多媒体格式专利费参考标准.docx

Multimedia Standards Licensing Fee

Introduction
This document outlines licensing costs and contact information for major multimedia standards and technology in use and in future use that is of interest to companies building next generation multimedia products.  Most licensing facts are referring to patent pools.  Additional costs would be high quality encoder and potentially decoder implementations. 
Most multimedia standards algorithms are asymmetrical.  In other words, the encoders require much more processing power and are much more complex both algorithmically as well as computationally.   This is due to the concept that encoder would be used on servers, in professional broadcast system in professional movie studio authoring tools; applications where the costs of encoding did not impact the delivery mechanism or cost to consumers.  Finally, in most standards, except for AC-3, the encoder is not normative.  In other words, companies and individuals are allowed to design the worse encoding technology possible as long as the algorithm generates a compliant bitstream. The highest quality encoder designs are often created by the very companies with patents in the various MPEG standards.  The main point of this discussion is to note the high variance in encoder technology quality and to also be aware that individual implementations of coders, should they be deemed necessary for this product will have additional licensing and fee costs. 
Standards
Each Standard is given a brief description, applications where it is in use, licensing terms and contact information.    Gnu Public License or any sort of algorithm or application that does not require payment is not addressed in this document.  Also not discussed are Real Networks formats or other lesser “player” technologies in the Internet Media space.

AUDIO
A.          MPEG1/2, Layer III (MP3)

1.   Type
MPEG1/2, layer III is a spectrum audio lossy compression algorithm.  Asymmetric encode and decode with a computational complexity ratio of 15:1.  Compression Ratio 12:1. 128kb/s stereo for 44.1 or 48kHz sampling rate is a typical bitrate in this technology.

2.   Description
Ubiquitous, MP3 is “The Internet standard” as well as portable player standard.  Cons:  very poor performance on speech signals, quality varies greatly depending upon implementation.  MPEG1, Layer III performs poorly in tandeming.    MPEG1 layer III is antiquated technology that is over 10 years old in the field of audio compression research.  Many different MPEG1-layer III encoders and compression tools are available via the Internet, thus quality varies from extremely good quality to encoders that will not even produce a proper bitstream. 

3.   Content
Illegal copying of content launched MP3 and today one can find almost anything available in free download using peer-to-peer software Kazaa in MP3 format.  NO Media service that does not support MP3 could be located in Q3/03.   
0.75USD
decode/unit
2.50-5.00USD
encode/unit

B.          MP3Pro

1.   Type
MP3Pro in architecture consists of modifications to the MPEG1-Layer III algorithm that add one additional DSP module.  This module, called SBR, reconstructs high frequency spectrum after decoding.  This piece of information implies that one can utilize an existing MPEG1 Layer III code base and potentially add an addition to the existing code to reduce code storage requirements. Asymmetric encode and decode in computational complexity.  Compression ratio is 14:1.  96kb/s stereo for 44.1 or 48kHz sampling rate is a typical bitrate in this technology.  Thompson advertises “half” the bitrate of MP3, yet investigation yields this bitrate as a typical setting.

2.   Description
MP3Pro gives very poor performance on speech signals yet there are few implementation versions, so quality is more consistent.  MP3Pro does not tandem.    The Spectral Band Replica technology, which SBR plus MP3 makes up MP3Pro, is under consideration in MPEG-4 audio standards bodies as well.  Thus, if one considered utilizing the MPEG-4 audio license pool, this portion of the MP3Pro algorithm is potentially available for use.  MP3Pro consistently ranks above MP3 in audio quality and often is competitive with AAC depending upon the encoder utilized, the profile chosen within AAC and the material.

3.   Content
Designed to replace MP3 with superior performance, MP3Pro has not achieved its goals.  Very little, content available on the Internet.  MP supposedly supports it, yet the database is almost exclusively MP3 files.

4.    Licensing
  1.25USD
decode/unit
5.00USD
encode/unit

C.          MPEG2-AAC
To obtain the licensing for patents only:
To obtain high quality implementations:      [e-mail]

1.   Type
The AAC specification supports up to 48 full-frequency range audio channels.  Asymmetric encode and decode in computational complexity.  Compression ratio is 18:1.  64kb/s stereo for 44.1 or 48kHz sampling rate is a typical bitrate in this technology.  128kb/s (as well as 96kb/s) stereo @ 44.1kHz was determined to be “indistinguishable” from CD quality audio in ITU standardized listening test results. 

