From marek Fri Oct 28 08:50:33 1994 Received: from oldnova.npac.syr.edu by spica.npac.syr.edu (4.1/I-1.98K) id AA17378; Fri, 28 Oct 94 08:50:27 EDT Message-Id: <9410281250.AA17675@oldnova.npac.syr.edu> Received: from localhost.syr.edu by oldnova.npac.syr.edu (4.1/N-0.12) id AA17675; Fri, 28 Oct 94 08:50:26 EDT To: gcf Cc: project-leaders, marek From: Marek Podgorny Subject: Re: MPEG-2 Compression In-Reply-To: Your message of "Fri, 28 Oct 94 07:58:37 EDT." <9410281158.AA17547@oldnova.npac.syr.edu> Date: Fri, 28 Oct 94 08:50:26 -0400 Sender: marek Status: RO when Ccube came out with the design of programmable video compression chip, they described MPEG-2 as ideally suitable for parallel implementation. the chip architecture is a combination of data parallel and piplines. i will probably be able to localize a BYTE article about it. > > >At a conversation at meeting today I was told that >not known how to implement MPEG-2 compression on SP-2 >as true parallel algorithm -- goal isreal-time MPEG2 >compression >I dont believe this as both princeton engine and Maspar >have implemented parallel MPEG (1 or 2?) >Does anybody know either the algorithm definition or a reference? >As desribed to me it seemed to be clear how to do efficiently ... >Thanks From furm@npac.syr.edu Fri Oct 28 13:59:04 1994 Received: from king.syr.edu by spica.npac.syr.edu (4.1/I-1.98K) id AA20542; Fri, 28 Oct 94 13:58:18 EDT From: furm@npac.syr.edu (Wojtek Furmanski) Received: (furm@localhost) by king.syr.edu (8.6.7/8.6.6) id NAA10052; Fri, 28 Oct 1994 13:58:16 -0400 Date: Fri, 28 Oct 1994 13:58:16 -0400 Message-Id: <199410281758.NAA10052@king.syr.edu> To: gcf@npac.syr.edu Subject: Re: MPEG-2 Compression Cc: furm@npac.syr.edu, project-leaders@npac.syr.edu Status: RO There are tons on MPEG related pages on the Web. I just include one as an example. It is MPEG FAQ (Frequently Asked Questions), nicely organized and apparently quite informative. Home page splits into "MPEG standard" and "MPEG-#" where # = 1,2,3,4. The only problem is that the contributing subpages come from all over the world and so it takes a while to read it all through. I didn't - just checked that there are links there to C-Cube chip (the one referred to by Marek) in MPEG-2 section as well a zillion other links to vendors, standard committees, package developers, media companies and so on. BTW, is our Web Miner already operational? If so, this particular space could a natural targer for an NPAC mirror. The home page also tells us that the last update was done Oct 27 - hence the material looks pretty alive and worth reading/mirroring. URL of MPEG FAQ is: http://www.crs4.it/HTML/LUIGI/MPEG/mpegfaq.html and the home page is attached below. I only checked MPEG-2 link structure: when clicked on, it gives you (after a while...) a list of questions only (in blue). If you click on a question, you get (after w longer while...) a large file with all MPEG-2 answers and the one you asked for on top of the page. Hence, questions are implemented in anchor addressing mode and the whole FAQ is likely to be organized as five big files with all addressess sitting there. Q/As are post-processed or developed by author so style is homogeneous and quality apparently high (the technical content looks similar to Ken's HPCS glossary i.e. lots of algorithmic concepts in blue). Please let me know in case I need to work more now on this file/issue. ----------------------------------------------------------------------------- MPEG Moving Picture Expert Group FAQ

Moving Picture Expert Group

This is a collection of frequently asked questions and other informations about the MPEG compression standard. It is organized as an hypertext in HTML format to be easily extensible and upgradable.
If you are just looking for MPEG related public domain software or documentation you might want to browse the MPEG FTP Repository.

MPEG Frequently Asked Questions with Answers

  1. The MPEG standard

  2. MPEG-1

  3. MPEG-2

  4. MPEG-3

  5. MPEG-4
Latest changes : October 27 1994.


Many sources contributed to this document. If you wish to contribute, correct any mistake or just send your comments and impressions please use the comment form or contact :

