From xvid-devel-bounces@xvid.org Mon Aug 1 12:54:47 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id A2A2A126D82 for ; Mon, 1 Aug 2005 12:54:47 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id E0EF315771; Mon, 1 Aug 2005 12:54:40 +0200 (CEST) X-Original-To: xvid-devel@xvid.org Delivered-To: xvid-devel@edu.bnhof.de Received: from s5.lansco.de (s5.lansco.de [212.63.130.170]) by edu.bnhof.de (Postfix) with ESMTP id 0A3EA15766 for ; Mon, 1 Aug 2005 12:54:37 +0200 (CEST) Received: from localhost (s5.lansco.de [212.63.130.170]) by s5.lansco.de (Postfix) with ESMTP id 5611925F3C for ; Mon, 1 Aug 2005 12:54:37 +0200 (CEST) Received: from p54B5C6E6.dip.t-dialin.net (p54B5C6E6.dip.t-dialin.net [84.181.198.230]) by www.lansco.de (IMP) with HTTP for ; Mon, 1 Aug 2005 12:54:37 +0200 Message-ID: <1122893677.42edff6d35782@www.lansco.de> Date: Mon, 1 Aug 2005 12:54:37 +0200 From: Michael Militzer To: xvid-devel@xvid.org Subject: Re: [XviD-devel] [PATCH] Long standing error in xvid_decraw for big endian hosts References: <20050726234136.GA6713@edgomez.kicks-ass.org> In-Reply-To: <20050726234136.GA6713@edgomez.kicks-ass.org> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.3 X-Originating-IP: 84.181.198.230 X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Applied your patches - thanks! Michael Quoting Edouard Gomez : > Please apply after my last patch. > > # HG changeset patch > # User ed.gomez@free.fr > # Node ID 4fea89b12adc00ce55a53bb01a06c58188461880 > # Parent 5397bb5706a7c402c6438c21e4641c5edb729943 > Long standing error in xvid_decraw for big endian hosts > > > >From ed.gomez, spoted by Paul Kurucz ptk9417 at ritvax.isc.rit.edu > * No width and height declared in write_tga bug. > > diff -r 5397bb5706a7 -r 4fea89b12adc examples/xvid_decraw.c > --- a/examples/xvid_decraw.c Mon Jul 25 19:45:25 2005 > +++ b/examples/xvid_decraw.c Tue Jul 26 23:32:52 2005 > @@ -574,7 +574,7 @@ > #else > { > int i; > - for (i=0; i + for (i=0; i if (BPP == 1) { > fputc(*(image+i), f); > } else if (BPP == 2) { > > -- > Edouard Gomez > _______________________________________________ > XviD-devel mailing list > XviD-devel@xvid.org > http://list.xvid.org/mailman/listinfo/xvid-devel > _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Mon Aug 1 22:35:02 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 99C6E126D82 for ; Mon, 1 Aug 2005 22:35:02 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id EDCED15762; Mon, 1 Aug 2005 22:34:58 +0200 (CEST) X-Original-To: xvid-devel@xvid.org Delivered-To: xvid-devel@edu.bnhof.de Received: from smtp12.wanadoo.fr (smtp12.wanadoo.fr [193.252.22.20]) by edu.bnhof.de (Postfix) with ESMTP id 2B8DA1575F for ; Mon, 1 Aug 2005 22:34:55 +0200 (CEST) Received: from me-wanadoo.net (localhost [127.0.0.1]) by mwinf1209.wanadoo.fr (SMTP Server) with ESMTP id 404471C000B2 for ; Mon, 1 Aug 2005 22:34:55 +0200 (CEST) Received: from localhost.localdomain (ASte-Genev-Bois-151-1-43-29.w83-114.abo.wanadoo.fr [83.114.153.29]) by mwinf1209.wanadoo.fr (SMTP Server) with ESMTP id E6BC41C000B0 for ; Mon, 1 Aug 2005 22:34:54 +0200 (CEST) X-ME-UUID: 20050801203454945.E6BC41C000B0@mwinf1209.wanadoo.fr Subject: Re: [XviD-devel] Multithreaded Motion Search - looking for benchmarker From: Pascal Massimino To: xvid-devel@xvid.org In-Reply-To: <42E7F1CD.4030905@leoninedev.com> References: <42E7F1CD.4030905@leoninedev.com> Content-Type: text/plain Message-Id: <1122928538.3316.6.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 (1.4.5-14) Date: Mon, 01 Aug 2005 22:35:38 +0200 Content-Transfer-Encoding: 7bit X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: skal@planet-d.net, xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hello Bryan, On Wed, 2005-07-27 at 22:42, Bryan Mayland wrote: > While I was lying in bed the other night thinking about how much time it > takes to encode my 720x480 MPEG2 television recordings in 2-pass XViD, I > considered a processor upgrade. hey, hope it was just bad dream ;) > The algorithm creates a second thread to help with the ME, although > there is no reason this could not be increased further. An array > [mb_width x mb_height] of "completed blocks" is initialized to 0, and an > array of "available blocks" is initialized to point to the block at > (0,0). Once the thread has completed the search, the block is flagged as > completed and then the algorithm checks to see if the block to the right > can be searched (which is true if block2's top right neighbor is > complete), and if the block to the bottom left can be searched (which is > true if block3's left neighbor is complete). These blocks are added to > the "available" list. If there are blocks on the list, the thread takes > one and starts again. This seems to me like a very good idea. Something equivalent is possible on the decoder side, too. Did you consider using a cascading condition/signal instead of using the array? Every row signals it is finished doing its current MB to the one underneath, which is possibly blocked at its condition (but fortunately not waiting for long). I would warmly recommend using the pthread API for that. There are a lot of equivalent to it under Win32. I could quickly crank up one if you want... very nice! -Skal (ps: @Tom Jacobs: unfortunately your attachment didn't make it to the ML. Do you have a replacement URL??) _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Tue Aug 2 19:55:26 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id B56A8126D84 for ; Tue, 2 Aug 2005 19:55:26 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 3431E15794; Tue, 2 Aug 2005 19:55:24 +0200 (CEST) X-Original-To: xvid-devel@xvid.org Delivered-To: xvid-devel@edu.bnhof.de Received: from mail.ln9.net (rrcs-24-73-193-138.se.biz.rr.com [24.73.193.138]) by edu.bnhof.de (Postfix) with ESMTP id 3468B15663 for ; Tue, 2 Aug 2005 19:55:21 +0200 (CEST) Received: from [192.168.100.164] (yarr.ad.ln9.net [192.168.100.164]) by mail.ln9.net (Postfix) with ESMTP id CDF8C4B014B; Tue, 2 Aug 2005 13:55:14 -0400 (EDT) Message-ID: <42EFB382.90008@leoninedev.com> Date: Tue, 02 Aug 2005 13:55:14 -0400 From: Bryan Mayland Organization: Leonine Development, Inc. User-Agent: Mozilla Thunderbird 1.0.6 (Windows/20050716) X-Accept-Language: en-us, en MIME-Version: 1.0 To: skal@planet-d.net, xvid-devel@xvid.org Subject: Re: [XviD-devel] Multithreaded Motion Search - looking for benchmarker References: <42E7F1CD.4030905@leoninedev.com> <1122928538.3316.6.camel@localhost.localdomain> In-Reply-To: <1122928538.3316.6.camel@localhost.localdomain> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Pascal Massimino wrote: > hey, hope it was just bad dream ;) > > And after I the cond_wait joined... I realized I was naked! > Did you consider using a cascading condition/signal > instead of using the array? Every row signals it is > finished doing its current MB to the one underneath, > which is possibly blocked at its condition (but > fortunately not waiting for long). > > I almost did. I've actually since rewritten the double-array thing into a single subscripted array which represents the progress of the row of MB's search. I considered doing something where each row had it's own wait and the thread would sit and wait for the block to the top left of its current positon to finish, but that seemed to me to be a lot of waits. The way I have it now checks to see if it can keep moving right, if not check if there's any other rows that are ready, and if not then block on a wait. With the dual PII-450s, I see about a 20% speed increase, more if you crank up VHQ. (This is of course without BVOPs, so expect a third of that in a real world