.6. chroma_planes . . . . . . . . . . . . . . . . . . . . 24
4.1.7. bits_per_raw_sample . . . . . . . . . . . . . . . . . 24
4.1.8. log2_h_chroma_subsample . . . . . . . . . . . . . . . 24
4.1.9. log2_v_chroma_subsample . . . . . . . . . . . . . . . 24
4.1.10. alpha_plane . . . . . . . . . . . . . . . . . . . . . 24
4.1.11. num_h_slices . . . . . . . . . . . . . . . . . . . . 25
4.1.12. num_v_slices . . . . . . . . . . . . . . . . . . . . 25
4.1.13. quant_table_set_count . . . . . . . . . . . . . . . . 25
4.1.14. states_coded . . . . . . . . . . . . . . . . . . . . 25
4.1.15. initial_state_delta . . . . . . . . . . . . . . . . . 25
Niedermayer, et al. Expires March 29, 2019 [Page 2]
Internet-Draft FFV1 September 2018
4.1.16. ec . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.1.17. intra . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2. Configuration Record . . . . . . . . . . . . . . . . . . 26
4.2.1. reserved_for_future_use . . . . . . . . . . . . . . . 26
4.2.2. configuration_record_crc_parity . . . . . . . . . . . 27
4.2.3. Mapping FFV1 into Containers . . . . . . . . . . . . 27
4.3. Frame . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4. Slice . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.5. Slice Header . . . . . . . . . . . . . . . . . . . . . . 29
4.5.1. slice_x . . . . . . . . . . . . . . . . . . . . . . . 30
4.5.2. slice_y . . . . . . . . . . . . . . . . . . . . . . . 30
4.5.3. slice_width . . . . . . . . . . . . . . . . . . . . . 30
4.5.4. slice_height . . . . . . . . . . . . . . . . . . . . 30
4.5.5. quant_table_set_index_count . . . . . . . . . . . . . 30
4.5.6. quant_table_set_index . . . . . . . . . . . . . . . . 31
4.5.7. picture_structure . . . . . . . . . . . . . . . . . . 31
4.5.8. sar_num . . . . . . . . . . . . . . . . . . . . . . . 31
4.5.9. sar_den . . . . . . . . . . . . . . . . . . . . . . . 31
4.5.10. reset_contexts . . . . . . . . . . . . . . . . . . . 31
4.5.11. slice_coding_mode . . . . . . . . . . . . . . . . . . 31
4.6. Slice Content . . . . . . . . . . . . . . . . . . . . . . 32
4.6.1. primary_color_count . . . . . . . . . . . . . . . . . 32
4.6.2. plane_pixel_height . . . . . . . . . . . . . . . . . 32
4.6.3. slice_pixel_height . . . . . . . . . . . . . . . . . 32
4.6.4. slice_pixel_y . . . . . . . . . . . . . . . . . . . . 33
4.7. Line . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.7.1. plane_pixel_width . . . . . . . . . . . . . . . . . . 33
4.7.2. slice_pixel_width . . . . . . . . . . . . . . . . . . 33
4.7.3. slice_pixel_x . . . . . . . . . . . . . . . . . . . . 33
4.7.4. sample_difference . . . . . . . . . . . . . . . . . . 33
4.8. Slice Footer . . . . . . . . . . . . . . . . . . . . . . 34
4.8.1. slice_size . . . . . . . . . . . . . . . . . . . . . 34
4.8.2. error_status . . . . . . . . . . . . . . . . . . . . 34
4.8.3. slice_crc_parity . . . . . . . . . . . . . . . . . . 34
4.9. Quantization Table Set . . . . . . . . . . . . . . . . . 34
4.9.1. quant_tables . . . . . . . . . . . . . . . . . . . . 36
4.9.2. context_count . . . . . . . . . . . . . . . . . . . . 36
5. Restrictions . . . . . . . . . . . . . . . . . . . . . . . . 36
6. Security Considerations . . . . . . . . . . . . . . . . . . . 36
7. Media Type Definition . . . . . . . . . . . . . . . . . . . . 37
8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 39
9. Appendixes . . . . . . . . . . . . . . . . . . . . . . . . . 39
9.1. Decoder implementation suggestions . . . . . . . . . . . 39
9.1.1. Multi-threading Support and Independence of Slices . 39
10. Changelog . . . . . . . . . . . . . . . . . . . . . . . . . . 39
11. References . . . . . . . . . . . . . . . . . . . . . . . . . 39
11.1. Normative References . . . . . . . . . . . . . . . . . . 39
11.2. Informative References . . . . . . . . . . . . . . . . . 40
Niedermayer, et al. Expires March 29, 2019 [Page 3]
Internet-Draft FFV1 September 2018
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 41
1. Introduction
This document describes FFV1, a lossless video encoding format. The
design of FFV1 considers the storage of image characteristics, data
fixity, and the optimized use of encoding time and storage
requirements. FFV1 is designed to support a wide range of lossless
video applications such as long-term audiovisual preservation,
scientific imaging, screen recording, and other video encoding
scenarios that seek to avoid the generational loss of lossy video
encodings.
This document defines a version 4 of FFV1. Prior versions of FFV1
are defined within [I-D.ietf-cellar-ffv1].
The latest version of this document is available at
<https://raw.github.com/FFmpeg/FFV1/master/ffv1.md>
This document assumes familiarity with mathematical and coding
concepts such as Range coding [range-coding] and YCbCr color spaces
[YCbCr].
2. Notation and Conventions
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC2119].
2.1. Definitions
"Frame": An encoded representation of a complete static image.
"Slice": A spatial sub-section of a "Frame" that is encoded
separately from an other region of the same frame.
"Container": Format that encapsulates "Frames" and (when required) a
"Configuration Record" into a bitstream.
"Sample": The smallest addressable representation of a color
component or a luma component in a "Frame". Examples of sample are
Luma, Blue Chrominance, Red Chrominance, Alpha, Red, Green, and Blue.
"Pixel": The smallest addressable representation of a color in a
"Frame". It is composed of 1 or more samples.
"ESC": An ESCape symbol to indicate that the symbol to be stored is
too large for normal storage and that an alternate storage method.
Niedermayer, et al. Expires March 29, 2019 [Page 4]
Internet-Draft FFV1 September 2018
"MSB": Most Significant Bit, the bit that can cause the largest
change in magnitude of the symbol.
"RCT": Reversible Color Transform, a near linear, exactly reversible
integer transform that converts between RGB and YCbCr representations
of a Pixel.
"VLC": Variable Length Code, a code that maps source symbols to a
variable number of bits.
"RGB": A reference to the method of storing the value of a Pixel by
using three numeric values that represent Red, Green, and Blue.
"YCbCr": A reference to the method of storing the value of a Pixel by
using three numeric values that represent the luma of the Pixel (Y)
and the chrominance of the Pixel (Cb and Cr). YCbCr word is used for
historical reasons and currently references any color space relying
on 1 luma sample and 2 chrominance samples e.g. YCbCr, YCgCo or
ICtCp. Exact meaning of the three numeric values is unspecified.
"TBA": To Be Announced. Used in reference to the development of
future iterations of the FFV1 specification.
2.2. Conventions
2.2.1. Pseudo-code
The FFV1 bitstream is described in this document using pseudo-code.
Note that the pseudo-code is used for clarity in order to illustrate
the structure of FFV1 and not intended to specify any particular
implementation. The pseudo-code used is based upon the C programming
language [ISO.9899.1990] and uses its "if/else", "while" and "for"
functions as well as functions defined within this document.
2.2.2. Arithmetic Operators
Note: the operators and the order of precedence are the same as used
in the C programming language [ISO.9899.1990].
"a + b" means a plus b.
"a - b" means a minus b.
"-a" means negation of a.
"a * b" means a multiplied by b.
"a / b" means a divided by b.
Niedermayer, et al. Expires March 29, 2019 [Page 5]
Internet-Draft FFV1 September 2018
"a & b" means bit-wise "and" of a and b.
"a | b" means bit-wise "or" of a and b.
"a >> b" means arithmetic right shift of two's complement integer
representation of a by b binary digits.
"a << b" means arithmetic left shift of two's complement integer
representation of a by b binary digits.
