You can subscribe to this list here.
2005 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(3) |
Dec
(4) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2006 |
Jan
(8) |
Feb
(7) |
Mar
(3) |
Apr
|
May
(4) |
Jun
(11) |
Jul
|
Aug
(4) |
Sep
(2) |
Oct
(1) |
Nov
(1) |
Dec
|
2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(1) |
Aug
|
Sep
(3) |
Oct
|
Nov
|
Dec
(1) |
2008 |
Jan
(1) |
Feb
(1) |
Mar
(1) |
Apr
(9) |
May
(9) |
Jun
(1) |
Jul
(1) |
Aug
|
Sep
(1) |
Oct
(14) |
Nov
(6) |
Dec
(4) |
2009 |
Jan
(1) |
Feb
(1) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(1) |
Nov
(2) |
Dec
(1) |
2010 |
Jan
(2) |
Feb
|
Mar
|
Apr
(2) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2012 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
|
2014 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(1) |
Dec
|
2015 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(1) |
Nov
(1) |
Dec
|
2016 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
(1) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
2017 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(1) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Schmeitz, H. <har...@ph...> - 2008-10-15 15:00:58
|
Hi Pawel, Your suggestion would indeed allow users to choose their interpretation. I assume the default will be "Binary offset", as this was the only mode in older versions of PV. Having the selection in PV does not solve the root cause: why are multiple vector interpretations used? Here at research we have defined motion vectors to be (I think it is also implemented in PV like this): - positive in the positive direction of time and frame coordinates (upper left is 0,0) - for 8 bit components: fixed point, with 1/4 pixel accuracy (6.2 fixed point) and offset 128 - for 16 bit components: fixed point, with 1/16 pixel accuracy (12.4 fixed point) and offset 32768 - for float components: floating point, gain=1, offset=0 Remember that a simple truncation of signed integers to signed bytes does not necessarily preserve the sign. Having to convert with an offset usually catches these overflow conditions (clipping should be applied of course). Regards, Harold ________________________________ From: Pawel Kruszynski [mailto:paw...@nx...] Sent: Wednesday 15 October 2008 16:00 To: pfs...@li... Subject: [Pfs...@sf...] pv - Vector interpretation modification Hi all, Present version of pv supports automatic detection of vector interpretation. There are two possible interpretations: two's complement (if bytes are signed), and binary offset, if values are calculated with respect to offset 128. The method used in present version of pv is based on statistics, and it is not always reliable. Therefore we are going to modify the processing of vector interpretation. Proposed modification consists in: 1. Disable auto-detection of the vector interpretation per frame and per component (function pfspd_Image::measure_pfspd_comp_two_comp) 2. Introduce user control from View menu: New menu item "Vector interpretation", with submenu items "Next vector interpretation", "Previous vector interpretation". Selecting one of them toggles between "Two's complement" and "Binary offset". Keys - "b" and "Shift b". The submenu items "Two's complement" and "Binary Offset" allow for direct selection. 3. Allow default value to be set in an init file and in command line by means of options "-binoffset" and "-2scomplement". Note - the vector interpretation selection is valid for all frames and components, unless changed by the user. Each new window starts with the default setting. Affected modules: cmdl_options.h format_convert.c frame.cc frame.h pfspd.cc pfspd.h settings.cc settings.h All comments and remarks are very much welcome Kind regards Pawel Kruszynski, Clemens Wust, Bram Rimens P. Kruszyński, Ph.D. NXP Semiconductors (founded by Philips) Corporate I&T/Research System & Algorithms High Tech Campus 32 room 1.38 5656 AE Eindhoven The Netherlands tel. +31(0)40-2725898 fax. +31402744639 gsm +31(0)629023208 ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Pawel K. <paw...@nx...> - 2008-10-15 14:02:31