situation, right?) > I would warmly recommend using the pthread API for that. > There are a lot of equivalent to it under Win32. > I could quickly crank up one if you want... > > Well my code needs to be rewritten anyway. I was going to just write some functions that would do win32 on win32 and pthread anywhere else. If pthread is supported on win32 I'll do that though. > > _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Wed Aug 3 11:17:06 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 6CF42126D84 for ; Wed, 3 Aug 2005 11:17:06 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id E44F115808; Wed, 3 Aug 2005 11:16:53 +0200 (CEST) X-Original-To: xvid-devel@xvid.org Delivered-To: xvid-devel@edu.bnhof.de Received: from smtp20.libero.it (smtp20.libero.it [193.70.192.147]) by edu.bnhof.de (Postfix) with ESMTP id 7E60115805 for ; Wed, 3 Aug 2005 11:16:51 +0200 (CEST) Received: from localhost (172.16.1.84) by smtp20.libero.it (7.0.027-DD01) id 41D02C9802F2B8E7 for xvid-devel@xvid.org; Wed, 3 Aug 2005 11:16:50 +0200 Received: from libero.it (172.16.1.101) by smtp0.libero.it (7.0.027-DD01) id 42E0A09D0017F156 for xvid-devel@xvid.org; Wed, 3 Aug 2005 11:16:49 +0200 Date: Wed, 3 Aug 2005 11:16:49 +0200 Message-Id: MIME-Version: 1.0 X-Sensitivity: 3 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable From: "carlo.bramix" To: "xvid-devel" X-XaM3-API-Version: 4.1 (B27pl2) X-type: 0 X-SenderIP: 151.25.157.115 X-Virus-Scanned: by amavisd-new at libero.it serv5 Subject: [XviD-devel] Changes to colorspace_yuyv_mmx.asm X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hello, I did some changes to colorspace_yuyv_mmx.asm. Here they are:=0D = 1) Into macro YUYV_TO_YV12, it shouldn't be required to mask with MM7 if = we are working with YUYV colorspace. The masking is actually done by PSR= LW instruction, since it already clears the upper byte of the MMX registe= r. 2) Into macro YV12_TO_YUYVI, I think it's better to recalculate exp= ected value for ESI and EDI instead of doing PUSH/POP on them. Sincere= ly, Carlo Bramini. Index: src/image/x86_asm/colorspace_yuyv_mmx= .asm =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D RCS file: /xvi= d/xvidcore/src/image/x86_asm/colorspace_yuyv_mmx.asm,v retrieving revisi= on 1.6 diff -u -r1.6 colorspace_yuyv_mmx.asm --- src/image/x86_asm/colo= rspace_yuyv_mmx.asm 29 Aug 2004 10:02:38 -0000 1.6 +++ src/image/x86_asm= /colorspace_yuyv_mmx.asm 3 Aug 2005 08:50:32 -0000 @@ -94,9 +94,10 @@=0D = %if %1 =3D=3D 0 ; yuyv psrlw mm4, 8 psrlw mm5, 8 -%endi= f +%else pand mm4, mm7 pand mm5, mm7 +%endif paddw mm4, mm5= movq mm5, mm1 @@ -104,9 +105,10 @@ %if %1 =3D=3D 0 ; y= uyv psrlw mm5, 8 psrlw mm6, 8 -%endif +%else pand mm5, mm7=0D = pand mm6, mm7 +%endif paddw mm5, mm6 paddw mm4, [mmx_one] = ; +1 rounding paddw mm5, [mmx_one] ; @@ -129,10 +131,11 @@=0D = %if %1 =3D=3D 0 ; yuyv psrlw mm4, 8 psrlw mm5, 8 -%= endif +%else pand mm4, mm7 pand mm5, mm7 %endif +%endif ;--= ------------------------------------------------------------------ = ; write y-component @@ -141,11 +144,12 @@ psrlw mm1, 8 psrlw = mm2, 8 psrlw mm3, 8 -%endif +%else pand mm0, mm7 pand mm1, = mm7 pand mm2, mm7 pand mm3, mm7 +%endif packuswb mm0, mm1=0D = packuswb mm2, mm3 @@ -247,8 +251,6 @@ movq [edi+edx], mm6 = movq [edi+edx+8], mm7 - push esi - push edi add esi, eax a= dd edi, edx movq mm4, [esi+eax] ; [yyyy|yyyy] ; y row 2 @@= -263,8 +265,8 @@ movq [edi+edx+8], mm5 movq [edi+2*edx], mm6 = movq [edi+2*edx+8], mm7 - pop edi - pop esi + sub esi, eax + sub= edi, edx %else ; UYVY movq mm2, [esi] ; [y= yyy|yyyy] ; y row 0 movq mm3, [esi+eax] ; [yyyy|yyyy] ; y r= ow 1 @@ -281,8 +283,6 @@ movq [edi+edx], mm6 movq [edi+edx+8], m= m7 - push esi - push edi add esi, eax add edi, edx mov= q mm2, [esi+eax] ; [yyyy|yyyy] ; y row 2 @@ -299,8 +299,8 @@=0D = movq [edi+edx+8], mm5 movq [edi+2*edx], mm6 movq [edi+2*edx+8]= , mm7 - pop edi - pop esi + sub esi, eax + sub edi, edx %endif=0D = %endmacro =0A=0A=0A=0A______________________________________________= ______________=0ALibero Flat, sempre a 4 Mega a 19,95 euro al mese! =0AAb= bonati subito su http://www.libero.it=0A=0A _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Wed Aug 3 12:48:21 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 27D45126D84 for ; Wed, 3 Aug 2005 12:48:21 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 911CA156A0; Wed, 3 Aug 2005 12:48:20 +0200 (CEST) X-Original-To: xvid-devel@xvid.org Delivered-To: xvid-devel@edu.bnhof.de Received: from ftp.ilog.fr (ftp.ilog.fr [81.80.162.195]) by edu.bnhof.de (Postfix) with ESMTP id 6552415683 for ; Wed, 3 Aug 2005 12:48:16 +0200 (CEST) Received: from laposte.ilog.fr (cerbere-qfe0 [81.80.162.193]) by ftp.ilog.fr (8.13.4/8.13.3) with ESMTP id j73AmDYF011440 for ; Wed, 3 Aug 2005 12:48:13 +0200 (MET DST) Received: from marbore.ilog.biz (marbore1.ilog.fr [172.17.2.61]) by laposte.ilog.fr (8.13.1/8.13.1) with ESMTP id j73Am8nP009859 for ; Wed, 3 Aug 2005 12:48:08 +0200 (MET DST) Received: from parmbx01.ilog.biz ([172.17.2.64]) by marbore.ilog.biz with Microsoft SMTPSVC(6.0.3790.211); Wed, 3 Aug 2005 12:49:42 +0200 Received: from 6ttgg1j ([172.17.4.138]) by parmbx01.ilog.biz with Microsoft SMTPSVC(6.0.3790.211); Wed, 3 Aug 2005 12:49:42 +0200 Subject: Re: [XviD-devel] More MMX improvements => funky cbp! From: Skal To: xvid-devel@xvid.org In-Reply-To: References: Content-Type: text/plain Message-Id: <1123066117.3419.18.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 (1.4.5-14) Date: Wed, 03 Aug 2005 12:48:37 +0200 Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 03 Aug 2005 10:49:42.0212 (UTC) FILETIME=[0DBE8040:01C59819] X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hello Carlo! On Fri, 2005-07-29 at 17:59, carlo.bramix wrote: > Hello, > thanks a lot for your replies. > Unfortunately, I thought those routines were speed critical for the fact they were rewritten in ASM. > I will try to do improvements on other parts of the codec. Actually, i've tried applying your patch, and experience some binary differences in the output (using xvid_encraw.c with forced used of MMX cpu). Could you cross check your function is Ok? If not, then xvid_bench.c should you be enhanced to remove this false-positive. But it might just be me that messed the test up. Anyway, do you feel like exercising a little, just for the sport of it? Yes? 'coz i've had a look at your ASM code, and the final bit-by-bit computation of the CBP could be sped up a little, IMHO. Attention, we're just talking about few % speed-up of few % cpu use, here, but that's just for the challenge (summer is sooo boring;) Here it goes: cbp computation (for the luma part) is in fact a scalar product: cbp_y = 1.a + 2.b + 4.c + 8.d, where a,b,c, and d are boolean values deduced from or'ing all the 8x8 (luma) DCT coeffs (with exception to the DC), and 'pcmpgtw'ing them to zero. Now, you can easily compute this scalar product with good ol' 32bits-mult as: cbp_y = ( 0xdcba * 0x1248 ) >> 24 where 0xdcba is the 32bit integer resulting from packing (packssdw/wb) the four bools as 0xdcba = (d<<24) | (c<<16) | (b<<8) | (a) This works because no overflow occur for each individual terms. Just write the actual mult (like in school) to see it: 0x d c b a * 0x 1 2 4 8 -------------------- + 8d 8c 8b 8a + 4d 4c 4b 4a + 2d 2c 2b 2a + 1d 1c 1b 1a --------------------- = .........