2.2.3. Assignment Operators
"a = b" means a is assigned b.
"a++" is equivalent to a is assigned a + 1.
"a--" is equivalent to a is assigned a - 1.
"a += b" is equivalent to a is assigned a + b.
"a -= b" is equivalent to a is assigned a - b.
"a *= b" is equivalent to a is assigned a * b.
2.2.4. Comparison Operators
"a > b" means a is greater than b.
"a >= b" means a is greater than or equal to b.
"a < b" means a is less than b.
"a <= b" means a is less than or equal b.
"a == b" means a is equal to b.
"a != b" means a is not equal to b.
"a && b" means Boolean logical "and" of a and b.
"a || b" means Boolean logical "or" of a and b.
"!a" means Boolean logical "not" of a.
"a ? b : c" if a is true, then b, otherwise c.
Niedermayer, et al. Expires March 29, 2019 [Page 6]
Internet-Draft FFV1 September 2018
2.2.5. Mathematical Functions
floor(a) the largest integer less than or equal to a
ceil(a) the smallest integer greater than or equal to a
sign(a) extracts the sign of a number, i.e. if a < 0 then -1, else if
a > 0 then 1, else 0
abs(a) the absolute value of a, i.e. abs(a) = sign(a)*a
log2(a) the base-two logarithm of a
min(a,b) the smallest of two values a and b
max(a,b) the largest of two values a and b
median(a,b,c) the numerical middle value in a data set of a, b, and
c, i.e. a+b+c-min(a,b,c)-max(a,b,c)
a_{b} the b-th value of a sequence of a
a_{b,c} the 'b,c'-th value of a sequence of a
2.2.6. Order of Operation Precedence
When order of precedence is not indicated explicitly by use of
parentheses, operations are evaluated in the following order (from
top to bottom, operations of same precedence being evaluated from
left to right). This order of operations is based on the order of
operations used in Standard C.
a++, a--
!a, -a
a * b, a / b, a % b
a + b, a - b
a << b, a >> b
a < b, a <= b, a > b, a >= b
a == b, a != b
a & b
a | b
a && b
a || b
a ? b : c
a = b, a += b, a -= b, a *= b
Niedermayer, et al. Expires March 29, 2019 [Page 7]
Internet-Draft FFV1 September 2018
2.2.7. Range
"a...b" means any value starting from a to b, inclusive.
2.2.8. NumBytes
"NumBytes" is a non-negative integer that expresses the size in 8-bit
octets of particular FFV1 "Configuration Record" or "Frame". FFV1
relies on its "Container" to store the "NumBytes" values, see
Section 4.2.3.
2.2.9. Bitstream Functions
2.2.9.1. remaining_bits_in_bitstream
"remaining_bits_in_bitstream( )" means the count of remaining bits
after the pointer in that "Configuration Record" or "Frame". It is
computed from the "NumBytes" value multiplied by 8 minus the count of
bits of that "Configuration Record" or "Frame" already read by the
bitstream parser.
2.2.9.2. byte_aligned
"byte_aligned( )" is true if "remaining_bits_in_bitstream( NumBytes
)" is a multiple of 8, otherwise false.
2.2.9.3. get_bits
"get_bits( i )" is the action to read the next "i" bits in the
bitstream, from most significant bit to least significant bit, and to
return the corresponding value. The pointer is increased by "i".
3. General Description
Samples within a plane are coded in raster scan order (left->right,
top->bottom). Each sample is predicted by the median predictor from
samples in the same plane and the difference is stored see
Section 3.8.
3.1. Border
A border is assumed for each coded slice for the purpose of the
predictor and context according to the following rules:
o one column of samples to the left of the coded slice is assumed as
identical to the samples of the leftmost column of the coded slice
shifted down by one row. The value of the topmost sample of the
Niedermayer, et al. Expires March 29, 2019 [Page 8]
Internet-Draft FFV1 September 2018
column of samples to the left of the coded slice is assumed to be
"0"
o one column of samples to the right of the coded slice is assumed
as identical to the samples of the rightmost column of the coded
slice
o an additional column of samples to the left of the coded slice and
two rows of samples above the coded slice are assumed to be "0"
The following table depicts a slice of samples "a,b,c,d,e,f,g,h,i"
along with its assumed border.
+---+---+---+---+---+---+---+---+
| 0 | 0 | | 0 | 0 | 0 | | 0 |
+---+---+---+---+---+---+---+---+
| 0 | 0 | | 0 | 0 | 0 | | 0 |
+---+---+---+---+---+---+---+---+
| | | | | | | | |
+---+---+---+---+---+---+---+---+
| 0 | 0 | | a | b | c | | c |
+---+---+---+---+---+---+---+---+
| 0 | a | | d | e | f | | f |
+---+---+---+---+---+---+---+---+
| 0 | d | | g | h | i | | i |
+---+---+---+---+---+---+---+---+
3.2. Samples
Positions used for context and median predictor are:
+---+---+---+---+
| | | T | |
+---+---+---+---+
| |tl | t |tr |
+---+---+---+---+
| L | l | X | |
+---+---+---+---+
"X" is the current processed Sample. The identifiers are made of the
first letters of the words Top, Left and Right.
3.3. Median Predictor
The prediction for any sample value at position "X" may be computed
based upon the relative neighboring values of "l", "t", and "tl" via
this equation:
Niedermayer, et al. Expires March 29, 2019 [Page 9]
Internet-Draft FFV1 September 2018
"median(l, t, l + t - tl)".
Note, this prediction template is also used in [ISO.14495-1.1999] and
[HuffYUV].
Exception for the media predictor: if "colorspace_type == 0 &&
bits_per_raw_sample == 16 && ( coder_type == 1 || coder_type == 2 )",
the following media predictor MUST be used:
"median(left16s, top16s, left16s + top16s - diag16s)"
where:
left16s = l >= 32768 ? ( l - 65536 ) : l
top16s = t >= 32768 ? ( t - 65536 ) : t
diag16s = tl >= 32768 ? ( tl - 65536 ) : tl
Background: a two's complement signed 16-bit signed integer was used
for storing sample values in all known implementations of FFV1
bitstream. So in some circumstances, the most significant bit was
wrongly interpreted (used as a sign bit instead of the 16th bit of an
unsigned integer). Note that when the issue is discovered, the only
configuration of all known implementations being impacted is 16-bit
YCbCr with no Pixel transformation with Range Coder coder, as other
potentially impacted configurations (e.g. 15/16-bit JPEG2000-RCT with
Range Coder coder, or 16-bit content with Golomb Rice coder) were
implemented nowhere [ISO.15444-1.2016]. In the meanwhile, 16-bit
JPEG2000-RCT with Range Coder coder was implemented without this
issue in one implementation and validated by one conformance checker.
It is expected (to be confirmed) to remove this exception for the
media predictor in the next version of the FFV1 bitstream.
3.4. Context
Relative to any sample "X", the Quantized Sample Differences "L-l",
"l-tl", "tl-t", "T-t", and "t-tr" are used as context:
context = Q_{0}[l - tl] +
Q_{1}[tl - t] +
Q_{2}[t - tr] +
Q_{3}[L - l] +
Q_{4}[T - t]
If "context >= 0" then "context" is used and the difference between
the sample and its predicted value is encoded as is, else "-context"
is used and the difference between the sample and its predicted value
is encoded with a flipped sign.
Niedermayer, et al. Expires March 29, 2019 [Page 10]
Internet-Draft FFV1 September 2018
3.5. Quantization Table Sets
The FFV1 bitstream contains 1 or more Quantization Table Sets. Each
Quantization Table Set contains exactly 5 Quantization Tables, each
Quantization Table corresponding to 1 of the 5 Quantized Sample
Differences. For each Quantization Table, both the number of
quantization steps and their distribution are stored in the FFV1
bitstream; each Quantization Table has exactly 256 entries, and the 8
least significant bits of the Quantized Sample Difference are used as
index:
Q_{j}[k] = quant_tables[i][j][k&255]
In this formula, "i" is the Quantization Table Set index, "j" is the
Quantized Table index, "k" the Quantized Sample Difference.