2.   Description
AAC performs fairly well with speech signals down to 24kb/s stereo.  yet,  when compared to state of the art speech compression, AAC is inferior to ACELP based vocoders.    AAC is not designed for tandeming yet performs better than MPEG1, Layer III.    MPEG2-AAC has consistently ranked as the best of show for audio compression algorithms.  Please note that audio quality is dependent upon the design of the encoder.  AT&T and FhG as well as Dolby encoders are the only versions that should be considered when evaluating AAC quality.
http://david.weekly.org/audio/iso-aac.pdf

3.   Content
MPEG-2 AAC is the audio format utilized in the Japanese Digital Broadcast system, known as ISDB (Integrated Services Digital Broadcasting). MPEG-2 AAC is also the basis of the audio coding technology used by XM Radio, one of two satellite radio services currently operating in the United States. Additionally, the technology can be used for streaming and downloading music via the Internet, and has been specified within the DVD-Audio Recordable (DVD-AR) format.
To use MPEG2-AAC, there is a one-time 15000USD fee.  Rates are per channel, so a stereo implementation selling 50k units per quarter is 1.80USD for both the encoder and decoder or 0.45USD for just the encoder.   I believe this implies if one sells 1 unit per quarter one pays 3.60USD per unit for decode and 7.20USD per unit for both encoder and decode. Professional and Consumer are defined as to which market the product is targeting.    Pro Audio sells less, but the compression algorithm implementations should be the same, although one could increase the encoding of AAC to non-real time on a 2.4GHz processor to obtain the optimum in bit allocation routines.  Pro Audio decoders are 1.80USD per unit and encoder are 18.00USD per unit.

4.   Licensing
Volume
Either Decoderor Encoder
Both Encoder & Decoder
(Quarterly reset)                      
Per channel
Per channel
1 to 100,000
$0.45
$0.90
100,001 to 500,000
$0.35
$0.80
500,001 to 1,000,000
$0.25
$0.70
1,000,001 to 5,000,000
$0.20
$0.65
5,000,001 to 10,000,000
$0.15
$0.65
10,000,001 or more
$0.10
$0.65

D.          MPEG-4 Audio

1.   Type
MPEG-4 audio incorporates AAC above with various “objects” of other compression technology, such as a text to speech algorithm, a CELP based speech coder and some other very low bitrate encoding tools.  Most products are using MPEG-4 audio “AAC” which is MPEG2-AAC plus a couple of new DSP modules and ignoring much of the rest of the MPEG-4 audio standard. 
I find it interesting that if one claims they are a software solution in a “PC like” environment, the licensing costs become reasonable.

2.   Description
MPEG-4 AAC has been specified as the high-quality general audio coder for 3G wireless terminals. Apple Computer has incorporated MPEG-4 AAC into QuickTime 6 and iTunes 4, as well as the latest version of its award-winning iPod portable music player. The Digital Radio Mondiale system (the next-generation digital replacement for radio broadcasting under 30 MHZ) builds on the audio coding of MPEG-4 AAC.

3.   Content 
This standard has potential for future.  TBD how it is being adopted.  Please note that this compression algorithm is slightly superior to MPEG2-AAC yet the licensing costs are slightly lower in large volumes.  Also, note that there is a discount to anyone making a cell phone.  There is an initial fee of $15,000 due upon execution of the license.  These rates apply to patent rights only.  Implementations are again additional cost depending upon vendor.

4.   Licensing
Volume
Either Decoder or Encoder
Both Encoder & Decoder
(Quarterly reset)                      
Per channel
Per channel
1 to 100,000
$0.50
$1.00
100,001 to 500,000
$0.37
$0.74
500,001 to 1,000,000
$0.27
$0.54
1,000,001 to 5,000,000
$0.22
$0.44
5,000,001 to 10,000,000
$0.17
$0.34
10,000,001 or more
$0.12
$0.24

E.          MLP or Meridian Lossless Packing

1.   Type
Lossless audio compression that supports six channels of 24-bit/96 kHz audio, including a separate low-frequency bass channel or alternatively hold two channels of 24-bit/192 kHz audio.

2.   Description
MLP is required to implement the DVD-A specification.  TBD add previous investigation here.

3.   Content
A few DVD-A disks are in place.  Yet, the ability to play any DVD disk might be importantly strategically.  More I see the ability to play “sort” of content in high fidelity, i.e. 24bit depth, 96kHz as well the ability to support 5.1 or 7.1 audio being a feature that no other portable player has. 
10,000USD one time fee is required.  Fee is waived if one also has a “Dolby Digital” (AC-3) license. The royalty rate for MLP is calculated on a sliding scale based on quarterly quantity.  Table 2 shows the royalty schedule for MLP decoding.  These rates will be reduced by two-thirds in areas where no patents are in effect or no patent applications are in process. 
The cost-of-living adjustment to apply to royalties in 2000 for consumer MLP is 1.154. For professional MLP licenses, the adjustment would be 1.027.

