FFV1 has been standardized at the IETF under RFC 9043.[7] TheEuropean Broadcasting Union (EBU) lists FFV1 under the codec-family index "31" in their combined list of video codec references.[8]
For long-term preservation of digital video sustainablecontainer formats as well as audio/video codecs are necessary. There is no consensus as of 2013[needs update] among the archival community as to which file format or codecs should be used for preservation purposes for digital video.[9] The previously proclaimed encodings wereMotion JPEG 2000[10] and uncompressed video.[11]
FFV1 proved to be a viable archival encoding and theU.S. Library of Congress began regarding it as a suitable option for preservation encoding in 2014.[2][12] Compared to lossless JPEG 2000, FFV1 features comparable compression ratios and lower computing requirements. As of 2014,[needs update] it is being used by archives, particularly where the collections do not feature extensive broadcast materials and instead consist of oral history and the like.[13][14]
Since around 2015, the European PREFORMA Project started working on the standardisation of FFV1 through theInternet Engineering Task Force (IETF).[15] It was standardised in August 2021 as RFC 9043.[7] The PERFORMA Project also implemented a conformance checker for FFV1 in theMatroska container.[16] Details of FFV1's standardization plan[17] have been prepared by MediaArea (authors ofMediaInfo) as part of their conformance checking tool Media CONCH.[18]
Within the video archiving domain, interest in FFV1 is increasing.[23][24][25][26][excessive citations][needs update] Companies are also picking up FFV1 support. For example, NOA announced support for the FFV1 in their product line in July 2013[27] and KEM-Studiotechnik released a film-scanner with FFV1 output in November 2013.[28]
"[...] for video, there are many choices when it comes to codecs (the way the bits are encoded/decoded to represent the visual data, e.g., ffv1, H.264, Apple ProRes) [...]"
In January 2013, the possible use and adoption of FFV1 as an archiving codec was addressed in the issue of PrestoCentre's[30]AV Insider magazine:[31]
"FFV1 has many beneficial technical features [...], but adoption rates are relatively low compared with alternatives, for example JPEG2000. [...] But holding back too long only serves to self-perpetuate the status of FFV1.
The adoption by Archivematica and the Austrian Mediathek with their active promotion of FFV1 along with others may start to break this vicious circle. This could lead to a virtuous circle of wider take-up, to shared development, to incorporation into commercial products and a host of other benefits for the community."
PACKED - the "Centre of Expertise in Digital Heritage" in Belgium, say in an article about video formats for archiving:[32]
"When removing the proprietary codecs from this list, only a few are left. [...] This basically leaves heritage institutions that want to use a lossless codec, with only two options: Jpeg2000 and FFV1."
meemoo [nl], the Flemish Institute for Archives announced in June 2020 that they would start a major operation transcoding and rewrapping their MXF wrapped JPEG2000 files resulting from digitisation projects for the cultural heritage sector material to MKV wrapped FFV1 files. In 2019 they had already announced that MKV-FFV1 would be the mezzanine format of their migration project for DV, DVCAM and DVCPRO cassettes.
The "Österreichische Mediathek" has also developed DVA-Profession aFree Software solution for archive-suitable mass video digitization, mainly using FFV1 as video encoding throughout the whole workflow, without transcoding.[52] Additionally, they have initiated the development of "FFV1.3" (=version 3 of FFV1) together with Michael Niedermayer (FFmpeg), Peter Bubestinger-Steindl and Dave Rice; see#Versions below.[53]
Here is a list of applications known to be able to read and/or write FFV1 video files, either natively or by installing codec packages.
Entries marked with "-" means that they generally only support either encodingor decoding.
The term"built-in" means that the application can handle FFV1 without the necessity to install additional codec packages.Applications that come with FFV1 supportout of the box, usually useFFmpeg's orLibav's libraries in order to do so.