3.6. Quantization Table Set Indexes
For each plane of each slice, a Quantization Table Set is selected
from an index:
o For Y plane, "quant_table_set_index [ 0 ]" index is used
o For Cb and Cr planes, "quant_table_set_index [ 1 ]" index is used
o For Alpha plane, "quant_table_set_index [ (version <= 3 ||
chroma_planes) ? 2 : 1 ]" index is used
Background: in first implementations of FFV1 bitstream, the index for
Cb and Cr planes was stored even if it is not used (chroma_planes set
to 0), this index is kept for version <= 3 in order to keep
compatibility with FFV1 bitstreams in the wild.
3.7. Color spaces
FFV1 supports two color spaces: YCbCr and RGB. Both color spaces
allow an optional Alpha plane that can be used to code transparency
data.
3.7.1. YCbCr
In YCbCr color space, the Cb and Cr planes are optional, but if used
then MUST be used together. Omitting the Cb and Cr planes codes the
frames in grayscale without color data. An FFV1 "Frame" using YCbCr
MUST use one of the following arrangements:
o Y
Niedermayer, et al. Expires March 29, 2019 [Page 11]
Internet-Draft FFV1 September 2018
o Y, Alpha
o Y, Cb, Cr
o Y, Cb, Cr, Alpha
The Y plane MUST be coded first. If the Cb and Cr planes are used
then they MUST be coded after the Y plane. If an Alpha
(transparency) plane is used, then it MUST be coded last.
3.7.2. RGB
JPEG2000-RCT is a Reversible Color Transform that codes RGB (red,
green, blue) planes losslessly in a modified YCbCr color space
[ISO.15444-1.2016]. Reversible Pixel transformations between YCbCr
and RGB use the following formulae.
Cb=b-g
Cr=r-g
Y=g+(Cb+Cr)>>2
g=Y-(Cb+Cr)>>2
r=Cr+g
b=Cb+g
Exception for the JPEG2000-RCT conversion: if bits_per_raw_sample is
between 9 and 15 inclusive and alpha_plane is 0, the following
formulae for reversible conversions between YCbCr and RGB MUST be
used instead of the ones above:
Cb=g-b
Cr=r-b
Y=b+(Cb+Cr)>>2
b=Y-(Cb+Cr)>>2
r=Cr+b
g=Cb+b
Background: At the time of this writing, in all known implementations
of FFV1 bitstream, when bits_per_raw_sample was between 9 and 15
Niedermayer, et al. Expires March 29, 2019 [Page 12]
Internet-Draft FFV1 September 2018
inclusive and alpha_plane is 0, GBR planes were used as BGR planes
during both encoding and decoding. In the meanwhile, 16-bit
JPEG2000-RCT was implemented without this issue in one implementation
and validated by one conformance checker. Methods to address this
exception for the transform are under consideration for the next
version of the FFV1 bitstream.
When FFV1 uses the JPEG2000-RCT, the horizontal lines are interleaved
to improve caching efficiency since it is most likely that the RCT
will immediately be converted to RGB during decoding. The
interleaved coding order is also Y, then Cb, then Cr, and then if
used Alpha.
As an example, a "Frame" that is two pixels wide and two pixels high,
could be comprised of the following structure:
+------------------------+------------------------+
| Pixel[1,1] | Pixel[2,1] |
| Y[1,1] Cb[1,1] Cr[1,1] | Y[2,1] Cb[2,1] Cr[2,1] |
+------------------------+------------------------+
| Pixel[1,2] | Pixel[2,2] |
| Y[1,2] Cb[1,2] Cr[1,2] | Y[2,2] Cb[2,2] Cr[2,2] |
+------------------------+------------------------+
In JPEG2000-RCT, the coding order would be left to right and then top
to bottom, with values interleaved by lines and stored in this order:
Y[1,1] Y[2,1] Cb[1,1] Cb[2,1] Cr[1,1] Cr[2,1] Y[1,2] Y[2,2] Cb[1,2]
Cb[2,2] Cr[1,2] Cr[2,2]
3.8. Coding of the Sample Difference
Instead of coding the n+1 bits of the Sample Difference with Huffman
or Range coding (or n+2 bits, in the case of RCT), only the n (or
n+1) least significant bits are used, since this is sufficient to
recover the original sample. In the equation below, the term "bits"
represents bits_per_raw_sample+1 for RCT or bits_per_raw_sample
otherwise:
coder_input =
[(sample_difference + 2^(bits-1)) & (2^bits - 1)] - 2^(bits-1)
3.8.1. Range Coding Mode
Early experimental versions of FFV1 used the CABAC Arithmetic coder
from H.264 as defined in [ISO.14496-10.2014] but due to the uncertain
patent/royalty situation, as well as its slightly worse performance,
Niedermayer, et al. Expires March 29, 2019 [Page 13]
Internet-Draft FFV1 September 2018
CABAC was replaced by a Range coder based on an algorithm defined by
G. Nigel and N. Martin in 1979 [range-coding].
3.8.1.1. Range Binary Values
To encode binary digits efficiently a Range coder is used. "C_{i}"
is the i-th Context. "B_{i}" is the i-th byte of the bytestream.
"b_{i}" is the i-th Range coded binary value, "S_{0,i}" is the i-th
initial state, which is 128. The length of the bytestream encoding n
binary symbols is "j_{n}" bytes.
r_{i} = floor( ( R_{i} * S_{i,C_{i}} ) / 2^8 )
S_{i+1,C_{i}} = zero_state_{S_{i,C_{i}}} XOR
l_i = L_i XOR
t_i = R_i - r_i <==
b_i = 0 <==>
L_i < R_i - r_i
S_{i+1,C_{i}} = one_state_{S_{i,C_{i}}} XOR
l_i = L_i - R_i + r_i XOR
t_i = r_i <==
b_i = 1 <==>
L_i >= R_i - r_i
S_{i+1,k} = S_{i,k} <== C_i != k
R_{i+1} = 2^8 * t_{i} XOR
L_{i+1} = 2^8 * l_{i} + B_{j_{i}} XOR
j_{i+1} = j_{i} + 1 <==
t_{i} < 2^8
R_{i+1} = t_{i} XOR
L_{i+1} = l_{i} XOR
j_{i+1} = j_{i} <==
t_{i} >= 2^8
R_{0} = 65280
L_{0} = 2^8 * B_{0} + B_{1}
j_{0} = 2
3.8.1.2. Range Non Binary Values
To encode scalar integers, it would be possible to encode each bit
separately and use the past bits as context. However that would mean
255 contexts per 8-bit symbol that is not only a waste of memory but
Niedermayer, et al. Expires March 29, 2019 [Page 14]
Internet-Draft FFV1 September 2018
also requires more past data to reach a reasonably good estimate of
the probabilities. Alternatively assuming a Laplacian distribution
and only dealing with its variance and mean (as in Huffman coding)
would also be possible, however, for maximum flexibility and
simplicity, the chosen method uses a single symbol to encode if a
number is 0 and if not encodes the number using its exponent,
mantissa and sign. The exact contexts used are best described by the
following code, followed by some comments.
pseudo-code | type
--------------------------------------------------------------|-----
void put_symbol(RangeCoder *c, uint8_t *state, int v, int \ |
is_signed) { |
int i; |
put_rac(c, state+0, !v); |
if (v) { |
int a= abs(v); |
int e= log2(a); |
|
for (i=0; i<e; i++) |
put_rac(c, state+1+min(i,9), 1); //1..10 |
|
put_rac(c, state+1+min(i,9), 0); |
for (i=e-1; i>=0; i--) |
put_rac(c, state+22+min(i,9), (a>>i)&1); //22..31 |
|
if (is_signed) |
put_rac(c, state+11 + min(e, 10), v < 0); //11..21|
} |
} |
3.8.1.3. Initial Values for the Context Model
At keyframes all Range coder state variables are set to their initial
state.