4.   Licensing
Volume
Decoder Only
(Quarterly reset)                      
Per unit
1 to 10,000
$0.45
10,001 to 50,000
$0.33
50,001 to 250,000
$0.17
250,000 to 1,000,000
$0.12
1,000,001 or more
$0.08
Based on the total number of products sold per quarter, the table below illustrates the average per-product royalty rates.
Volume
DVD Player with MLP
(Quarterly reset)
AVERAGE rate/Per unit
< 10,000
$0.45
50,000
$0.35
100,000
$0.26
250,000
$0.21
500,000
$0.16
1,000,000
$0.14
2,500,00
$0.10

F.           Dolby Digital (AC-3)

1.   Type
AC-3 is a 5.1 compression algorithm supporting six channels of audio encoding. 
(Five discrete full spectrum channels and one channel carrying the low frequency response of the entire audio signal).  Typical bitrates for 5.1 compression are 320kb/s @ 48kHz sampling rate.  AC-3 does not support 96kHz-sampling rates.

2.   Description
AC-3 does poor in tandeming.  Dolby has created a new compression algorithm, Dolby E, which does tandem and is used as an intermediary step in audio production studios.  All content is finally encoded to AC-3 compliant bitstreams.
Dolby AC-3 decoder has the capability to downmix a 5.1 audio stream into stereo output.  Spatialization algorithms can then operate on the stereo downmix for stereo reproduction systems.

3.   Content
All DVD-V content has AC-3 encoding in the United States as well as some laserdisc.
AC-3 is the standard for HDTV audio broadcast.
Below are patents only, not the software for use of AC-3 (Dolby Digital).  Software implementations will be an additional cost.

4.   Licensing
0.40USD (stereo) to 0.60USD (5.1)
decode/unit
TBD from memory 8.00USD
encode/unit

VIDEO
G.         MPEG-2

1.   Type
MPEG-2 is a “generic” lossy video compression algorithm with typical bitrates from 2 to 15 Mbits/sec.

2.   Description
MPEG-2 is a generic video coding system supporting a diverse range of applications. Different algorithmic 'tools', developed for many applications, have been integrated into the full standard. To implement all the features of the standard in all decoders is unnecessarily complex and a waste of bandwidth, so a small number of subsets of the full standard, known as profiles and levels, have been defined. A profile is a subset of algorithmic tools and a level identifies a set of constraints on parameter values (such as picture size and bit rate). Video playback or decoders are only required to support the corresponding subset of the full standard defined by the profile.
For example, MPEG-2 Main Profile, Main Level allows frame sizes between 16 to 720 pixels horizontally and 16 to 576 pixels vertically. DVD, however only supports a subset of frame sizes.

3.   Content
MPEG-2 video and transport layer is part of the DVD-V specification, DBS system as well as the most common algorithm for Digital Cable and the choice of HDTV.   MPEG-2 video is the most common video compression algorithm in use.

4.   Licensing
$2.50/unit decode
   $2.50/unit encode        (personal use)
H.          MPEG-4

1.   Type
Set of video compression tools or objects to compress video and synthetic video in a variety of applications.  Introduces object-oriented video encoding which can compress arbitrary shapes within a video image as well as superimpose images in a composite tool. 

2.   Description
MPEG-4 Video supports a wide range of bitrates from the very low 5kbits/sec to a newly added Digital Cinema profile.  MPEG has ratified versions 1 and 2 but new additions are emerging in further versions of the MPEG-4 standard.  MPEG-4, visual part 10 is the work of most interest to content providers.  This profile promises DVD video quality at 1Mb/s in tests.  This fact implies that reasonable quality video can be streamed at lower broadband rates and rush the Internet based VoD market.

3.   Content
Apple as well as DivX use MPEG-4 but it is not the newer part 10 version that all are excited about.  Much of the intellectual property of the part 10 video profile is currently in Windows Media Player.  Due to MPEG-4 based DivX being the “new MP3” format of media, MPEG-4 video will surely be important in the future.

4.   Licensing
Licensing for MPEG-4 visual is currently in a state of flux.  MPEG-4, visual part 10 is an eme