|
Hi all, Present version of pv supports automatic detection of vector interpretation. There are two possible interpretations: two's complement (if bytes are signed), and binary offset, if values are calculated with respect to offset 128. The method used in present version of pv is based on statistics, and it is not always reliable. Therefore we are going to modify the processing of vector interpretation. Proposed modification consists in: 1. Disable auto-detection of the vector interpretation per frame and per component (function pfspd_Image::measure_pfspd_comp_two_comp) 2. Introduce user control from View menu: New menu item "Vector interpretation", with submenu items "Next vector interpretation", "Previous vector interpretation". Selecting one of them toggles between "Two's complement" and "Binary offset". Keys - "b" and "Shift b". The submenu items "Two's complement" and "Binary Offset" allow for direct selection. 3. Allow default value to be set in an init file and in command line by means of options "-binoffset" and "-2scomplement". Note - the vector interpretation selection is valid for all frames and components, unless changed by the user. Each new window starts with the default setting. Affected modules: cmdl_options.h format_convert.c frame.cc frame.h pfspd.cc pfspd.h settings.cc settings.h All comments and remarks are very much welcome Kind regards Pawel Kruszynski, Clemens Wust, Bram Rimens P. Kruszyński, Ph.D. NXP Semiconductors (founded by Philips) Corporate I&T/Research System & Algorithms High Tech Campus 32 room 1.38 5656 AE Eindhoven The Netherlands tel. +31(0)40-2725898 fax. +31402744639 gsm +31(0)629023208 |
From: Schmeitz, H. <har...@ph...> - 2008-10-15 09:42:19
|
Ik heb zelf geen MSVS 2008, dus kan ik jouw probleem niet reproduceren. Wat zal helpen bij debuggen is het zetten van #define FIO_TRACE (line 384 van cpfspd_fio.c) om de asynchrone file IO events te loggen. Een snelle test om te achterhalen of normale file IO wel lukt: FIO_WIN32_FILE niet definen (hiervoor moet je ook de #error in cpfspd_fio.h verwijderen) Nu zal cpfspd de standaard C read/write functies gebruiken (met daarbij horende file size limiet van 2 GB) Succes, Harold ________________________________ From: Vleuten, Rene van der [mailto:ren...@ph...] Sent: Tuesday 14 October 2008 15:08 To: pfs...@li... Subject: [Pfs...@sf...] FW: cpfspd onder VisualStudio 2008 ________________________________ From: Vleuten, Rene van der Sent: Tuesday 14 October 2008 14:57 To: Mevissen, Perry Cc: 'pfs...@sf...' Subject: RE: cpfspd onder VisualStudio 2008 Hoi Perry, Misschien weet Harold het maar die is er vandaag niet; daarom post ik hem ook maar even op de "officiele" lijst. Vriendelijke groeten, René. ________________________________ From: Mevissen, Perry Sent: Tuesday 14 October 2008 14:38 To: Geurts, Gert Cc: Vleuten, Rene van der; Schmeitz, Harold Subject: cpfspd onder VisualStudio 2008 Hallo, Ik hoop dat ik goe zit met m'n vraag bij jullie Heb jullie namen van gforge Ik ben op dit moment bezig met een project in Visual Studio 2008 Echter onvervind ik problemen bij het linken van de laatste pfspd library. Ik krijg vele conflicten tussen LIBCMT, MSVCRT en MSVCRTD. Als ik geen exclusions doe heeft de linker dubbele definities en dus errors. Als ik een van de libs exclude ( /NODEFAULTLIB optie ) dan kan de linker bepaalde functies niet vinden. Ik heb geprobeerd cpfspd zelf te compilen met __MSVC__ en FIO_WIN32_FILE flags Dit complieert en linkt allemaal zonder problemen alleen krijg ik nu altijd een P_READ_FAILED als ik bijvoorbeeld een pfspd file probeer te lezen in de applicatie die de lib begruikt. hints: * Ik heb vrij ver gedebugt en kom vaak de functie _fio tegen in meldingen. * Heeft het misschien te maken met het verplaatsen van functies uit LIBCMT naar MSVCRT door Miscrosoft ? Kunnen jullie helpen ? Best regards, Perry _____________________________________________ Perry Mevissen, Specialist Digital Video Applications Philips Consumer Lifestyle, Innovation Lab. High Tech Campus 37, office 8.039 5656 AE Eindhoven tel: +31 (0)6 1354 0669 mailto: Per...@ph... The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Vleuten, R. v. d. <ren...@ph...> - 2008-10-14 13:08:38