3.8.1.4. State Transition Table
one_state_{i} =
default_state_transition_{i} + state_transition_delta_{i}
zero_state_{i} = 256 - one_state_{256-i}
3.8.1.5. default_state_transition
Niedermayer, et al. Expires March 29, 2019 [Page 15]
Internet-Draft FFV1 September 2018
0, 0, 0, 0, 0, 0, 0, 0, 20, 21, 22, 23, 24, 25, 26, 27,
28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 37, 38, 39, 40, 41, 42,
43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 56, 57,
58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73,
74, 75, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88,
89, 90, 91, 92, 93, 94, 94, 95, 96, 97, 98, 99,100,101,102,103,
104,105,106,107,108,109,110,111,112,113,114,114,115,116,117,118,
119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,133,
134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,
150,151,152,152,153,154,155,156,157,158,159,160,161,162,163,164,
165,166,167,168,169,170,171,171,172,173,174,175,176,177,178,179,
180,181,182,183,184,185,186,187,188,189,190,190,191,192,194,194,
195,196,197,198,199,200,201,202,202,204,205,206,207,208,209,209,
210,211,212,213,215,215,216,217,218,219,220,220,222,223,224,225,
226,227,227,229,229,230,231,232,234,234,235,236,237,238,239,240,
241,242,243,244,245,246,247,248,248, 0, 0, 0, 0, 0, 0, 0,
3.8.1.6. Alternative State Transition Table
The alternative state transition table has been built using iterative
minimization of frame sizes and generally performs better than the
default. To use it, the coder_type MUST be set to 2 and the
difference to the default MUST be stored in the "Parameters", see
Section 4.1. The reference implementation of FFV1 in FFmpeg uses
this table by default at the time of this writing when Range coding
is used.
Niedermayer, et al. Expires March 29, 2019 [Page 16]
Internet-Draft FFV1 September 2018
0, 10, 10, 10, 10, 16, 16, 16, 28, 16, 16, 29, 42, 49, 20, 49,
59, 25, 26, 26, 27, 31, 33, 33, 33, 34, 34, 37, 67, 38, 39, 39,
40, 40, 41, 79, 43, 44, 45, 45, 48, 48, 64, 50, 51, 52, 88, 52,
53, 74, 55, 57, 58, 58, 74, 60,101, 61, 62, 84, 66, 66, 68, 69,
87, 82, 71, 97, 73, 73, 82, 75,111, 77, 94, 78, 87, 81, 83, 97,
85, 83, 94, 86, 99, 89, 90, 99,111, 92, 93,134, 95, 98,105, 98,
105,110,102,108,102,118,103,106,106,113,109,112,114,112,116,125,
115,116,117,117,126,119,125,121,121,123,145,124,126,131,127,129,
165,130,132,138,133,135,145,136,137,139,146,141,143,142,144,148,
147,155,151,149,151,150,152,157,153,154,156,168,158,162,161,160,
172,163,169,164,166,184,167,170,177,174,171,173,182,176,180,178,
175,189,179,181,186,183,192,185,200,187,191,188,190,197,193,196,
197,194,195,196,198,202,199,201,210,203,207,204,205,206,208,214,
209,211,221,212,213,215,224,216,217,218,219,220,222,228,223,225,
226,224,227,229,240,230,231,232,233,234,235,236,238,239,237,242,
241,243,242,244,245,246,247,248,249,250,251,252,252,253,254,255,
3.8.2. Golomb Rice Mode
This coding mode uses Golomb Rice codes. The VLC is split into 2
parts, the prefix stores the most significant bits and the suffix
stores the k least significant bits or stores the whole number in the
ESC case. The end of the bitstream of the "Frame" is filled with
0-bits until that the bitstream contains a multiple of 8 bits.
3.8.2.1. Prefix
Niedermayer, et al. Expires March 29, 2019 [Page 17]
Internet-Draft FFV1 September 2018
+----------------+-------+
| bits | value |
+----------------+-------+
| 1 | 0 |
| 01 | 1 |
| ... | ... |
| 0000 0000 0001 | 11 |
| 0000 0000 0000 | ESC |
+----------------+-------+
3.8.2.2. Suffix
+-------+-----------------------------------------------------------+
| non | the k least significant bits MSB first |
| ESC | |
| ESC | the value - 11, in MSB first order, ESC may only be used |
| | if the value cannot be coded as non ESC |
+-------+-----------------------------------------------------------+
3.8.2.3. Examples
+-----+-------------------------+-------+
| k | bits | value |
+-----+-------------------------+-------+
| 0 | "1" | 0 |
| 0 | "001" | 2 |
| 2 | "1 00" | 0 |
| 2 | "1 10" | 2 |
| 2 | "01 01" | 5 |
| any | "000000000000 10000000" | 139 |
+-----+-------------------------+-------+
3.8.2.4. Run Mode
Run mode is entered when the context is 0 and left as soon as a non-0
difference is found. The level is identical to the predicted one.
The run and the first different level are coded.
3.8.2.5. Run Length Coding
The run value is encoded in 2 parts, the prefix part stores the more
significant part of the run as well as adjusting the run_index that
determines the number of bits in the less significant part of the
run. The 2nd part of the value stores the less significant part of
the run as it is. The run_index is reset for each plane and slice to
0.
Niedermayer, et al. Expires March 29, 2019 [Page 18]
Internet-Draft FFV1 September 2018
pseudo-code | type
--------------------------------------------------------------|-----
log2_run[41]={ |
0, 0, 0, 0, 1, 1, 1, 1, |
2, 2, 2, 2, 3, 3, 3, 3, |
4, 4, 5, 5, 6, 6, 7, 7, |
8, 9,10,11,12,13,14,15, |
16,17,18,19,20,21,22,23, |
24, |
}; |
|
if (run_count == 0 && run_mode == 1) { |
if (get_bits(1)) { |
run_count = 1 << log2_run[run_index]; |
if (x + run_count <= w) |
run_index++; |
} else { |
if (log2_run[run_index]) |
run_count = get_bits(log2_run[run_index]); |
else |
run_count = 0; |
if (run_index) |
run_index--; |
run_mode = 2; |
} |
} |
The log2_run function is also used within [ISO.14495-1.1999].
3.8.2.6. Level Coding
Level coding is identical to the normal difference coding with the
exception that the 0 value is removed as it cannot occur:
if (diff>0) diff--;
encode(diff);
Note, this is different from JPEG-LS, which doesn't use prediction in
run mode and uses a different encoding and context model for the last
difference On a small set of test samples the use of prediction
slightly improved the compression rate.
4. Bitstream
Niedermayer, et al. Expires March 29, 2019 [Page 19]
Internet-Draft FFV1 September 2018
+--------+----------------------------------------------------------+
| Symbol | Definition |
+--------+----------------------------------------------------------+
| u(n) | unsigned big endian integer using n bits |
| sg | Golomb Rice coded signed scalar symbol coded with the |
| | method described in Section 3.8.2 |
| br | Range coded Boolean (1-bit) symbol with the method |
| | described in Section 3.8.1.1 |
| ur | Range coded unsigned scalar symbol coded with the method |
| | described in Section 3.8.1.2 |
| sr | Range coded signed scalar symbol coded with the method |
| | described in Section 3.8.1.2 |
+--------+----------------------------------------------------------+
The same context that is initialized to 128 is used for all fields in
the header.
The following MUST be provided by external means during
initialization of the decoder:
"frame_pixel_width" is defined as "Frame" width in pixels.
"frame_pixel_height" is defined as "Frame" height in pixels.
Default values at the decoder initialization phase:
"ConfigurationRecordIsPresent" is set to 0.
4.1. Parameters
The "Parameters" section contains significant characteristics used
for all instances of "Frame". The pseudo-code below describes the
contents of the bitstream.
Niedermayer, et al. Expires March 29, 2019 [Page 20]
Internet-Draft FFV1 September 2018
pseudo-code | type
--------------------------------------------------------------|-----
Parameters( ) { |
version | ur
if (version >= 3) |
micro_version | ur
coder_type | ur
if (coder_type > 1) |
for (i = 1; i < 256; i++) |
state_transition_delta[ i ] | sr
colorspace_type | ur
if (version >= 1) |
bits_per_raw_sample | ur
chroma_planes | br
log2_h_chroma_subsample | ur
log2_v_chroma_subsample | ur
alpha_plane | br
if (version >= 3) { |
num_h_slices - 1 | ur
num_v_slices - 1 | ur
quant_table_set_count | ur
} |
for( i = 0; i < quant_table_set_count; i++ ) |
QuantizationTableSet( i ) |
if (version >= 3) { |
for( i = 0; i < quant_table_set_count; i++ ) { |
states_coded | br
if (states_coded) |
for( j = 0; j < context_count[ i ]; j++ ) |
for( k = 0; k < CONTEXT_SIZE; k++ ) |
initial_state_delta[ i ][ j ][ k ] | sr
} |
ec | ur
intra | ur
} |
} |
4.1.1. version
"version" specifies the version of the FFV1 bitstream.