Luigi.Filippini@crs4.it
Scientific Visualization Group
CRS4 From njm Fri Oct 28 12:11:02 1994 Received: from cepheus.npac.syr.edu by spica.npac.syr.edu (4.1/I-1.98K) id AA19648; Fri, 28 Oct 94 12:10:57 EDT Message-Id: <9410281610.AA06168@cepheus.npac.syr.edu> Received: from localhost.syr.edu by cepheus.npac.syr.edu (4.1/N-0.12) id AA06168; Fri, 28 Oct 94 12:10:52 EDT To: gcf, marek Cc: njm Subject: mpeg descriptions Date: Fri, 28 Oct 94 12:10:48 -0400 From: njm Status: RO I looked at the MPEG FAQ site on Mosaic - that's where I got a basic description of the MPEG 1 algorithm when I was preparing a lecture on it last summer. So I checked to see if there was anything similar now for MPEG 2, but couldn't find anything. Here's the URL: http://www.cis.ohio-state.edu/hypertext/faq/usenet/compression-faq/part2/faq-doc-2.html If you go up from there, there are some more, mostly uninteresting, MPEG FAQ's, and if you go up from there you get a large file of all the usenet faq's. Marek: Here's the similar one for JPEG, and I put a hard copy of the introduction to MPEG in your office. http://www.cis.ohio-state.edu/hypertext/faq/usenet/compression-faq/part2/faq-doc-6.html Nancy From marek Fri Oct 28 12:48:09 1994 Received: from cygnus.npac.syr.edu by spica.npac.syr.edu (4.1/I-1.98K) id AA19967; Fri, 28 Oct 94 12:48:06 EDT Message-Id: <9410281648.AA08178@cygnus.npac.syr.edu> Received: from localhost.syr.edu by cygnus.npac.syr.edu (4.1/N-0.12) id AA08178; Fri, 28 Oct 94 12:48:03 EDT To: choudhar@cat.syr.edu Cc: gcf@npac.syr.edu (Geoffrey Fox), marek From: Marek Podgorny Subject: Re: MPEG-2 Compression In-Reply-To: Your message of "Fri, 28 Oct 94 11:51:33 EDT." <9410281551.AA04670@cat.syr.edu> Date: Fri, 28 Oct 94 12:48:01 -0400 Sender: marek Status: RO the most time consuming part of the algorithm is a "pattern-matching procedure. the compression is done for pixel blocks, say 16x16. before calculating the difference, the algorithm looks for a closest match of this pixel block in the next or previous frame, following movin objects, in a sense. i will try to localize the artice over the weekend. > > >> Date: Fri, 28 Oct 94 07:58:37 EDT >> From: gcf@npac.syr.edu (Geoffrey Fox) >> To: project-leaders@npac.syr.edu >> Subject: MPEG-2 Compression > >> >> >> At a conversation at meeting today I was told that >> not known how to implement MPEG-2 compression on SP-2 >> as true parallel algorithm -- goal isreal-time MPEG2 >> compression >> I dont believe this as both princeton engine and Maspar >> have implemented parallel MPEG (1 or 2?) >> Does anybody know either the algorithm definition or a reference? >> As desribed to me it seemed to be clear how to do efficiently ... >> Thanks > >I believe algorithm involves difference of consecutive frames, Discrete Cosin >e >Transform, quantization .. (in a sequence) > >I would be intersted in Marek's article. >Alok >============================================================================= >= >Alok Choudhary Phone: 315-443-4280 (Off.) >Associate Professor 315-423-8065(home) >ECE, CIS and NPAC (Fax): 315-443-2583 >121 Link Hall choudhar@cat.syr.edu (preferred) >Syracuse University choudhar@nova.npac.syr.edu >Syracuse, NY 13244. http://www.cat.syr.edu/~choudhar > >============================================================================= From choudhar@cat.syr.edu Fri Oct 28 11:57:26 1994 Received: from cat.syr.edu (peach.ece.syr.EDU) by spica.npac.syr.edu (4.1/I-1.98K) id AA19545; Fri, 28 Oct 94 11:56:40 EDT Received: from localhost.syr.edu by cat.syr.edu (4.1/1.0-6/5/90) id AA04670; Fri, 28 Oct 94 11:51:36 EDT Message-Id: <9410281551.AA04670@cat.syr.edu> To: gcf@npac.syr.edu (Geoffrey Fox) Cc: project-leaders@npac.syr.edu Subject: Re: MPEG-2 Compression In-Reply-To: Your message of Fri, 28 Oct 94 07:58:37 -0500. <9410281158.AA17547@oldnova.npac.syr.edu> Date: Fri, 28 Oct 94 11:51:33 -0400 From: choudhar@cat.syr.edu Status: RO > Date: Fri, 28 Oct 94 07:58:37 EDT > From: gcf@npac.syr.edu (Geoffrey Fox) > To: project-leaders@npac.syr.edu > Subject: MPEG-2 Compression > > > At a conversation at meeting today I was told that > not known how to implement MPEG-2 compression on SP-2 > as true parallel algorithm -- goal isreal-time MPEG2 > compression > I dont believe this as both princeton engine and Maspar > have implemented parallel MPEG (1 or 2?) > Does anybody know either the algorithm definition or a reference? > As desribed to me it seemed to be clear how to do efficiently ... > Thanks I believe algorithm involves difference of consecutive frames, Discrete Cosine Transform, quantization .. (in a sequence) I would be intersted in Marek's article. Alok ============================================================================== Alok Choudhary Phone: 315-443-4280 (Off.) Associate Professor 315-423-8065(home) ECE, CIS and NPAC (Fax): 315-443-2583 121 Link Hall choudhar@cat.syr.edu (preferred) Syracuse University choudhar@nova.npac.syr.edu Syracuse, NY 13244. http://www.cat.syr.edu/~choudhar =============================================================================