|
________________________________ From: Vleuten, Rene van der Sent: Tuesday 14 October 2008 14:57 To: Mevissen, Perry Cc: 'pfs...@sf...' Subject: RE: cpfspd onder VisualStudio 2008 Hoi Perry, Misschien weet Harold het maar die is er vandaag niet; daarom post ik hem ook maar even op de "officiele" lijst. Vriendelijke groeten, René. ________________________________ From: Mevissen, Perry Sent: Tuesday 14 October 2008 14:38 To: Geurts, Gert Cc: Vleuten, Rene van der; Schmeitz, Harold Subject: cpfspd onder VisualStudio 2008 Hallo, Ik hoop dat ik goe zit met m'n vraag bij jullie Heb jullie namen van gforge Ik ben op dit moment bezig met een project in Visual Studio 2008 Echter onvervind ik problemen bij het linken van de laatste pfspd library. Ik krijg vele conflicten tussen LIBCMT, MSVCRT en MSVCRTD. Als ik geen exclusions doe heeft de linker dubbele definities en dus errors. Als ik een van de libs exclude ( /NODEFAULTLIB optie ) dan kan de linker bepaalde functies niet vinden. Ik heb geprobeerd cpfspd zelf te compilen met __MSVC__ en FIO_WIN32_FILE flags Dit complieert en linkt allemaal zonder problemen alleen krijg ik nu altijd een P_READ_FAILED als ik bijvoorbeeld een pfspd file probeer te lezen in de applicatie die de lib begruikt. hints: * Ik heb vrij ver gedebugt en kom vaak de functie _fio tegen in meldingen. * Heeft het misschien te maken met het verplaatsen van functies uit LIBCMT naar MSVCRT door Miscrosoft ? Kunnen jullie helpen ? Best regards, Perry _____________________________________________ Perry Mevissen, Specialist Digital Video Applications Philips Consumer Lifestyle, Innovation Lab. High Tech Campus 37, office 8.039 5656 AE Eindhoven tel: +31 (0)6 1354 0669 mailto: Per...@ph... The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Bram R. <bra...@nx...> - 2008-10-14 07:09:18
|
Hi Madalina, all, Here some comments: 1. have the behavior of ESC back in PV in order to exit the application Not a good plan. Esc is used to get out of full-screen mode. In this mode, there is no menu, or whatsoever, so there is no way to help the user understand how to return to normal window mode. Many other application use esc in such case. For this reason people simply may try esc. LEts keep the current implementation. 1. combine the new behavior of TAB with the old one (if the user does not select a certain color method, TAB shall be used to browse the images while preserving the zoom factor) I'm confused, isn't that the case? Tab / sift-tab move to next/previous file in the directory while keeping current settings (incl. zoom). Analogous to space/backspace that cycle through the files opening each file with default settings. 1. navigate through a video at each 1 second using Ctrl-PageUp and Ctrl-PageDown Good suggestion. 1. have a shortcut key to set the default values of the Component Mapping Good suggestion. Maybe take shift-M as a shortcut of the M key that pops up the component mapping dialog? 1. exclude ICC from "Color display method" when pressing "C" or "shift-C" Hm, isn't that confusing? How to visualise this in the "color display method" submenu? Or do you propose to remove it from that submenu? But then, where to put it? For me, the whole ICC mode is not yet fully understood. I feel the need for a better understanding of ICC and how to integrate this in PV, such that it supports people who need it well without distracting/disturbing others. Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... Phone: +31 40 27 25910; Fax: +31 40 27 28504 High Tech Campus 32 (floor 1, office 138) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. From: Iota, Madalina [mailto:mad...@ph...] Sent: Monday 13 October 2008 16:13 To: pfs...@li... Subject: [Pfs...