^^ and look at the sum in the fourth column. (yes, multiplication really is a convolution). Shifting this column to LSBits (with >>24), you get the cbp_y result with very few instructions. haf phun, -Skal (this mult trick is used in the GMC code, btw) _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Thu Aug 4 23:16:23 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id D908F126D89 for ; Thu, 4 Aug 2005 23:16:22 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 8BBAB15660; Thu, 4 Aug 2005 23:16:01 +0200 (CEST) X-Original-To: xvid-devel@xvid.org Delivered-To: xvid-devel@edu.bnhof.de Received: from smtp3.libero.it (smtp3.libero.it [193.70.192.127]) by edu.bnhof.de (Postfix) with ESMTP id 41EBA1565E for ; Thu, 4 Aug 2005 23:06:41 +0200 (CEST) Received: from localhost (172.16.1.79) by smtp3.libero.it (7.0.027-DD01) id 42C445C5006CFCC0 for xvid-devel@xvid.org; Thu, 4 Aug 2005 23:06:41 +0200 Received: from libero.it (172.16.1.104) by smtp2.libero.it (7.0.027-DD01) id 42C48C65004BFD63 for xvid-devel@xvid.org; Thu, 4 Aug 2005 23:06:40 +0200 Date: Thu, 4 Aug 2005 23:06:40 +0200 Message-Id: Subject: Re: [XviD-devel] More MMX improvements => funky cbp! MIME-Version: 1.0 X-Sensitivity: 3 Content-Type: multipart/mixed; boundary="_=__=_XaM3_.1123189600.2A.172210.42.7191.52.42.007.21659" From: "carlo.bramix" To: "xvid-devel" X-XaM3-API-Version: 4.1 (B27pl2) X-type: 0 X-SenderIP: 151.25.158.66 X-Virus-Scanned: by amavisd-new at libero.it serv1 X-Mailman-Approved-At: Thu, 04 Aug 2005 23:15:58 +0200 X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org --_=__=_XaM3_.1123189600.2A.172210.42.7191.52.42.007.21659 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Hello, after receiving your message, I tested again that MMX function.=0D = I extracted some parts from the benchmark and I did a small program for c= ross testing, like you suggested me. As you can see from the attached so= urce, I also added an additional loop for testing all steps of the routin= e. All results are compared with the ones obtained from the C implementa= tion. Everything works fine here. I don't know how the original bench t= est has been studied, but if the second loop doesn't generate errors, the= n the routine should really be ok. I perfectly understood your algorit= hm, but I had some doubts it could give us more speed. So, after an afte= rnoon of work I was able coded this alternative function. It uses the me= thod you suggested (just a bit modified because MMX instructions capabili= ties). It's a completely different method than currently used one, becau= se we need all terms for doing that multiply. MMX can do just 16x16 bit = multiply. So I used two accumulated 16x16 bit multiply and I got the sam= e thing. Unfortunately, the final result can't be obtained so easily as = you thought. Since there are six terms to accumulate, we need a shift-an= d-add anyways, like previous method. I did some benchmark for looking = how it could be fast, so I wrote a very simple DOS program and I compiled= it with DJGPP: with RDTSC, I read the elapsed time. These are the numbe= r of ticks: Current CVS function: 15005-14767 Latest patched functio= n: 13127 This new method: 11953 However, I must say that I was = almost forced to remove the loop and load all terms sequentially. Otherw= ise it runs even slower with 14231 ticks. I hope this could be helpful= . Sincerely, Carlo Bramini. PS: I hope I passed the test on my c= oding skills with this exercise ^_- ---------- Initial Header --------= --- >From : xvid-devel-bounces@xvid.org To : xvid-devel@= xvid.org Cc : Date : Wed, 03 Aug 2005 12:48:37 +0200 Su= bject : Re: [XviD-devel] More MMX improvements =3D> funky cbp! =0D = > > Hello Carlo! > > On Fri, 2005-07-29 at 17:59, carlo.brami= x wrote: > > Hello, > > thanks a lot for your replies. > > Unfortunate= ly, I thought those routines were speed critical for the fact they were r= ewritten in ASM. > > I will try to do improvements on other parts of the= codec. > > Actually, i've tried applying your patch, > and experi= ence some binary differences in > the output (using xvid_encraw.c with = forced > used of MMX cpu). > > Could you cross check your function = is Ok? > If not, then xvid_bench.c should you be > enhanced to remove= this false-positive. > But it might just be me that messed the test up= . > > > Anyway, do you feel like exercising a little, > just for = the sport of it? Yes? > > 'coz i've had a look at your ASM code, and=0D = > the final bit-by-bit computation of the CBP > could be sped up a lit= tle, IMHO. > Attention, we're just talking about few % > speed-up of = few % cpu use, here, but that's > just for the challenge (summer is soo= o boring;) > > Here it goes: > > cbp computation (for the luma pa= rt) is in fact > a scalar product: > > cbp_y =3D 1.a + 2.b + 4.c + = 8.d, > > where a,b,c, and d are boolean values > deduced from or'in= g all the 8x8 (luma) DCT coeffs > (with exception to the DC), and 'pcmp= gtw'ing > them to zero. > > Now, you can easily compute this scalar= product > with good ol' 32bits-mult as: > > cbp_y =3D ( 0xdcba * 0= x1248 ) >> 24 > > where 0xdcba is the 32bit integer resulting > fro= m packing (packssdw/wb) the four bools as > > 0xdcba =3D (d<<24) | (c= <<16) | (b<<8) | (a) > > This works because no overflow occur for eac= h > individual terms. Just write the actual mult > (like in school) t= o see it: > > 0x d c b a > * 0x 1 2 4 8=0D = > -------------------- > + 8d 8c 8b 8a > + 4d 4c = 4b 4a > + 2d 2c 2b 2a > + 1d 1c 1b 1a > --------------------- > =3D= .........^^ > > and look at the sum in the fourth column. > (yes, = multiplication really is a convolution). > > Shifting this column to = LSBits (with >>24), you get > the cbp_y result with very few instructio= ns. > > haf phun, > > -Skal > > (this mult trick is used in the= GMC code, btw) > > > > > ______________________________________= _________ > XviD-devel mailing list > XviD-devel@xvid.org > http://lis= t.xvid.org/mailman/listinfo/xvid-devel > =0A=0A=0A=0A___________________= _________________________________________=0A6X velocizzare la tua navigaz= ione a 56k? 6X Web Accelerator di Libero!=0AScaricalo su INTERNET GRATIS = 6X http://www.libero.it=0A --_=__=_XaM3_.1123189600.2A.172210.42.7191.52.42.007.21659 Content-Type: text/plain; name="=?iso-8859-1?Q?cbp=5Fmmx.asm?=" Content-Transfer-Encoding: quoted-printable Content-Disposition: attachment; filename="=?iso-8859-1?Q?cbp=5Fmmx.asm?=" ;/***********************************************************************= ***** ; * ; * XVID MPEG-4 VIDEO CODEC ; * - MMX CBP computation -=0D = ; * ; * Copyright (C) 2001-2003 Peter Ross ; * = 2002-2003 Pascal Massimino ; * ; * This p= rogram is free software ; you can redistribute it and/or modify ; * it = under the terms of the GNU General Public License as published by ; * t= he Free Software Foundation ; either version 2 of the License, or ; * (= at your option) any later version. ; * ; * This program is distributed= in the hope that it will be useful, ; * but WITHOUT ANY WARRANTY ; wit= hout even the implied warranty of ; * MERCHANTABILITY or FITNESS FOR A = PARTICULAR PURPOSE. See the ; * GNU General Public License for more de= tails. ; * ; * You should have received a copy of the GNU General Publ= ic License ; * along with this program ; if not, write to the Free Soft= ware ; * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 0211= 1-1307 USA ; * ; * $Id: cbp_mmx.asm,v 1.12 2004/08/29 10:02:38 edgomez = Exp $ ; * ; ***********************************************************= ****************/ BITS 32 ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D ; Macros ;=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D %macro cg= lobal 1 %ifdef PREFIX %ifdef MARK_FUNCS global _%1:function %1.e= ndfunc-%1 %define %1 _%1:function %1.endfunc-%1 %else global _= %1 %define %1 _%1 %endif %else %ifdef MARK_FUNCS global %= 1:function %1.endfunc-%1 %else global %1 %endif %endif %endm= acro ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D ; Local data ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D %ifdef FORMAT_COFF SECTION .r= odata %else SECTION .rodata align=3D16 %endif ALIGN 16 mult_mask= : db 0x10,0x20,0x04,0x08,0x01,0x02,0x00,0x00 ignore_dc: dw 0, = -1, -1, -1 ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D ; Code ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D SECTION .text cglobal _calc= _cbp_mmx %macro MAKE_LOAD 1 por mm0, [eax-128*1+%1*8]=0D = por mm1, [eax+128*0+%1*8] por mm2, [eax+128*1+%1*8] por mm3, [eax= +128*2+%1*8] por mm4, [eax+128*3+%1*8] por mm5, [eax+128*4+%1*8]=0D = %endmacro ;-----------------------------------------------------------= ------------------ ; uint32_t calc_cbp_mmx(const int16_t coeff[6][64]);=0D = ;------------------------------------------------------------------------= ----- ALIGN 16 _calc_cbp_mmx: mov eax, [esp + 4] ; coef= f movq mm7, [ignore_dc] psubd mm6, mm6 ; used only= for comparing movq mm0, [eax+128*0] movq mm1, [eax+128*1] movq = mm2, [eax+128*2] movq mm3, [eax+128*3] movq mm4, [eax+128*4] mov= q mm5, [eax+128*5] add eax, 8+128 pand mm0, mm7 pand mm1, mm7=0D = pand mm2, mm7 pand mm3, mm7 pand mm4, mm7 pand mm5, mm7 M= AKE_LOAD 0 MAKE_LOAD 1 MAKE_LOAD 2 MAKE_LOAD 3 MAKE_LOAD 4=0D = MAKE_LOAD 5 MAKE_LOAD 6 MAKE_LOAD 7 MAKE_LOAD 8 MAKE_LOAD 9= MAKE_LOAD 10 MAKE_LOAD 11 MAKE_LOAD 12 MAKE_LOAD 13 MAKE_= LOAD 14 movq mm7, [mult_mask] packssdw mm0, mm1 packssdw mm2, = mm3 packssdw mm4, mm5 packssdw mm0, mm2 packssdw mm4, mm6 pcm= pgtw mm0, mm6 pcmpgtw mm4, mm6 psrlw mm0, 15 psrlw mm4, 15 pa= ckuswb mm0, mm4 pmaddwd mm0, mm7 movq mm1, mm0 psrlq mm1, 32=0D = paddusb mm0, mm1 movd eax, mm0 shr eax, 8 and eax, 0x3F r= et .endfunc --_=__=_XaM3_.1123189600.2A.172210.42.7191.52.42.007.21659 Content-Type: text/plain; name="=?iso-8859-1?Q?Test.c?=" Content-Transfer-Encoding: quoted-printable Content-Disposition: attachment; filename="=?iso-8859-1?Q?Test.c?=" #include #define int16_t signed short int #define DECLARE= _ALIGNED_MATRIX(name, x, y, tp, al) \ tp name[x*y] int calc_cbp_mmx(c= onst int16_t *coeff); /* naive C */ int calc_cbp_plain(const int16_t = codes[6 * 64]) { int i, j, cbp =3D 0; for (i =3D 0; i < 6; i= ++) { for (j=3D1; j<64;j++) { if (codes[64*i+j]) {=0D = cbp |=3D 1 << (5-i); break; = } } } return cbp; } int test_cbp(const int16_t *c= oeff) { int cbp =3D calc_cbp_mmx(coeff); int res =3D calc_cbp_= plain(coeff); printf("calc_cbp#1 cbp=3D0x%02x %s\n", cbp, (cbp!=3D= res)?"| ERROR": ""); } int main() { int tst, cbp; int i;=0D = DECLARE_ALIGNED_MATRIX(Src1, 6, 64, int16_t, 16); DECLARE_ALIGNE= D_MATRIX(Src2, 6, 64, int16_t, 16); DECLARE_ALIGNED_MATRIX(Src3, 6, = 64, int16_t, 16); DECLARE_ALIGNED_MATRIX(Src4, 6, 64, int16_t, 16);=0D = DECLARE_ALIGNED_MATRIX(Src5, 6, 64, int16_t, 16); DECLARE_ALIGNE= D_MATRIX(Src6, 6, 64, int16_t, 16); printf( "\n =3D=3D=3D=3D=3D t= est cbp =3D=3D=3D=3D=3D\n" ); for(i=3D0; i<6*64; ++i) { S= rc1[i] =3D (i*i*3/8192)&(i/64)&1; /* 'random' */ Src2[i] =3D (i= <3*64); /* half-full */ Src3[i] =3D ((i+32)>3*64);= Src4[i] =3D (i=3D=3D(3*64+2) || i=3D=3D(5*64+9)); } t= est_cbp(Src1); test_cbp(Src2); test_cbp(Src3); test_cbp(Sr= c4); for (tst=3D0; tst<64; tst++) { for(i=3D0; i<6*64; ++= i) { Src1[i] =3D (i=3D=3Dtst+64*0) ? 1 : 0; = Src2[i] =3D (i=3D=3Dtst+64*1) ? 1 : 0; Src3[i] =3D = (i=3D=3Dtst+64*2) ? 1 : 0; Src4[i] =3D (i=3D=3Dtst+64*3)= ? 1 : 0; Src5[i] =3D (i=3D=3Dtst+64*4) ? 1 : 0; = Src6[i] =3D (i=3D=3Dtst+64*5) ? 1 : 0; } test= _cbp(Src1); test_cbp(Src2); test_cbp(Src3); te= st_cbp(Src4); test_cbp(Src5); test_cbp(Src6); }=0D = return 0; } --_=__=_XaM3_.1123189600.2A.172210.42.7191.52.42.007.21659 Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel --_=__=_XaM3_.1123189600.2A.172210.42.7191.52.42.007.21659-- From xvid-devel-bounces@xvid.org Wed Aug 10 18:24:17 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 63011126D83 for ; Wed, 10 Aug 2005 18:24:17 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 0624813EA1; Wed, 10 Aug 2005 18:23:58 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 523ECD9AD for ; Wed, 10 Aug 2005 18:23:51 +0200 (CEST) Received: from GWOUT.thalesgroup.com (gwout.thalesgroup.com [195.101.39.227]) by s6.lansco.de (Postfix) with ESMTP id DEDF74546 for ; Wed, 10 Aug 2005 18:23:50 +0200 (CEST) Received: from GWOUT.thalesgroup.com (200.3.3.3) by GWOUT.thalesgroup.com (NPlex 6.5.026) id 42F6D2A60002C40B for xvid-devel@xvid.org; Wed, 10 Aug 2005 18:23:56 +0200 Received: from thalescan.corp.thales ([200.3.2.3]) by GWOUT with InterScan Messaging Security Suite; Wed, 10 Aug 2005 18:23:56 +0200 Received: from msw001.uk.trt.thales ([192.168.224.28]) by thalescan with InterScan Messaging Security Suite; Wed, 10 Aug 2005 18:23:37 +0200 Received: from NTS013.uk.trt.thales (unverified) by msw001.uk.trt.thales (Content Technologies SMTPRS 4.2.10) with ESMTP id for ; Wed, 10 Aug 2005 17:23:22 +0100 Received: by nts013.uk.trt.thales with Internet Mail Service (5.5.2656.59) id ; Wed, 10 Aug 2005 17:23:21 +0100 Message-ID: <51BF576D5A02CC4CB2591F50994FD7660141C307@nts013.uk.trt.thales> From: "Yong, Chee Yeew " To: "'xvid-devel@xvid.org'" Date: Wed, 10 Aug 2005 17:23:21 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2656.59) Content-Type: text/plain; charset="iso-8859-1" Subject: [XviD-devel] Changing the frequency of I-Frame X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hello, I am interested in changing the frequency of the I-Frames generated by the XviD encoder. Is there an easy and/or 'accepted' way of doing this? The reason I am asking is because I am developing an application that does videoconferencing, and lost I-Frames causes the XviD decoder a long time to recover. I think that if I reduce the number of I-frames, then the video would recover quicker. However, this may change the target bitrate, am I right? Thanks for any pointers. Chee _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Wed Aug 10 18:30:44 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id ADA3B126D84 for ; Wed, 10 Aug 2005 18:30:44 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 935531465C; Wed, 10 Aug 2005 18:30:34 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id D7F6F85CD for ; Wed, 10 Aug 2005 18:30:30 +0200 (CEST) Received: from nproxy.gmail.com (nproxy.gmail.com [64.233.182.205]) by s6.lansco.de (Postfix) with ESMTP id 682D445FF for ; Wed, 10 Aug 2005 18:30:15 +0200 (CEST) Received: by nproxy.gmail.com with SMTP id g2so43902nfe for ; Wed, 10 Aug 2005 09:30:14 -0700 (PDT) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=TcWtyXqmEljcsRS5D/n1DshmIOV07rQ8M5ARzoteD+W2WYouZYXxoFJIAimpDfRZe0IJdzwcKZSQlmPp0gIVWnN5dSVHhFAJ8haK03dEgNXOebhYiMvgTaR4rJ16m7ndPq5BTP8PH136JNxtICzXd5F6g6qlOHPL1FBV7DpUtjE= Received: by 10.48.144.1 with SMTP id r1mr32725nfd; Wed, 10 Aug 2005 09:30:14 -0700 (PDT) Received: by 10.48.4.16 with HTTP; Wed, 10 Aug 2005 09:30:14 -0700 (PDT) Message-ID: <4e03026a0508100930261880b1@mail.gmail.com> Date: Wed, 10 Aug 2005 18:30:14 +0200 From: Guillaume POIRIER To: xvid-devel@xvid.org Subject: Re: [XviD-devel] Changing the frequency of I-Frame In-Reply-To: <51BF576D5A02CC4CB2591F50994FD7660141C307@nts013.uk.trt.thales> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline References: <51BF576D5A02CC4CB2591F50994FD7660141C307@nts013.uk.trt.thales> X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hi, On 8/10/05, Yong, Chee Yeew wrote: > Hello, >=20 > I am interested in changing the frequency of the I-Frames generat= ed > by the XviD encoder. Is there an easy and/or 'accepted' way of doing this= ? With MEncoder (part of MPlayer), it's the option "min_key_interval". Check the file libmpcodecs/ve_xvid4.c for more informations as to which is the XviD intenal option behind it. Guillaume --=20 A legend is an old man with a cane known for what he used to do. I'm still doing it. -- Miles Davis _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Wed Aug 10 18:37:36 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 1B4D9126D83 for ; Wed, 10 Aug 2005 18:37:36 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id E805E14FB5; Wed, 10 Aug 2005 18:37:25 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 36D0F14F9E for ; Wed, 10 Aug 2005 18:37:22 +0200 (CEST) Received: from mail.ln9.net (rrcs-24-73-193-138.se.biz.rr.com [24.73.193.138]) by s6.lansco.de (Postfix) with ESMTP id 13AC645F7 for ; Wed, 10 Aug 2005 18:37:07 +0200 (CEST) Received: from [192.168.100.164] (yarr.ad.ln9.net [192.168.100.164]) by mail.ln9.net (Postfix) with ESMTP id 908FC4B015D for ; Wed, 10 Aug 2005 12:37:05 -0400 (EDT) Message-ID: <42FA2D31.8070306@leoninedev.com> Date: Wed, 10 Aug 2005 12:37:05 -0400 From: Bryan Mayland Organization: Leonine Development, Inc. User-Agent: Mozilla Thunderbird 1.0.6 (Windows/20050716) X-Accept-Language: en-us, en MIME-Version: 1.0 To: xvid-devel@xvid.org Subject: Re: [XviD-devel] Changing the frequency of I-Frame References: <51BF576D5A02CC4CB2591F50994FD7660141C307@nts013.uk.trt.thales> In-Reply-To: <51BF576D5A02CC4CB2591F50994FD7660141C307@nts013.uk.trt.thales> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Yong, Chee Yeew wrote: > I am interested in changing the frequency of the I-Frames generated >by the XviD encoder. Is there an easy and/or 'accepted' way of doing this? > > Just change the max_key_interval you pass in the encoding struct in enc_create. >However, this may change the target bitrate, am I right? > > Well it doesn't change your target bitrate, but yes I-frames take a significant chunk of bitrate so you'll end up with a more lossy encode for the same bitrate (assuming you don't have any headroom left). _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Thu Aug 11 12:11:56 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 0DCEF126D83 for ; Thu, 11 Aug 2005 12:11:56 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 60AF214FBB; Thu, 11 Aug 2005 12:11:39 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 311AD14FB4 for ; Thu, 11 Aug 2005 12:11:36 +0200 (CEST) Received: from hotmail.com (bay19-f29.bay19.hotmail.com [64.4.53.79]) by s6.lansco.de (Postfix) with ESMTP id 7B1424621 for ; Thu, 11 Aug 2005 12:11:35 +0200 (CEST) Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Thu, 11 Aug 2005 03:11:17 -0700 Message-ID: Received: from 202.175.54.18 by by19fd.bay19.hotmail.msn.com with HTTP; Thu, 11 Aug 2005 10:11:16 GMT X-Originating-IP: [202.175.54.18] X-Originating-Email: [lawbreakeron@hotmail.com] X-Sender: lawbreakeron@hotmail.com From: "Law Ron" To: xvid-devel@xvid.org Date: Thu, 11 Aug 2005 10:11:16 +0000 Mime-Version: 1.0 Content-Type: text/plain; charset=big5; format=flowed X-OriginalArrivalTime: 11 Aug 2005 10:11:17.0443 (UTC) FILETIME=[034CA930:01C59E5D] Subject: [XviD-devel] Dynamically set the xvid zone quantizer? X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hi all, I'm trying to change xvid vfw zone quantizer simultaneously as the vfw is running, but it can't react simultaneously while the zone quantizer is changed. How should I modify the vfw so that it can compress with the changed zone quantizer? Thanks. _________________________________________________________________ ¨Ï¥Î MSN Messenger¡A»PªB¤Í¦b½u¤W²á¤Ñ: http://messenger.microsoft.com/tc _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Fri Aug 12 11:07:35 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id DC567126D83 for ; Fri, 12 Aug 2005 11:07:35 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 8B00114FBA; Fri, 12 Aug 2005 11:07:21 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 13CE014F9D for ; Fri, 12 Aug 2005 11:07:18 +0200 (CEST) Received: from GWOUT.thalesgroup.com (gwout.thalesgroup.com [195.101.39.227]) by s6.lansco.de (Postfix) with ESMTP id 41CA54626 for ; Fri, 12 Aug 2005 11:07:08 +0200 (CEST) Received: from GWOUT.thalesgroup.com (200.3.3.3) by GWOUT.thalesgroup.com (NPlex 6.5.026) id 42F6D2A600042116 for xvid-devel@xvid.org; Fri, 12 Aug 2005 11:06:54 +0200 Received: from thalescan.corp.thales ([200.3.2.3]) by GWOUT with InterScan Messaging Security Suite; Fri, 12 Aug 2005 11:06:54 +0200 Received: from msw001.uk.trt.thales ([192.168.224.28]) by thalescan with InterScan Messaging Security Suite; Fri, 12 Aug 2005 11:06:34 +0200 Received: from NTS013.uk.trt.thales (unverified) by msw001.uk.trt.thales (Content Technologies SMTPRS 4.2.10) with ESMTP id for ; Fri, 12 Aug 2005 10:06:17 +0100 Received: by nts013.uk.trt.thales with Internet Mail Service (5.5.2656.59) id ; Fri, 12 Aug 2005 10:06:16 +0100 Message-ID: <51BF576D5A02CC4CB2591F50994FD7660141C30B@nts013.uk.trt.thales> From: "Yong, Chee Yeew " To: "'xvid-devel@xvid.org'" Date: Fri, 12 Aug 2005 10:06:09 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2656.59) Content-Type: text/plain; charset="iso-8859-1" Subject: [XviD-devel] Decoded only 30 bytes from an I-frame - what causes it? X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hello all, I am experiencing a bizarre effect from the Xvid decoder. Whenever I try to decode an I-frame (by calling xvid_decore), I first get 30 bytes of decoded 'data', and then the rest of the I-frame gets decoded (the next time I call xvid_decore). To make it clearer, I have a buffer that keeps all the frames that are to be decoded. Whenever an I-frame is encountered, xvid_decore decodes only the first 30 bytes. I then shift the buffer forward by 30 bytes, leaving the rest of the I-frame in the buffer. The next time I call xvid_decore, only then the rest of the I-frame gets decoded. This does not occur on P-frames. There are no B-frames in my stream. Is this normal? What is the significance of the 30 bytes value? Any pointers *at all* would be welcome. Cheers, Chee _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Fri Aug 12 11:20:06 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 80902126D83 for ; Fri, 12 Aug 2005 11:20:06 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id E05D41581A; Fri, 12 Aug 2005 11:19:54 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 02324157FE for ; Fri, 12 Aug 2005 11:19:51 +0200 (CEST) Received: from modena.elinuxservers.com (modena.elinuxservers.com [64.235.243.136]) by s6.lansco.de (Postfix) with ESMTP id 50C0A4669 for ; Fri, 12 Aug 2005 11:19:36 +0200 (CEST) Received: from mp1-249-160.dialup.emory.edu ([170.140.249.160] helo=[10.0.1.3]) by modena.elinuxservers.com with esmtpa (Exim 4.44) id 1E3VhM-0003cj-80; Fri, 12 Aug 2005 02:19:33 -0700 Mime-Version: 1.0 (Apple Message framework v733) In-Reply-To: <51BF576D5A02CC4CB2591F50994FD7660141C30B@nts013.uk.trt.thales> References: <51BF576D5A02CC4CB2591F50994FD7660141C30B@nts013.uk.trt.thales> Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Message-Id: Content-Transfer-Encoding: 7bit From: Alexander Strange Subject: Re: [XviD-devel] Decoded only 30 bytes from an I-frame - what causes it? Date: Fri, 12 Aug 2005 05:19:24 -0400 To: xvid-devel@xvid.org, "Yong, Chee Yeew" X-Mailer: Apple Mail (2.733) X-AntiAbuse: This header was added to track abuse, please include it with any abuse report X-AntiAbuse: Primary Hostname - modena.elinuxservers.com X-AntiAbuse: Original Domain - xvid.org X-AntiAbuse: Originator/Caller UID/GID - [47 12] / [47 12] X-AntiAbuse: Sender Address Domain - ithinksw.com