The list is far from being complete, and will be augmented over time:
FFV1 is not strictly an intra-frame format; despite not using inter-frame prediction, it allows the context model to adapt over multiple frames. This can be useful for compression due to the very large size of the context table, but can be disabled to force the encoder to generate a strictly intra-frame bitstream. As the gained compression seems to decrease[66] with later versions of FFV1 (version 2,3), the use of GOP size greater than "1" might disappear in the future.
During progressive scanning of a frame, the difference between a current pixel and its predicted value, judging by neighboring pixels, is sent to the entropy-coding process. The prediction is done as follows:
prediction = Median(Top,Left,Top +Left -TopLeft)
The third value,Top + Left - TopLeft, is effectively equivalent to applying the "top" predictor to the current and the left sample, followed by applying the left predictor to the prediction residual of the top predictor. This method, also known as the gradient, exploits both horizontal and vertical redundancy. So in simple terms the prediction is themedian of the top, left, and gradient prediction methods. For improved performance and simplicity, the edges of the frame are assumed to be zero to avoid special cases. The prediction in encoding and decoding is managed using aring buffer.[67]
The residuals are coded using eitherGolomb-Rice coding[68] orrange coding. Both options use a very large context model. The "small" context model uses(11×11×11+1)/2=666 contexts based on the neighboring values of(Left −TopLeft),(TopLeft-Top), and(Top −TopRight). The "large" context model uses(11×11×5×5×5+1)/2=7563 contexts based on the same values as before, but also(TopTop −Top) and(LeftLeft −Left), whereTopTop is the pixel two above the current one vertically, andLeftLeft is the pixel two to the left of the current one. In range coding, each "context" actually has 32 sub-contexts used for various portions of coding each residual, resulting in a grand total of 242,016 contexts for the "large" model.
Early experimental versions of FFV1 used the CABAC Arithmetic coder from H.264, but due to the uncertain patent/royalty situation, as well as its slightly worse performance, CABAC was replaced by range coding.[69]
On April 16, 2006, a commit-message by Michael Niedermayer confirmed that the bitstream of FFV1 (version 1) is frozen:[70]
"ffv1 and ffvhuff haven't changed since a long time and no one proposed any changes within 1 month after my warning so they are officially no longerexperimental and we will guarantee decodability of files encoded with the current ffv1/ffvhuff in the future"
The bitstream of version 1 is frozen and has been considered stable for production use since April 2006.[70]
The remark "experimental" in the source code was overlooked back then and removed in March 2010.[71]
Version 2 (FFV1.2)
Version 2 was an intermediate version that was never officially released and should not be used for production purposes.
Version 3 (FFV1.3)
The bitstream of version 3 is frozen since August 3, 2013.[72] The final commit marking this version as officially released for production usage was on August 26, 2013.[53]
FFV1.3 contains improvements and new features such as support for multi-threaded encoding/decoding, error resilience and integrity validation by CRC checksums, storing of display aspect ratio (DAR) and field order. It was tested for over 1 year,[73] and officially released stable for production in August 2013.[53]
In August 2016, support for48bit/16bpc (=bits per component) inRGB was added to the reference codec.[74] Before that, 16bpc in FFV1 were only supported inYCbCr and RGB was limited to 14bpc.
There is still no VFW multithreaded encoder of FFV1.3 for Windows in 2017. FFdshow can encode only an FFV1.1 stream with a single CPU core.
Version 4 (FFV1.4)
Improvements beyond FFV1.3 are works in progress and being discussed on the IETF "CELLAR" mailing list.[75]
Planned are additional support for color-handling, especially non-linear/logarithmiccolor spaces.
The Draft standard is hosted on GitHub[1] and IETF Datatracker.[76]
^Austen, Barbara; Bowling, Melissa; Deakyne, Holly; Evans, Ryan (7 August 2013)."Electronic Records Project"(PDF). Society of American Archivists. Retrieved21 October 2014.
^Marsh, Alex (2021-03-12)."FFV1: The Gains of Lossless".Bitstreams: Notes from the digital collections team. Duke University Libraries.Archived from the original on 2021-04-06. Retrieved6 April 2021.