@sf...] pv: add new features Hi all, As you have been informed, a new rendering mode was added to PV. This mode displays the image using the ICC profile of the panel. Currently, ICC is not fully functional, reason why I want to continue improving it. One of the improvements would be the creating of a virtual profiler for YUV images. Besides this I would like to: 1. have the behavior of ESC back in PV in order to exit the application 2. combine the new behavior of TAB with the old one (if the user does not select a certain color method, TAB shall be used to browse the images while preserving the zoom factor) 3. navigate through a video at each 1 second using Ctrl-PageUp and Ctrl-PageDown 4. have a shortcut key to set the default values of the Component Mapping 5. exclude ICC from "Color display method" when pressing "C" or "shift-C" What is your opinion about this? Best regards, Madalina TASS Software Professionals Ir. Madalina Iota PDEng T+31 40 2503200 | F +31 40 2503201 W www.tass.nl<http://www.tass.nl> | M mad...@ta...<mailto:mad...@ta...> Temporary address: Philips Research Location HTC Phone: +31 40 274 1604 E-mail: mad...@ph...<mailto:mad...@ph...> ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Bram R. <bra...@nx...> - 2008-10-14 06:46:24
|
Hi Madalina, all, For me it is not clear what motivates this proposal. Is there a significant advantage? I have the followign comments: - Note that current API supports reading a frame, also for interlaced data the p_read_frame_xxx() call then actually reads two fields and combines it in a frame buffer. - If we decide to combine frame&field calls, then we shall do this for all read/write calls: this holds for the 8 bit versions, and 16 bit version; it holds for the 2 and 3 component calls, etc. this way the interface remains orthogonal and consistent. - Note that all cpfspd api calls must be accompanied by tests, so this will also add a substantial amount of test code. - Your proposed call suffers from unclarity: stride per frame line or per field line? - I suspect you will use a frame height, not a field height, since frame height is defined in all cases, whereas field height is not. As far as I can see, this introduces more issues than it solves. In particular, I prefer to see a really good reason to add interface calls. Obviously, you can always create such wrapper around some cpfspd calls in your application. Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... Phone: +31 40 27 25910; Fax: +31 40 27 28504 High Tech Campus 32 (floor 1, office 138) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. From: Iota, Madalina [mailto:mad...@ph...] Sent: Monday 13 October 2008 15:49 To: pfs...@li... Subject: [Pfs...@sf...] cpfspd: combine the call of p_read_frame_comp_16() and p_read_field_comp_16() in one method Hi all, I would like to have two functions to read and write image data per frame/field/component. The prototype of these functions would be: pT_status p_read_comp_16 (const char *filename, pT_header *header, int frame, int field, int comp, unsigned short *c_fld, int read_mode, int width, int fld_height, int stride); pT_status p_write_comp_16(const char *filename, pT_header *header, int frame, int field, int comp, const unsigned short *c_fld, int write_mode, int width, int fld_height, int stride); where: filename file to read/write header pointer to pT_header struct frame frame number field 0: frame, 1/2: field access. comp the component id. i.e. the index in the header structure. c_fld pointer to field/frame buffer. read_mode mem_data_fmt write_mode mem_data_fmt width width of frame in memory. fld_height height of frame in memory. stride stride of frame in memory. The idea is to combine the reading/writing of a frame and of a field in one method. According to the given value of field, p_read_comp_16() method will call the p_read_frame_comp_16() and p_read_field_comp_16(). The same holds for p_write_comp_16() method. Do you agree with adding them to cpfspd_rwi.c ? Best regards, Madalina TASS Software Professionals Ir. Madalina Iota PDEng T+31 40 2503200 | F +31 40 2503201 W www.tass.nl<http://www.tass.nl> | M mad...@ta...<mailto:mad...@ta...> Temporary address: Philips Research Location HTC Phone: +31 40 274 1604 E-mail: mad...@ph...<mailto:mad...@ph...> ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Vleuten, R. v. d. <ren...@ph...> - 2008-10-13 15:08:52