X-Source: X-Source-Args: X-Source-Dir: Cc: X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org On Aug 12, 2005, at 5:06 AM, Yong, Chee Yeew wrote: > > Is this normal? > What is the significance of the 30 bytes value? This is probably the VOL, which describes some things about the encoded video (size, what optional features it uses, etc). There's nothing much useful you can do with it. If you pass in an xvid_dec_stats_t, the 'type' field will be set to XVID_TYPE_VOL. _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Fri Aug 12 11:56:28 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 59A8F126D82 for ; Fri, 12 Aug 2005 11:56:28 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 9C49714FB0; Fri, 12 Aug 2005 11:56:15 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id CEB141498A for ; Fri, 12 Aug 2005 11:56:12 +0200 (CEST) Received: from GWOUT.thalesgroup.com (gwout.thalesgroup.com [195.101.39.227]) by s6.lansco.de (Postfix) with ESMTP id EDF554696 for ; Fri, 12 Aug 2005 11:56:06 +0200 (CEST) Received: from GWOUT.thalesgroup.com (200.3.3.3) by GWOUT.thalesgroup.com (NPlex 6.5.026) id 42F6D2A600042FD9 for xvid-devel@xvid.org; Fri, 12 Aug 2005 11:56:03 +0200 Received: from thalescan.corp.thales ([200.3.2.3]) by GWOUT with InterScan Messaging Security Suite; Fri, 12 Aug 2005 11:56:03 +0200 Received: from msw001.uk.trt.thales ([192.168.224.28]) by thalescan with InterScan Messaging Security Suite; Fri, 12 Aug 2005 11:55:44 +0200 Received: from NTS013.uk.trt.thales (unverified) by msw001.uk.trt.thales (Content Technologies SMTPRS 4.2.10) with ESMTP id ; Fri, 12 Aug 2005 10:55:28 +0100 Received: by nts013.uk.trt.thales with Internet Mail Service (5.5.2656.59) id ; Fri, 12 Aug 2005 10:55:28 +0100 Message-ID: <51BF576D5A02CC4CB2591F50994FD7660141C30D@nts013.uk.trt.thales> From: "Yong, Chee Yeew " To: 'Alexander Strange' , xvid-devel@xvid.org, "Yong, Chee Yeew " Subject: RE: [XviD-devel] Decoded only 30 bytes from an I-frame - what cau ses it? Date: Fri, 12 Aug 2005 10:55:18 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2656.59) Content-Type: text/plain; charset="iso-8859-1" Cc: X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Thanks for the reply. Is it possible to strip this out from the Xvid encoder? Chee -----Original Message----- From: Alexander Strange [mailto:astrange@ithinksw.com] Sent: Friday, August 12, 2005 10:19 AM To: xvid-devel@xvid.org; Yong, Chee Yeew Subject: Re: [XviD-devel] Decoded only 30 bytes from an I-frame - what causes it? On Aug 12, 2005, at 5:06 AM, Yong, Chee Yeew wrote: > > Is this normal? > What is the significance of the 30 bytes value? This is probably the VOL, which describes some things about the encoded video (size, what optional features it uses, etc). There's nothing much useful you can do with it. If you pass in an xvid_dec_stats_t, the 'type' field will be set to XVID_TYPE_VOL. _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Fri Aug 12 12:03:57 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 123F9126D83 for ; Fri, 12 Aug 2005 12:03:57 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id F081315000; Fri, 12 Aug 2005 12:03:44 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 30F6AACE9 for ; Fri, 12 Aug 2005 12:03:42 +0200 (CEST) Received: from modena.elinuxservers.com (modena.elinuxservers.com [64.235.243.136]) by s6.lansco.de (Postfix) with ESMTP id A729B4666 for ; Fri, 12 Aug 2005 12:03:41 +0200 (CEST) Received: from mp1-249-160.dialup.emory.edu ([170.140.249.160] helo=[10.0.1.3]) by modena.elinuxservers.com with esmtpa (Exim 4.44) id 1E3WNy-0006ok-Nz; Fri, 12 Aug 2005 03:03:35 -0700 Mime-Version: 1.0 (Apple Message framework v733) In-Reply-To: <51BF576D5A02CC4CB2591F50994FD7660141C30D@nts013.uk.trt.thales> References: <51BF576D5A02CC4CB2591F50994FD7660141C30D@nts013.uk.trt.thales> Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Message-Id: <1CE339D8-C19A-42E5-AAE9-EC81184DA3A1@ithinksw.com> Content-Transfer-Encoding: 7bit From: Alexander Strange Subject: Re: [XviD-devel] Decoded only 30 bytes from an I-frame - what cau ses it? Date: Fri, 12 Aug 2005 06:03:28 -0400 To: "Yong, Chee Yeew" , xvid-devel@xvid.org X-Mailer: Apple Mail (2.733) X-AntiAbuse: This header was added to track abuse, please include it with any abuse report X-AntiAbuse: Primary Hostname - modena.elinuxservers.com X-AntiAbuse: Original Domain - xvid.org X-AntiAbuse: Originator/Caller UID/GID - [47 12] / [47 12] X-AntiAbuse: Sender Address Domain - ithinksw.com X-Source: X-Source-Args: X-Source-Dir: Cc: X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org On Aug 12, 2005, at 5:55 AM, Yong, Chee Yeew wrote: > Thanks for the reply. > Is it possible to strip this out from the Xvid encoder? No, the video can't be decoded without it. (That's correct for avi, at least... you do have to take it out of the I-frames if you're creating .mp4, but you still keep it somewhere) _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Sat Aug 13 12:39:09 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 423C0126D82 for ; Sat, 13 Aug 2005 12:39:09 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id C98AA158B1; Sat, 13 Aug 2005 12:38:50 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 646D5158A7 for ; Sat, 13 Aug 2005 12:38:47 +0200 (CEST) Received: from mail52.ha.ovh.net (b6.ovh.net [213.186.33.56]) by s6.lansco.de (Postfix) with ESMTP id 577CB46D4 for ; Sat, 13 Aug 2005 12:38:36 +0200 (CEST) Received: from b0.ovh.net (HELO queue-out) (213.186.33.50) by b0.ovh.net with SMTP; 13 Aug 2005 10:38:27 -0000 Received: from mail52.ha.ovh.net (10.0.50.52) by mail52.ha.ovh.net with SMTP; 13 Aug 2005 10:38:25 -0000 Received: from b0.ovh.net (HELO queue-pre) (213.186.33.50) by b0.ovh.net with SMTP; 13 Aug 2005 10:38:25 -0000 Received: from pur31-3-82-240-189-110.fbx.proxad.net (HELO KCSOFT) (82.240.189.110) by ns0.ovh.net with SMTP; 13 Aug 2005 10:38:25 -0000 Message-ID: <008801c59ff3$27d4e460$0200a8c0@KCSOFT> From: "Kyle Katarn" To: Date: Sat, 13 Aug 2005 12:38:34 +0200 MIME-Version: 1.0 X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.2670 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2670 X-Ovh-Remote: 82.240.189.110 (pur31-3-82-240-189-110.fbx.proxad.net) X-Ovh-Local: 213.186.33.20 (ns0.ovh.net) X-Spam-Check: fait|type 1&3|0.8|H 0.500723 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable X-Content-Filtered-By: Mailman/MimeDel 2.1.4 Subject: [XviD-devel] Xvid systematic crash when encoding using VfW X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hello, I'm writing an application which uses Vfw for encoding videos. This works very good for most codec but systematically crashes with = Xvid. Who should I call for assistance ? Looks like the official xVid forum = does not accept new members... _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Sun Aug 14 16:53:26 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 62886126D83 for ; Sun, 14 Aug 2005 16:53:26 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 21EDBA884; Sun, 14 Aug 2005 16:53:11 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 10496A837 for ; Sun, 14 Aug 2005 16:53:06 +0200 (CEST) Received: from web52505.mail.yahoo.com (web52505.mail.yahoo.com [206.190.48.188]) by s6.lansco.de (Postfix) with SMTP id A052243FB for ; Sun, 14 Aug 2005 16:53:05 +0200 (CEST) Received: (qmail 89185 invoked by uid 60001); 14 Aug 2005 14:53:04 -0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com.ar; h=Message-ID:Received:Date:From:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=HA6MxJANh9r+CPv6rfmWLP6u2ASXzjDa8NRPFqEosGBdISCFhkDZFBH/Tcj2SajXznuORyxa+Uv3jTCfZMYQ78J4r0IGTykmcYu/w8eicamXADFDOEISCzDo2YDYnhewYIx7vEs6YQKAbk8h1PgP8stx6y4CIABBNvdgyWyW9ao= ; Message-ID: <20050814145304.89183.qmail@web52505.mail.yahoo.com> Received: from [200.42.34.199] by web52505.mail.yahoo.com via HTTP; Sun, 14 Aug 2005 11:53:04 ART Date: Sun, 14 Aug 2005 11:53:04 -0300 (ART) From: Dark Sylinc Subject: Re: [XviD-devel] Xvid systematic crash when encoding using VfW To: xvid-devel@xvid.org In-Reply-To: <008801c59ff3$27d4e460$0200a8c0@KCSOFT> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Are you using ICCompress()? If so, tell me the parameters you are using Dark_Sylinc --- Kyle Katarn escribió: > Hello, > > I'm writing an application which uses Vfw for > encoding videos. > This works very good for most codec but > systematically crashes with Xvid. > > Who should I call for assistance ? Looks like the > official xVid forum does not accept new members... > _______________________________________________ > XviD-devel mailing list > XviD-devel@xvid.org > http://list.xvid.org/mailman/listinfo/xvid-devel > __________________________________________________ Correo Yahoo! Espacio para todos tus mensajes, antivirus y antispam ¡gratis! ¡Abrí tu cuenta ya! - http://correo.yahoo.com.ar _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Sun Aug 14 18:19:16 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 08183126D83 for ; Sun, 14 Aug 2005 18:19:16 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 45BA4DF2A; Sun, 14 Aug 2005 18:19:03 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 9EFFEDF28 for ; Sun, 14 Aug 2005 18:19:00 +0200 (CEST) Received: from 25.mail-out.ovh.net (25.mail-out.ovh.net [213.186.37.103]) by s6.lansco.de (Postfix) with ESMTP id 8EC51459F for ; Sun, 14 Aug 2005 18:18:59 +0200 (CEST) Received: (qmail 25648 invoked by uid 503); 14 Aug 2005 16:18:44 -0000 Received: from b6.ovh.net (HELO mail135.ha.ovh.net) (213.186.33.56) by 25.mail-out.ovh.net with SMTP; 14 Aug 2005 16:18:44 -0000 Received: from b0.ovh.net (HELO queue-out) (213.186.33.50) by b0.ovh.net with SMTP; 14 Aug 2005 16:18:41 -0000 Received: from AMarigot-102-1-8-101.w80-15.abo.wanadoo.fr (AMarigot-102-1-8-101.w80-15.abo.wanadoo.fr [80.15.170.101]) by ssl0.ovh.net (IMP) with HTTP for ; Sun, 14 Aug 2005 18:18:41 +0200 Message-ID: <1124036321.42ff6ee129e5d@ssl0.ovh.net> Date: Sun, 14 Aug 2005 18:18:41 +0200 From: kyle.katarn@kcsoftwares.com To: xvid-devel@xvid.org, Dark Sylinc Subject: Re: [XviD-devel] Xvid systematic crash when encoding using VfW References: <20050814145304.89183.qmail@web52505.mail.yahoo.com> In-Reply-To: <20050814145304.89183.qmail@web52505.mail.yahoo.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.6 X-Originating-IP: 80.15.170.101 Cc: xvid-devel@xvid.org X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org I'm using AVIFileOpen and so on..., not ICCOmpress. Kyle > Are you using ICCompress()? > If so, tell me the parameters you are using > > Dark_Sylinc > > --- Kyle Katarn > escribió: > > > Hello, > > > > I'm writing an application which uses Vfw for > > encoding videos. > > This works very good for most codec but > > systematically crashes with Xvid. > > > > Who should I call for assistance ? Looks like the > > official xVid forum does not accept new members... > > _______________________________________________ > > XviD-devel mailing list > > XviD-devel@xvid.org > > http://list.xvid.org/mailman/listinfo/xvid-devel > > > > > __________________________________________________ > Correo Yahoo! > Espacio para todos tus mensajes, antivirus y antispam ¡gratis! > ¡Abrí tu cuenta ya! - http://correo.yahoo.com.ar > _______________________________________________ > XviD-devel mailing list > XviD-devel@xvid.org > http://list.xvid.org/mailman/listinfo/xvid-devel > _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Wed Aug 17 00:12:11 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id DE096126D89 for ; Wed, 17 Aug 2005 00:12:11 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id 44A29C0D8; Wed, 17 Aug 2005 00:11:59 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id D8561AF68 for ; Wed, 17 Aug 2005 00:11:55 +0200 (CEST) Received: from cmsout03.mbox.net (cmsout03.mbox.net [165.212.64.33]) by s6.lansco.de (Postfix) with ESMTP id 7CE1B4696 for ; Wed, 17 Aug 2005 00:11:49 +0200 (CEST) Received: from cmsout03.mbox.net (cmsout03.mbox.net [165.212.64.33]) by cmsout03.mbox.net (Postfix) with ESMTP id AE3708B for ; Tue, 16 Aug 2005 22:11:42 +0000 (GMT) Received: from uadvg129.cms.usa.net [165.212.11.129] by cmsout03.mbox.net via smtad (C8.MAIN.3.21U); Tue, 16 Aug 2005 22:11:42 GMT X-USANET-Source: 165.212.11.129 IN avcp-xvidmail@usa.net uadvg129.cms.usa.net X-USANET-MsgId: XID166JHPwLQ6605X03 Received: from mumba [205.247.13.101] by uadvg129.cms.usa.net (ASMTP/codepro@usa.net) via mtad (C8.MAIN.3.21E) with ESMTP id 182JHPwLO0470M29; Tue, 16 Aug 2005 22:11:40 GMT X-USANET-Auth: 205.247.13.101 AUTH codepro@usa.net mumba From: "Alex Volkov" To: Date: Tue, 16 Aug 2005 18:11:08 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Office Outlook, Build 11.0.6353 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1506 Thread-Index: AcWir2cEKYXQ4QKSTQ+Xo7t0CB3F0w== Message-ID: <182JHPwLO0470M29@uadvg129.cms.usa.net> Z-USANET-MsgId: XID182JHPwLp0470X29 Subject: [XviD-devel] Bug+patch: mem corruption w/ unaligned bitstream writes X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Hello, everyone. I am new to the XviD development and this list, so please forgive me if this is not the most appropriate place for this (I could not find a functioning bugtracker, or a FAQ on how to post patches). There is a bug in 1.1-CVS Bitstream init/writing functions that causes memory corruption when a bitstream passed to enc_encode() is not aligned on 32bit boundary. This wreaks havoc on my encoder output buffer. I do not write out the data immediately after enc_encode() returns it if the returned data is too small (for example, the dx50-compat empty marker, which is 7 bytes long). Instead, I simply move the bitstream ptr by 7 bytes. On the next enc_encode() op the last 3 bytes (which contain the time_increment) get clobbered. This of course causes severe block artifacts during decoding. The patch to fix unaligned bitstream inits: Index: src/bitstream/bitstream.h =================================================================== RCS file: /xvid/xvidcore/src/bitstream/bitstream.h,v retrieving revision 1.22 diff -u -r1.22 bitstream.h --- src/bitstream/bitstream.h 23 May 2005 09:29:43 -0000 1.22 +++ src/bitstream/bitstream.h 16 Aug 2005 21:40:03 -0000 @@ -193,8 +193,12 @@ #endif bs->bufb = tmp; - bs->buf = 0; bs->pos = bs->initpos = bitpos*8; + /* preserve the intervening bytes */ + if (bs->initpos > 0) + bs->buf = bs->bufa & ~((1 << (32 - bs->initpos)) - 1); + else + bs->buf = 0; bs->length = length; } @@ -220,7 +224,11 @@ #endif bs->bufb = tmp; - bs->buf = 0; + /* preserve the intervening bytes */ + if (bs->initpos > 0) + bs->buf = bs->bufa & ~((1 << (32 - bs->initpos)) - 1); + else + bs->buf = 0; bs->pos = bs->initpos; } _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Tue Aug 23 17:13:05 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 638E8126D91 for ; Tue, 23 Aug 2005 17:13:05 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id A7EF4DF44; Tue, 23 Aug 2005 17:12:22 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 8AE9BDF41 for ; Tue, 23 Aug 2005 17:12:12 +0200 (CEST) Received: from scamper.lut.ac.uk (scamper.lut.ac.uk [158.125.1.219]) by s6.lansco.de (Postfix) with ESMTP id F1E7A45BB for ; Tue, 23 Aug 2005 17:12:11 +0200 (CEST) Received: from [158.125.1.221] (helo=magic.lut.ac.uk) by scamper.lut.ac.uk with