Each version is incompatible with others versions: decoders SHOULD
reject a file due to unknown version.
Decoders SHOULD reject a file with version <= 1 &&
ConfigurationRecordIsPresent == 1.
Decoders SHOULD reject a file with version >= 3 &&
ConfigurationRecordIsPresent == 0.
Niedermayer, et al. Expires March 29, 2019 [Page 21]
Internet-Draft FFV1 September 2018
+-------+-------------------------+
| value | version |
+-------+-------------------------+
| 0 | FFV1 version 0 |
| 1 | FFV1 version 1 |
| 2 | reserved* |
| 3 | FFV1 version 3 |
| 4 | FFV1 version 4 |
| Other | reserved for future use |
+-------+-------------------------+
* Version 2 was never enabled in the encoder thus version 2 files
SHOULD NOT exist, and this document does not describe them to keep
the text simpler.
4.1.2. micro_version
"micro_version" specifies the micro-version of the FFV1 bitstream.
After a version is considered stable (a micro-version value is
assigned to be the first stable variant of a specific version), each
new micro-version after this first stable variant is compatible with
the previous micro-version: decoders SHOULD NOT reject a file due to
an unknown micro-version equal or above the micro-version considered
as stable.
Meaning of micro_version for version 3:
+-------+-------------------------+
| value | micro_version |
+-------+-------------------------+
| 0...3 | reserved* |
| 4 | first stable variant |
| Other | reserved for future use |
+-------+-------------------------+
* development versions may be incompatible with the stable variants.
Meaning of micro_version for version 4 (note: at the time of writing
of this specification, version 4 is not considered stable so the
first stable version value is to be announced in the future):
+---------+-------------------------+
| value | micro_version |
+---------+-------------------------+
| 0...TBA | reserved* |
| TBA | first stable variant |
| Other | reserved for future use |
+---------+-------------------------+
Niedermayer, et al. Expires March 29, 2019 [Page 22]
Internet-Draft FFV1 September 2018
* development versions which may be incompatible with the stable
variants.
4.1.3. coder_type
"coder_type" specifies the coder used.
+-------+-------------------------------------------------+
| value | coder used |
+-------+-------------------------------------------------+
| 0 | Golomb Rice |
| 1 | Range Coder with default state transition table |
| 2 | Range Coder with custom state transition table |
| Other | reserved for future use |
+-------+-------------------------------------------------+
4.1.4. state_transition_delta
"state_transition_delta" specifies the Range coder custom state
transition table.
If state_transition_delta is not present in the FFV1 bitstream, all
Range coder custom state transition table elements are assumed to be
0.
4.1.5. colorspace_type
"colorspace_type" specifies color space losslessly encoded, Pixel
transformation used by the encoder, as well as interleave method.
+-------+---------------------+------------------+------------------+
| value | color space | transformation | interleave |
| | losslessly encoded | | method |
+-------+---------------------+------------------+------------------+
| 0 | YCbCr | No Pixel | plane then line |
| | | transformation | |
| 1 | RGB | JPEG2000-RCT | line then plane |
| Other | reserved for future | reserved for | reserved for |
| | use | future use | future use |
+-------+---------------------+------------------+------------------+
Restrictions:
If "colorspace_type" is 1, then "chroma_planes" MUST be 1,
"log2_h_chroma_subsample" MUST be 0, and "log2_v_chroma_subsample"
MUST be 0.
Niedermayer, et al. Expires March 29, 2019 [Page 23]
Internet-Draft FFV1 September 2018
4.1.6. chroma_planes
"chroma_planes" indicates if chroma (color) planes are present.
+-------+-------------------------------+
| value | presence |
+-------+-------------------------------+
| 0 | chroma planes are not present |
| 1 | chroma planes are present |
+-------+-------------------------------+
4.1.7. bits_per_raw_sample
"bits_per_raw_sample" indicates the number of bits for each sample.
Inferred to be 8 if not present.
+-------+---------------------------------+
| value | bits for each sample |
+-------+---------------------------------+
| 0 | reserved* |
| Other | the actual bits for each sample |
+-------+---------------------------------+
* Encoders MUST NOT store bits_per_raw_sample = 0 Decoders SHOULD
accept and interpret bits_per_raw_sample = 0 as 8.
4.1.8. log2_h_chroma_subsample
"log2_h_chroma_subsample" indicates the subsample factor, stored in
powers to which the number 2 must be raised, between luma and chroma
width ("chroma_width = 2^(-log2_h_chroma_subsample) * luma_width").
4.1.9. log2_v_chroma_subsample
"log2_v_chroma_subsample" indicates the subsample factor, stored in
powers to which the number 2 must be raised, between luma and chroma
height ("chroma_height=2^(-log2_v_chroma_subsample) * luma_height").
4.1.10. alpha_plane
"alpha_plane" indicates if a transparency plane is present.
+-------+-----------------------------------+
| value | presence |
+-------+-----------------------------------+
| 0 | transparency plane is not present |
| 1 | transparency plane is present |
+-------+-----------------------------------+
Niedermayer, et al. Expires March 29, 2019 [Page 24]
Internet-Draft FFV1 September 2018
4.1.11. num_h_slices
"num_h_slices" indicates the number of horizontal elements of the
slice raster.
Inferred to be 1 if not present.
4.1.12. num_v_slices
"num_v_slices" indicates the number of vertical elements of the slice
raster.
Inferred to be 1 if not present.
4.1.13. quant_table_set_count
"quant_table_set_count" indicates the number of Quantization
Table Sets.
Inferred to be 1 if not present.
MUST NOT be 0.
4.1.14. states_coded
"states_coded" indicates if the respective Quantization Table Set has
the initial states coded.
Inferred to be 0 if not present.
+-------+-----------------------------------------------------------+
| value | initial states |
+-------+-----------------------------------------------------------+
| 0 | initial states are not present and are assumed to be all |
| | 128 |
| 1 | initial states are present |
+-------+-----------------------------------------------------------+
4.1.15. initial_state_delta
"initial_state_delta[ i ][ j ][ k ]" indicates the initial Range
coder state, it is encoded using "k" as context index and
pred = j ? initial_states[ i ][j - 1][ k ] : 128
initial_state[ i ][ j ][ k ] =
( pred + initial_state_delta[ i ][ j ][ k ] ) & 255
4.1.16. ec
"ec" indicates the error detection/correction type.
Niedermayer, et al. Expires March 29, 2019 [Page 25]
Internet-Draft FFV1 September 2018
+-------+--------------------------------------------+
| value | error detection/correction type |
+-------+--------------------------------------------+
| 0 | 32-bit CRC on the global header |
| 1 | 32-bit CRC per slice and the global header |
| Other | reserved for future use |
+-------+--------------------------------------------+
4.1.17. intra
"intra" indicates the relationship between the instances of "Frame".
Inferred to be 0 if not present.
+-------+-----------------------------------------------------------+
| value | relationship |
+-------+-----------------------------------------------------------+
| 0 | Frames are independent or dependent (keyframes and non |
| | keyframes) |
| 1 | Frames are independent (keyframes only) |
| Other | reserved for future use |
+-------+-----------------------------------------------------------+
4.2. Configuration Record
In the case of a FFV1 bitstream with "version >= 3", a "Configuration
Record" is stored in the underlying "Container", at the track header
level. It contains the "Parameters" used for all instances of
"Frame". The size of the "Configuration Record", "NumBytes", is
supplied by the underlying "Container".
pseudo-code | type
--------------------------------------------------------------|-----
ConfigurationRecord( NumBytes ) { |
ConfigurationRecordIsPresent = 1 |
Parameters( ) |
while( remaining_bits_in_bitstream( NumBytes ) > 32 ) |
reserved_for_future_use | u(1)
configuration_record_crc_parity | u(32)
} |
4.2.1. reserved_for_future_use
"reserved_for_future_use" has semantics that are reserved for future
use.