|
Hi All, Since pts is not in the public domain, let's discuss this via regular mail. Best regards, René. ________________________________ From: Iota, Madalina [mailto:mad...@ph...] Sent: Monday 13 October 2008 16:16 To: pfs...@li... Subject: [Pfs...@sf...] PTS: accept XYZ color format and float buffers Hi all, Based on the fact that cpfspd library accepts now XYZ color format and float buffers, I would like to update also the PTS-tools. Apart from this, I would like to change 1. pts_convert and add the conversions from xyz_to_xvycc and xvycc_to_xyz. xvycc files will be stored as YUV444 integer and XYZ files will be stored as XYZ integer or floats (according to the user choice). 2. pts_cmp in order to compare frame-by-frame / field-by-field, and compare all common components 3. pts_map by recreate it in C without limitations Do you have any other suggestion? Best regards, Madalina TASS Software Professionals Ir. Madalina Iota PDEng T+31 40 2503200 | F +31 40 2503201 W www.tass.nl<http://www.tass.nl> | M mad...@ta...<mailto:mad...@ta...> Temporary address: Philips Research Location HTC Phone: +31 40 274 1604 E-mail: mad...@ph...<mailto:mad...@ph...> ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Vleuten, R. v. d. <ren...@ph...> - 2008-10-13 14:43:32
|
Hi all, Ø 3. navigate through a video at each 1 second using Ctrl-PageUp and Ctrl-PageDown I like this suggestion because it would enable faster browsing on large files: the current 10 frames is really not much when browsing a large file. However, neither is 1 second. I would prefer the step to be a percentage of the number of frames in the file, e.g. such that you could step through the whole file in about 25 to 30 steps. What is your opinion about this? Best regards, René ________________________________ From: Iota, Madalina [mailto:mad...@ph...] Sent: Monday 13 October 2008 16:13 To: pfs...@li... Subject: [Pfs...@sf...] pv: add new features Hi all, As you have been informed, a new rendering mode was added to PV. This mode displays the image using the ICC profile of the panel. Currently, ICC is not fully functional, reason why I want to continue improving it. One of the improvements would be the creating of a virtual profiler for YUV images. Besides this I would like to: 1. have the behavior of ESC back in PV in order to exit the application 2. combine the new behavior of TAB with the old one (if the user does not select a certain color method, TAB shall be used to browse the images while preserving the zoom factor) 3. navigate through a video at each 1 second using Ctrl-PageUp and Ctrl-PageDown 4. have a shortcut key to set the default values of the Component Mapping 5. exclude ICC from "Color display method" when pressing "C" or "shift-C" What is your opinion about this? Best regards, Madalina TASS Software Professionals Ir. Madalina Iota PDEng T+31 40 2503200 | F +31 40 2503201 W www.tass.nl<http://www.tass.nl> | M mad...@ta...<mailto:mad...@ta...> Temporary address: Philips Research Location HTC Phone: +31 40 274 1604 E-mail: mad...@ph...<mailto:mad...@ph...> ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Iota, M. <mad...@ph...> - 2008-10-13 14:16:26
|
Hi all, Based on the fact that cpfspd library accepts now XYZ color format and float buffers, I would like to update also the PTS-tools. Apart from this, I would like to change 1. pts_convert and add the conversions from xyz_to_xvycc and xvycc_to_xyz. xvycc files will be stored as YUV444 integer and XYZ files will be stored as XYZ integer or floats (according to the user choice). 2. pts_cmp in order to compare frame-by-frame / field-by-field, and compare all common components 3. pts_map by recreate it in C without limitations Do you have any other suggestion? Best regards, Madalina TASS Software Professionals Ir. Madalina Iota PDEng T+31 40 2503200 | F +31 40 2503201 W www.tass.nl<http://www.tass.nl> | M mad...@ta...<mailto:mad...@ta...> Temporary address: Philips Research Location HTC Phone: +31 40 274 1604 E-mail: mad...@ph...<mailto:mad...@ph...> ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Iota, M. <mad...@ph...> - 2008-10-13 14:12:57