esmtp (Exim 4.44) id 1E7aRa-000829-Fq for xvid-devel@xvid.org; Tue, 23 Aug 2005 16:12:06 +0100 Received: from apache by magic.lut.ac.uk with local (Exim 4.44) id 1E7aRa-0005Xs-A3 for xvid-devel@xvid.org; Tue, 23 Aug 2005 16:12:06 +0100 Received: from elvc-linux3.lut.ac.uk (elvc-linux3.lut.ac.uk [158.125.51.81]) by student-webmail.lboro.ac.uk (IMP) with HTTP for ; Tue, 23 Aug 2005 16:12:06 +0100 Message-ID: <1124809926.430b3cc6450b0@student-webmail.lboro.ac.uk> Date: Tue, 23 Aug 2005 16:12:06 +0100 From: Tom Jacobs To: xvid-devel@xvid.org Subject: Re: [XviD-devel] Multithreaded Motion Search - looking for benchmarker References: <42E7F1CD.4030905@leoninedev.com> <1122928538.3316.6.camel@localhost.localdomain> In-Reply-To: <1122928538.3316.6.camel@localhost.localdomain> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.5 X-Originating-IP: 158.125.51.81 X-Scan-Signature: baffe82f75d80d7b8fef1d828701ef89 X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org sorry for the late reply, just got back from holiday here is the link to the paper that got stripped by the mailing list http://www-staff.lboro.ac.uk/~eltrj/jacobstr_thread-level-parallel-xvid.pdf hope this sheds light on all my questions to the list on the last year or so Tom Quoting Pascal Massimino : > > Hello Bryan, > > On Wed, 2005-07-27 at 22:42, Bryan Mayland wrote: > > While I was lying in bed the other night thinking about how much time > it > > takes to encode my 720x480 MPEG2 television recordings in 2-pass XViD, > I > > considered a processor upgrade. > > hey, hope it was just bad dream ;) > > > The algorithm creates a second thread to help with the ME, although > > there is no reason this could not be increased further. An array > > [mb_width x mb_height] of "completed blocks" is initialized to 0, and > an > > array of "available blocks" is initialized to point to the block at > > (0,0). Once the thread has completed the search, the block is flagged > as > > completed and then the algorithm checks to see if the block to the > right > > can be searched (which is true if block2's top right neighbor is > > complete), and if the block to the bottom left can be searched (which > is > > true if block3's left neighbor is complete). These blocks are added to > > the "available" list. If there are blocks on the list, the thread > takes > > one and starts again. > > This seems to me like a very good idea. Something > equivalent is possible on the decoder side, too. > > Did you consider using a cascading condition/signal > instead of using the array? Every row signals it is > finished doing its current MB to the one underneath, > which is possibly blocked at its condition (but > fortunately not waiting for long). > > I would warmly recommend using the pthread API for that. > There are a lot of equivalent to it under Win32. > I could quickly crank up one if you want... > > very nice! > -Skal > > (ps: @Tom Jacobs: unfortunately your attachment didn't make > it to the ML. Do you have a replacement URL??) > > > > > _______________________________________________ > XviD-devel mailing list > XviD-devel@xvid.org > http://list.xvid.org/mailman/listinfo/xvid-devel > _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Wed Aug 24 16:47:48 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id AB123126D92 for ; Wed, 24 Aug 2005 16:47:48 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id AEC7FDF11; Wed, 24 Aug 2005 16:47:35 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id D32EBC0D8 for ; Wed, 24 Aug 2005 16:47:32 +0200 (CEST) Received: from scamper.lut.ac.uk (scamper.lut.ac.uk [158.125.1.219]) by s6.lansco.de (Postfix) with ESMTP id 1576C4557 for ; Wed, 24 Aug 2005 16:47:32 +0200 (CEST) Received: from [158.125.1.221] (helo=magic.lut.ac.uk) by scamper.lut.ac.uk with esmtp (Exim 4.44) id 1E7wXL-0005xW-Qq for xvid-devel@xvid.org; Wed, 24 Aug 2005 15:47:31 +0100 Received: from apache by magic.lut.ac.uk with local (Exim 4.44) id 1E7wXL-0007ap-LA for xvid-devel@xvid.org; Wed, 24 Aug 2005 15:47:31 +0100 Received: from elvc-linux3.lut.ac.uk (elvc-linux3.lut.ac.uk [158.125.51.81]) by student-webmail.lboro.ac.uk (IMP) with HTTP for ; Wed, 24 Aug 2005 15:47:31 +0100 Message-ID: <1124894851.430c888398834@student-webmail.lboro.ac.uk> Date: Wed, 24 Aug 2005 15:47:31 +0100 From: Tom Jacobs To: "xvid-devel@xvid.org" MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.5 X-Originating-IP: 158.125.51.81 X-Scan-Signature: d35d600df6cb489d365aebfb10718d88 Subject: [XviD-devel] porting xvid X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org hi after noting how old my current version of xvid i threaded is i have tried to compile the newest (1.1.0-beta2) for my test system, simplescalar. this prolly shows how long ago it was but when i did it last xvid wasnt a lbrary but instead simply produced a simple binary which encompased the whole xvid core. this was nice and simple to compile for since i only had to change the CC and CCFLAGs. now i have hit a bit of a problem. the configure obviously wont produce the correct platform.inc file for my desired system. having looked through plaform.inc.in i can see the obvious setting to add like CC RANLIB CCFLAG but i am not sure about the rest. i dont need asm optimalisations or need to install the lib. i simple need to be able to compile the lib and use xvidencraw as a binary to test my threaded mods. can anyone suggest which settings i need to change thanks Tom _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel From xvid-devel-bounces@xvid.org Tue Aug 30 09:25:57 2005 Return-Path: X-Original-To: itdp@fh-biergarten.de Delivered-To: itdp@localhost Received: from edu.bnhof.de (edu.bnhof.de [213.167.167.52]) by mail.kliche.org (Postfix) with ESMTP id 77031126D89 for ; Tue, 30 Aug 2005 09:25:57 +0200 (CEST) Received: from edu.bnhof.de (localhost [127.0.0.1]) by edu.bnhof.de (Postfix) with ESMTP id F19F915862; Tue, 30 Aug 2005 09:25:46 +0200 (CEST) X-Original-To: xvid-devel@edu.bnhof.de Delivered-To: xvid-devel@edu.bnhof.de Received: from s6.lansco.de (s6.lansco.de [213.239.218.110]) by edu.bnhof.de (Postfix) with ESMTP id 48D9EDF54 for ; Sun, 28 Aug 2005 23:19:34 +0200 (CEST) Received: from www.thegratwicks.com (unknown [64.246.26.79]) by s6.lansco.de (Postfix) with ESMTP id BACDD461A for ; Sun, 28 Aug 2005 23:19:33 +0200 (CEST) Received: (qmail 30516 invoked from network); 28 Aug 2005 21:05:24 -0000 Received: from 212-41-101-212.adsl.solnet.ch (HELO xpc2) (212.41.101.212) by ev1s-64-246-26-79.ev1servers.net with SMTP; 28 Aug 2005 21:05:24 -0000 Message-ID: <002201c5ac16$3a43d870$2501a8c0@xpc2> From: "Doom9 Feedback Hotline" To: References: <42E7F1CD.4030905@leoninedev.com><1122928538.3316.6.camel@localhost.localdomain> <42EFB382.90008@leoninedev.com> Subject: Re: [XviD-devel] Multithreaded Motion Search - looking for benchmarker Date: Sun, 28 Aug 2005 23:19:19 +0200 MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=response Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.2670 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2670 X-BeenThere: xvid-devel@xvid.org X-Mailman-Version: 2.1.4 Precedence: list Reply-To: xvid-devel@xvid.org List-Id: xvid-devel.xvid.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: xvid-devel-bounces@xvid.org Errors-To: xvid-devel-bounces@xvid.org Bryan It's been a while, but if you're looking for beta testers, I'm sure you'll find some with suitable machines in my XviD forum: http://forum.doom9.org/forumdisplay.php?f=52 . Regards Doom9 _______________________________________________ XviD-devel mailing list XviD-devel@xvid.org http://list.xvid.org/mailman/listinfo/xvid-devel