Encoders conforming to this version of this specification SHALL NOT
write this value.
Decoders conforming to this version of this specification SHALL
ignore its value.
Niedermayer, et al. Expires March 29, 2019 [Page 26]
Internet-Draft FFV1 September 2018
4.2.2. configuration_record_crc_parity
"configuration_record_crc_parity" 32 bits that are chosen so that the
"Configuration Record" as a whole has a crc remainder of 0.
This is equivalent to storing the crc remainder in the 32-bit parity.
The CRC generator polynomial used is the standard IEEE CRC polynomial
(0x104C11DB7) with initial value 0.
4.2.3. Mapping FFV1 into Containers
This "Configuration Record" can be placed in any file format
supporting "Configuration Records", fitting as much as possible with
how the file format uses to store "Configuration Records". The
"Configuration Record" storage place and "NumBytes" are currently
defined and supported by this version of this specification for the
following formats:
4.2.3.1. AVI File Format
The "Configuration Record" extends the stream format chunk ("AVI ",
"hdlr", "strl", "strf") with the ConfigurationRecord bitstream.
See [AVI] for more information about chunks.
"NumBytes" is defined as the size, in bytes, of the strf chunk
indicated in the chunk header minus the size of the stream format
structure.
4.2.3.2. ISO Base Media File Format
The "Configuration Record" extends the sample description box
("moov", "trak", "mdia", "minf", "stbl", "stsd") with a "glbl" box
that contains the ConfigurationRecord bitstream. See
[ISO.14496-12.2015] for more information about boxes.
"NumBytes" is defined as the size, in bytes, of the "glbl" box
indicated in the box header minus the size of the box header.
4.2.3.3. NUT File Format
The codec_specific_data element (in "stream_header" packet) contains
the ConfigurationRecord bitstream. See [NUT] for more information
about elements.
"NumBytes" is defined as the size, in bytes, of the
codec_specific_data element as indicated in the "length" field of
codec_specific_data
Niedermayer, et al. Expires March 29, 2019 [Page 27]
Internet-Draft FFV1 September 2018
4.2.3.4. Matroska File Format
FFV1 SHOULD use "V_FFV1" as the Matroska "Codec ID". For FFV1
versions 2 or less, the Matroska "CodecPrivate" Element SHOULD NOT be
used. For FFV1 versions 3 or greater, the Matroska "CodecPrivate"
Element MUST contain the FFV1 "Configuration Record" structure and no
other data. See [Matroska] for more information about elements.
"NumBytes" is defined as the "Element Data Size" of the
"CodecPrivate" Element.
4.3. Frame
A "Frame" consists of the keyframe field, "Parameters" (if version
<=1), and a sequence of independent slices.
pseudo-code | type
--------------------------------------------------------------|-----
Frame( NumBytes ) { |
keyframe | br
if (keyframe && !ConfigurationRecordIsPresent |
Parameters( ) |
while ( remaining_bits_in_bitstream( NumBytes ) ) |
Slice( ) |
} |
Architecture overview of slices in a "Frame":
+-----------------------------------------------------------------+
| first slice header |
| first slice content |
| first slice footer |
| --------------------------------------------------------------- |
| second slice header |
| second slice content |
| second slice footer |
| --------------------------------------------------------------- |
| ... |
| --------------------------------------------------------------- |
| last slice header |
| last slice content |
| last slice footer |
+-----------------------------------------------------------------+
Niedermayer, et al. Expires March 29, 2019 [Page 28]
Internet-Draft FFV1 September 2018
4.4. Slice
pseudo-code | type
--------------------------------------------------------------|-----
Slice( ) { |
if (version >= 3) |
SliceHeader( ) |
SliceContent( ) |
if (coder_type == 0) |
while (!byte_aligned()) |
padding | u(1)
if (version <= 1) { |
while (remaining_bits_in_bitstream( NumBytes ) != 0 ) |
reserved | u(1)
} |
if (version >= 3) |
SliceFooter( ) |
} |
"padding" specifies a bit without any significance and used only for
byte alignment. MUST be 0.
"reserved" specifies a bit without any significance in this revision
of the specification and may have a significance in a later revision
of this specification.
Encoders SHOULD NOT fill these bits.
Decoders SHOULD ignore these bits.
Note in case these bits are used in a later revision of this
specification: any revision of this specification SHOULD care about
avoiding to add 40 bits of content after "SliceContent" for version 0
and 1 of the bitstream. Background: due to some non conforming
encoders, some bitstreams where found with 40 extra bits
corresponding to "error_status" and "slice_crc_parity", a decoder
conforming to the revised specification could not do the difference
between a revised bitstream and a buggy bitstream.
4.5. Slice Header
Niedermayer, et al. Expires March 29, 2019 [Page 29]
Internet-Draft FFV1 September 2018
pseudo-code | type
--------------------------------------------------------------|-----
SliceHeader( ) { |
slice_x | ur
slice_y | ur
slice_width - 1 | ur
slice_height - 1 | ur
for( i = 0; i < quant_table_set_index_count; i++ ) |
quant_table_set_index [ i ] | ur
picture_structure | ur
sar_num | ur
sar_den | ur
if (version >= 4) { |
reset_contexts | br
slice_coding_mode | ur
} |
} |
4.5.1. slice_x
"slice_x" indicates the x position on the slice raster formed by
num_h_slices.
Inferred to be 0 if not present.
4.5.2. slice_y
"slice_y" indicates the y position on the slice raster formed by
num_v_slices.
Inferred to be 0 if not present.
4.5.3. slice_width
"slice_width" indicates the width on the slice raster formed by
num_h_slices.
Inferred to be 1 if not present.
4.5.4. slice_height
"slice_height" indicates the height on the slice raster formed by
num_v_slices.
Inferred to be 1 if not present.
4.5.5. quant_table_set_index_count
"quant_table_set_index_count" is defined as "1 + ( ( chroma_planes ||
version \<= 3 ) ? 1 : 0 ) + ( alpha_plane ? 1 : 0 )".
Niedermayer, et al. Expires March 29, 2019 [Page 30]
Internet-Draft FFV1 September 2018
4.5.6. quant_table_set_index
"quant_table_set_index" indicates the Quantization Table Set index to
select the Quantization Table Set and the initial states for the
slice.
Inferred to be 0 if not present.
4.5.7. picture_structure
"picture_structure" specifies the temporal and spatial relationship
of each line of the "Frame".
Inferred to be 0 if not present.
+-------+-------------------------+
| value | picture structure used |
+-------+-------------------------+
| 0 | unknown |
| 1 | top field first |
| 2 | bottom field first |
| 3 | progressive |
| Other | reserved for future use |
+-------+-------------------------+
4.5.8. sar_num
"sar_num" specifies the sample aspect ratio numerator.
Inferred to be 0 if not present.
MUST be 0 if sample aspect ratio is unknown.
4.5.9. sar_den
"sar_den" specifies the sample aspect ratio denominator.
Inferred to be 0 if not present.
MUST be 0 if sample aspect ratio is unknown.
4.5.10. reset_contexts
"reset_contexts" indicates if slice contexts must be reset.
Inferred to be 0 if not present.
4.5.11. slice_coding_mode
"slice_coding_mode" indicates the slice coding mode.
Inferred to be 0 if not present.
Niedermayer, et al. Expires March 29, 2019 [Page 31]
Internet-Draft FFV1 September 2018
+-------+-----------------------------+
| value | slice coding mode |
+-------+-----------------------------+
| 0 | Range Coding or Golomb Rice |
| 1 | raw PCM |
| Other | reserved for future use |
+-------+-----------------------------+
4.6. Slice Content
pseudo-code | type
--------------------------------------------------------------|-----
SliceContent( ) { |
if (colorspace_type == 0) { |
for( p = 0; p < primary_color_count; p++ ) |
for( y = 0; y < plane_pixel_height[ p ]; y++ ) |
Line( p, y ) |
} else if (colorspace_type == 1) { |
for( y = 0; y < slice_pixel_height; y++ ) |
for( p = 0; p < primary_color_count; p++ ) |
Line( p, y ) |
} |
} |
4.6.1. primary_color_count
"primary_color_count" is defined as "1 + ( chroma_planes ? 2 : 0 ) +
( alpha_plane ? 1 : 0 )".
