From caidww at czdn755.net Sun Oct 5 07:50:59 2008 From: caidww at czdn755.net (=?GB2312?B?sszA7sfg?=) Date: Sun Oct 5 07:51:07 2008 Subject: [ale] Chinese hand-made hats Message-ID: [1]Home [2]Lady's hats [3]Men's hats [4]Other hats [5]Contact us We, Guangzhou Kingdu Hats Co., Ltd., recommend Chinese hand-made hats to you. These hats are exported to the US, Japan, EU. and other parts of the world. Our company is located in Guangzhou, China, which is a developed city of hand-made manufacturing. We can supply low cost & high quality products due to the cheaper hand-made employees. If you have your own brand, logo, artwork or sample that you want to develop and import from us, please send them to us by e-mail attachment or air parcel, we'll do our best to develop the samples including designing them for you. And we promise you that we are the best supplier based on our long-time experience, products of high quality, competitive prices, versatile designs and on-time delivery, and with the principle of, "Customer First". If you wish to form a powerful, long-term business relationship, please do not hesitate to contact us. And we'll reply to your inquiry as soon as possible. (Please send your inquiry to the email address on our website. ) [b_2B4B2C0ED85E013ED8D61CC852E00F8C.jpg] [b_0A5646F65A47F33B6D8850C972D450BE.jpg] [b_D215234E77328FF035CBF721D04258B5.jpg] [b_0C8C4EC69DAECAC6266EB04061265FE4.jpg] Guangzhou Kingdu Hats Co., Ltd. Tel£º+86-20-83617056 Fax£º+86-20-87279459 Add£ºRoom 1908, 320, shataibei Road, GuangZhou City, China 510515 If you can not refer to the link and the photos above, pls visit this website£ºhttp://hi.baidu.com/exporthats/blog/item/37b611168fbdb9074a90 a7f5.html References 1. http://hi.baidu.com/exporthats/blog/item/2c4d3afb00f77b8f9e5146e8.html 2. http://hi.baidu.com/exporthats/blog/item/37b611168fbdb9074a90a7f5.html 3. http://hi.baidu.com/exporthats/blog/item/5c4544fa0dd2ab62034f56f5.html 4. http://hi.baidu.com/exporthats/blog/item/ab09f4f3d5519315b07ec5f6.html 5. http://hi.baidu.com/exporthats/blog/item/9f43c63cb28f3bc19f3d62f6.html From sopues at gmail.com Wed Oct 8 06:38:33 2008 From: sopues at gmail.com (=?ISO-8859-1?Q?John_B=E4ckstrand?=) Date: Wed Oct 8 06:38:44 2008 Subject: [ale] HTC Dream phone, no AA filter Message-ID: Check this photo: http://www.flickr.com/photos/codexile/2887815959/sizes/o/ If its really from the HTC Dream, I will assume it has no AA filter (does mobile phone cams normally?). Check the aliasing on the bar pattern in the rear window. Like made for doing super-resolution on? -- John B?ckstrand From zds at iki.fi Thu Oct 30 06:53:58 2008 From: zds at iki.fi (Jari Juslin) Date: Thu Oct 30 06:54:36 2008 Subject: [ale] 3D depth restoration concurrency? Message-ID: <4909BC76.3030709@iki.fi> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello! Has there been any plans to make the 3D depth restoration scale to multiple cores? It's so slow even trivial testing with it takes ages.. so dropping the execution time would make it more feasible to test. I tried with down-scaled images, 622x466 pixel version took like 48 CPU hours to render, and that's quite a lot :-). -Jari -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkkJvHYACgkQ9FQS+GEmmz7SzwCeJm3TMwkKc+qRUahSlNILzQzo FcIAnRPVIYtafEL9yAFE5dZUp41E9IFE =Fn7/ -----END PGP SIGNATURE----- From dhilvert at gmail.com Thu Oct 30 17:32:13 2008 From: dhilvert at gmail.com (David Hilvert) Date: Thu Oct 30 17:33:57 2008 Subject: [ale] 3D depth restoration concurrency? In-Reply-To: <4909BC76.3030709@iki.fi> References: <4909BC76.3030709@iki.fi> Message-ID: <20081031013213.b4518ca9.dhilvert@gmail.com> On Thu, 30 Oct 2008 15:53:58 +0200 Jari Juslin wrote: > Has there been any plans to make the 3D depth restoration scale to > multiple cores? It's so slow even trivial testing with it takes ages.. > so dropping the execution time would make it more feasible to test. > > I tried with down-scaled images, 622x466 pixel version took like 48 CPU > hours to render, and that's quite a lot :-). I'm currently working on adapting the 2D code for GPUs, and it would be nice if the 3D code could eventually be adapted in this way as well, but I haven't looked into the details of this yet. If you're interested in the CPU approach, which should require less effort and time, you could try adapting the current thread class (see thread.h, and places where thread is derived) to areas of the 3D code that are taking the most time.