4.6.2. plane_pixel_height
"plane_pixel_height[ p ]" is the height in pixels of plane p of the
slice.
"plane_pixel_height[ 0 ]" and "plane_pixel_height[ 1 + (
chroma_planes ? 2 : 0 ) ]" value is "slice_pixel_height".
If "chroma_planes" is set to 1, "plane_pixel_height[ 1 ]" and
"plane_pixel_height[ 2 ]" value is "ceil(slice_pixel_height /
log2_v_chroma_subsample)".
4.6.3. slice_pixel_height
"slice_pixel_height" is the height in pixels of the slice.
Its value is "floor(( slice_y + slice_height ) * slice_pixel_height /
num_v_slices) - slice_pixel_y".
Niedermayer, et al. Expires March 29, 2019 [Page 32]
Internet-Draft FFV1 September 2018
4.6.4. slice_pixel_y
"slice_pixel_y" is the slice vertical position in pixels.
Its value is "floor(slice_y * frame_pixel_height / num_v_slices)".
4.7. Line
pseudo-code | type
--------------------------------------------------------------|-----
Line( p, y ) { |
if (colorspace_type == 0) { |
for( x = 0; x < plane_pixel_width[ p ]; x++ ) |
sample_difference[ p ][ y ][ x ] |
} else if (colorspace_type == 1) { |
for( x = 0; x < slice_pixel_width; x++ ) |
sample_difference[ p ][ y ][ x ] |
} |
} |
4.7.1. plane_pixel_width
"plane_pixel_width[ p ]" is the width in pixels of plane p of the
slice.
"plane_pixel_width[ 0 ]" and "plane_pixel_width[ 1 + ( chroma_planes
? 2 : 0 ) ]" value is "slice_pixel_width".
If "chroma_planes" is set to 1, "plane_pixel_width[ 1 ]" and
"plane_pixel_width[ 2 ]" value is "ceil(slice_pixel_width / (1 <<
log2_h_chroma_subsample))".
4.7.2. slice_pixel_width
"slice_pixel_width" is the width in pixels of the slice.
Its value is "floor(( slice_x + slice_width ) * slice_pixel_width /
num_h_slices) - slice_pixel_x".
4.7.3. slice_pixel_x
"slice_pixel_x" is the slice horizontal position in pixels.
Its value is "floor(slice_x * frame_pixel_width / num_h_slices)".
4.7.4. sample_difference
"sample_difference[ p ][ y ][ x ]" is the sample difference for
sample at plane "p", y position "y", and x position "x". The sample
value is computed based on prediction and context described in
Section 3.2.
Niedermayer, et al. Expires March 29, 2019 [Page 33]
Internet-Draft FFV1 September 2018
4.8. Slice Footer
Note: slice footer is always byte aligned.
pseudo-code | type
--------------------------------------------------------------|-----
SliceFooter( ) { |
slice_size | u(24)
if (ec) { |
error_status | u(8)
slice_crc_parity | u(32)
} |
} |
4.8.1. slice_size
"slice_size" indicates the size of the slice in bytes.
Note: this allows finding the start of slices before previous slices
have been fully decoded, and allows parallel decoding as well as
error resilience.
4.8.2. error_status
"error_status" specifies the error status.
+-------+--------------------------------------+
| value | error status |
+-------+--------------------------------------+
| 0 | no error |
| 1 | slice contains a correctable error |
| 2 | slice contains a uncorrectable error |
| Other | reserved for future use |
+-------+--------------------------------------+
4.8.3. slice_crc_parity
"slice_crc_parity" 32 bits that are chosen so that the slice as a
whole has a crc remainder of 0.
This is equivalent to storing the crc remainder in the 32-bit parity.
The CRC generator polynomial used is the standard IEEE CRC polynomial
(0x104C11DB7) with initial value 0.
4.9. Quantization Table Set
The Quantization Table Sets are stored by storing the number of equal
entries -1 of the first half of the table (represented as "len - 1"
in the pseudo-code below) using the method described in
Section 3.8.1.2. The second half doesn't need to be stored as it is
Niedermayer, et al. Expires March 29, 2019 [Page 34]
Internet-Draft FFV1 September 2018
identical to the first with flipped sign. "scale" and "len_count[ i
][ j ]" are temporary values used for the computing of
"context_count[ i ]" and are not used outside Quantization Table Set
pseudo-code.
example:
Table: 0 0 1 1 1 1 2 2 -2 -2 -2 -1 -1 -1 -1 0
Stored values: 1, 3, 1
pseudo-code | type
--------------------------------------------------------------|-----
QuantizationTableSet( i ) { |
scale = 1 |
for( j = 0; j < MAX_CONTEXT_INPUTS; j++ ) { |
QuantizationTable( i, j, scale ) |
scale *= 2 * len_count[ i ][ j ] - 1 |
} |
context_count[ i ] = ceil ( scale / 2 ) |
} |
MAX_CONTEXT_INPUTS is 5.
pseudo-code | type
--------------------------------------------------------------|-----
QuantizationTable(i, j, scale) { |
v = 0 |
for( k = 0; k < 128; ) { |
len - 1 | ur
for( a = 0; a < len; a++ ) { |
quant_tables[ i ][ j ][ k ] = scale* v |
k++ |
} |
v++ |
} |
for( k = 1; k < 128; k++ ) { |
quant_tables[ i ][ j ][ 256 - k ] = \ |
-quant_tables[ i ][ j ][ k ] |
} |
quant_tables[ i ][ j ][ 128 ] = \ |
-quant_tables[ i ][ j ][ 127 ] |
len_count[ i ][ j ] = v |
} |
Niedermayer, et al. Expires March 29, 2019 [Page 35]
Internet-Draft FFV1 September 2018
4.9.1. quant_tables
"quant_tables[ i ][ j ][ k ]" indicates the quantification table
value of the Quantized Sample Difference "k" of the Quantization
Table "j" of the Set Quantization Table Set "i".
4.9.2. context_count
"context_count[ i ]" indicates the count of contexts for Quantization
Table Set "i".
5. Restrictions
To ensure that fast multithreaded decoding is possible, starting
version 3 and if frame_pixel_width * frame_pixel_height is more than
101376, slice_width * slice_height MUST be less or equal to
num_h_slices * num_v_slices / 4. Note: 101376 is the frame size in
pixels of a 352x288 frame also known as CIF ("Common Intermediate
Format") frame size format.
For each "Frame", each position in the slice raster MUST be filled by
one and only one slice of the "Frame" (no missing slice position, no
slice overlapping).
For each "Frame" with keyframe value of 0, each slice MUST have the
same value of slice_x, slice_y, slice_width, slice_height as a slice
in the previous "Frame", except if reset_contexts is 1.
6. Security Considerations
Like any other codec, (such as [RFC6716]), FFV1 should not be used
with insecure ciphers or cipher-modes that are vulnerable to known
plaintext attacks. Some of the header bits as well as the padding
are easily predictable.
Implementations of the FFV1 codec need to take appropriate security
considerations into account, as outlined in [RFC4732]. It is
extremely important for the decoder to be robust against malicious
payloads. Malicious payloads must not cause the decoder to overrun
its allocated memory or to take an excessive amount of resources to
decode. Although problems in encoders are typically rarer, the same
applies to the encoder. Malicious video streams must not cause the
encoder to misbehave because this would allow an attacker to attack
transcoding gateways. A frequent security problem in image and video
codecs is also to not check for integer overflows in Pixel count
computations, that is to allocate width * height without considering
that the multiplication result may have overflowed the arithmetic
types range.
Niedermayer, et al. Expires March 29, 2019 [Page 36]
Internet-Draft FFV1 September 2018
The reference implementation [REFIMPL] contains no known buffer
overflow or cases where a specially crafted packet or video segment
could cause a significant increase in CPU load.
The reference implementation [REFIMPL] was validated in the following
conditions:
o Sending the decoder valid packets generated by the reference
encoder and verifying that the decoder's output matches the
encoder's input.
o Sending the decoder packets generated by the reference encoder and
then subjected to random corruption.
o Sending the decoder random packets that are not FFV1.
In all of the conditions above, the decoder and encoder was run
inside the [VALGRIND] memory debugger as well as clangs address
sanitizer [Address-Sanitizer], which track reads and writes to
invalid memory regions as well as the use of uninitialized memory.
There were no errors reported on any of the tested conditions.
7. Media Type Definition
This registration is done using the template defined in [RFC6838] and
following [RFC4855].
Type name: video
Subtype name: FFV1
Required parameters: None.
Optional parameters:
This parameter is used to signal the capabilities of a receiver
implementation. This parameter MUST NOT be used for any other
purpose.
version: The version of the FFV1 encoding as defined by
Section 4.1.1.
micro_version: The micro_version of the FFV1 encoding as defined by
Section 4.1.2.
coder_type: The coder_type of the FFV1 encoding as defined by
Section 4.1.3.
Niedermayer, et al. Expires March 29, 2019 [Page 37]
Internet-Draft FFV1 September 2018
colorspace_type: The colorspace_type of the FFV1 encoding as defined
by Section 4.1.5.
bits_per_raw_sample: The version of the FFV1 encoding as defined by
Section 4.1.7.
max-slices: The value of max-slices is an integer indicating the
maximum count of slices with a frames of the FFV1 encoding.
Encoding considerations:
This media type is defined for encapsulation in several audiovisual
container formats and contains binary data; see Section 4.2.3. This
media type is framed binary data Section 4.8 of [RFC4288].
Security considerations:
See Section 6 of this document.
Interoperability considerations: None.
Published specification:
[I-D.ietf-cellar-ffv1] and RFC XXXX.
[RFC Editor: Upon publication as an RFC, please replace "XXXX" with
the number assigned to this document and remove this note.]
Applications which use this media type:
Any application that requires the transport of lossless video can use
this media type. Some examples are, but not limited to screen
recording, scientific imaging, and digital video preservation.
Fragment identifier considerations: N/A.
Additional information: None.
Person & email address to contact for further information: Michael
Niedermayer <mailto:michael@niedermayer.cc>
Intended usage: COMMON
Restrictions on usage: None.
Author: Dave Rice <mailto:dave@dericed.com>
Change controller: IETF cellar working group delegated from the IESG.
Niedermayer, et al. Expires March 29, 2019 [Page 38]
Internet-Draft FFV1 September 2018
8. IANA Considerations
The IANA is requested to register the following values:
o Media type registration as described in Section 7.
9. Appendixes
9.1. Decoder implementation suggestions
9.1.1. Multi-threading Support and Independence of Slices
The FFV1 bitstream is parsable in two ways: in sequential order as
described in this document or with the pre-analysis of the footer of
each slice. Each slice footer contains a slice_size field so the
boundary of each slice is computable without having to parse the
slice content. That allows multi-threading as well as independence
of slice content (a bitstream error in a slice header or slice
content has no impact on the decoding of the other slices).
After having checked keyframe field, a decoder SHOULD parse
slice_size fields, from slice_size of the last slice at the end of
the "Frame" up to slice_size of the first slice at the beginning of
the "Frame", before parsing slices, in order to have slices
boundaries. A decoder MAY fallback on sequential order e.g. in case
of a corrupted "Frame" (frame size unknown, slice_size of slices not
coherent...) or if there is no possibility of seek into the stream.
10. Changelog
See <https://github.com/FFmpeg/FFV1/commits/master>
11. References
11.1. Normative References
[I-D.ietf-cellar-ffv1]
Niedermayer, M., Rice, D., and J. Martinez, "FFV1 Video
Coding Format Version 0, 1, and 3", draft-ietf-cellar-
ffv1-04 (work in progress), July 2018.
[ISO.15444-1.2016]
International Organization for Standardization,
"Information technology -- JPEG 2000 image coding system:
Core coding system", October 2016.
Niedermayer, et al. Expires March 29, 2019 [Page 39]
Internet-Draft FFV1 September 2018
[ISO.9899.1990]
International Organization for Standardization,
"Programming languages - C", ISO Standard 9899, 1990.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119,
DOI 10.17487/RFC2119, March 1997,
<https://www.rfc-editor.org/info/rfc2119>.
[RFC4288] Freed, N. and J. Klensin, "Media Type Specifications and
Registration Procedures", RFC 4288, DOI 10.17487/RFC4288,
December 2005, <https://www.rfc-editor.org/info/rfc4288>.
[RFC4732] Handley, M., Ed., Rescorla, E., Ed., and IAB, "Internet
Denial-of-Service Considerations", RFC 4732,
DOI 10.17487/RFC4732, December 2006,
<https://www.rfc-editor.org/info/rfc4732>.
[RFC4855] Casner, S., "Media Type Registration of RTP Payload
Formats", RFC 4855, DOI 10.17487/RFC4855, February 2007,
<https://www.rfc-editor.org/info/rfc4855>.
[RFC6716] Valin, JM., Vos, K., and T. Terriberry, "Definition of the
Opus Audio Codec", RFC 6716, DOI 10.17487/RFC6716,
September 2012, <https://www.rfc-editor.org/info/rfc6716>.
[RFC6838] Freed, N., Klensin, J., and T. Hansen, "Media Type
Specifications and Registration Procedures", BCP 13,
RFC 6838, DOI 10.17487/RFC6838, January 2013,
<https://www.rfc-editor.org/info/rfc6838>.
11.2. Informative References
[Address-Sanitizer]
The Clang Team, "ASAN AddressSanitizer website", undated,
<https://clang.llvm.org/docs/AddressSanitizer.html>.
[AVI] Microsoft, "AVI RIFF File Reference", undated,
<https://msdn.microsoft.com/en-us/library/windows/desktop/
dd318189%28v=vs.85%29.aspx>.
[HuffYUV] Rudiak-Gould, B., "HuffYUV", December 2003,
<https://web.archive.org/web/20040402121343/
http://cultact-server.novi.dk/kpo/huffyuv/huffyuv.html>.
Niedermayer, et al. Expires March 29, 2019 [Page 40]
Internet-Draft FFV1 September 2018
[ISO.14495-1.1999]
International Organization for Standardization,
"Information technology -- Lossless and near-lossless
compression of continuous-tone still images: Baseline",
December 1999.
[ISO.14496-10.2014]
International Organization for Standardization,
"Information technology -- Coding of audio-visual objects
-- Part 10: Advanced Video Coding", September 2014.
[ISO.14496-12.2015]
International Organization for Standardization,
"Information technology -- Coding of audio-visual objects
-- Part 12: ISO base media file format", December 2015.
[Matroska]
IETF, "Matroska", 2016, <https://datatracker.ietf.org/doc/
draft-lhomme-cellar-matroska/>.
[NUT] Niedermayer, M., "NUT Open Container Format", December
2013, <https://ffmpeg.org/~michael/nut.txt>.
[range-coding]
Nigel, G. and N. Martin, "Range encoding: an algorithm for
removing redundancy from a digitised message.", Proc.
Institution of Electronic and Radio Engineers
International Conference on Video and Data Recording ,
July 1979.
[REFIMPL] Niedermayer, M., "The reference FFV1 implementation / the
FFV1 codec in FFmpeg", undated, <https://ffmpeg.org>.
[VALGRIND]
Valgrind Developers, "Valgrind website", undated,
<https://valgrind.org/>.
[YCbCr] Wikipedia, "YCbCr", undated,
<https://en.wikipedia.org/w/index.php?title=YCbCr>.
Authors' Addresses
Michael Niedermayer
Email: michael@niedermayer.cc
Niedermayer, et al. Expires March 29, 2019 [Page 41]
Internet-Draft FFV1 September 2018
Dave Rice
Email: dave@dericed.com
Jerome Martinez
Email: jerome@mediaarea.net
Niedermayer, et al. Expires March 29, 2019 [Page 42]