|
Hi all, As you have been informed, a new rendering mode was added to PV. This mode displays the image using the ICC profile of the panel. Currently, ICC is not fully functional, reason why I want to continue improving it. One of the improvements would be the creating of a virtual profiler for YUV images. Besides this I would like to: 1. have the behavior of ESC back in PV in order to exit the application 2. combine the new behavior of TAB with the old one (if the user does not select a certain color method, TAB shall be used to browse the images while preserving the zoom factor) 3. navigate through a video at each 1 second using Ctrl-PageUp and Ctrl-PageDown 4. have a shortcut key to set the default values of the Component Mapping 5. exclude ICC from "Color display method" when pressing "C" or "shift-C" What is your opinion about this? Best regards, Madalina TASS Software Professionals Ir. Madalina Iota PDEng T+31 40 2503200 | F +31 40 2503201 W www.tass.nl<http://www.tass.nl> | M mad...@ta...<mailto:mad...@ta...> Temporary address: Philips Research Location HTC Phone: +31 40 274 1604 E-mail: mad...@ph...<mailto:mad...@ph...> ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Iota, M. <mad...@ph...> - 2008-10-13 13:50:17
|
Hi all, I would like to have two functions to read and write image data per frame/field/component. The prototype of these functions would be: pT_status p_read_comp_16 (const char *filename, pT_header *header, int frame, int field, int comp, unsigned short *c_fld, int read_mode, int width, int fld_height, int stride); pT_status p_write_comp_16(const char *filename, pT_header *header, int frame, int field, int comp, const unsigned short *c_fld, int write_mode, int width, int fld_height, int stride); where: filename file to read/write header pointer to pT_header struct frame frame number field 0: frame, 1/2: field access. comp the component id. i.e. the index in the header structure. c_fld pointer to field/frame buffer. read_mode mem_data_fmt write_mode mem_data_fmt width width of frame in memory. fld_height height of frame in memory. stride stride of frame in memory. The idea is to combine the reading/writing of a frame and of a field in one method. According to the given value of field, p_read_comp_16() method will call the p_read_frame_comp_16() and p_read_field_comp_16(). The same holds for p_write_comp_16() method. Do you agree with adding them to cpfspd_rwi.c ? Best regards, Madalina TASS Software Professionals Ir. Madalina Iota PDEng T+31 40 2503200 | F +31 40 2503201 W www.tass.nl<http://www.tass.nl> | M mad...@ta...<mailto:mad...@ta...> Temporary address: Philips Research Location HTC Phone: +31 40 274 1604 E-mail: mad...@ph...<mailto:mad...@ph...> ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Harold S. <har...@ph...> - 2008-09-01 15:37:29
|
All, A new rendering mode has been added to PV. This mode uses the littlecms library to convert from file to display using an ICC profile. To activate: press 0 (zero) Ctrl-0 will show information about which profiles were selected. In practice, this means that if you supply the ICC profile for your monitor, the display will be the correct color. This important for perceptual tests and color matching between monitors. ICC profiles can be selected via command line options. If left blank, PV will search for the profile in 1) currect directory, 2) executable directory, 3) generate a 'virtual' profile The automatically read profiles must be named: YUV601.icm YUV709.icm sRGB.icm XYZ.icm A set of default profiles can be downloaded from http://www.littlecms.com/profiles.zip For testing purposes, precompiled Linux and cygwin versions can be found at \\cos\home\pfspd\pv\devel\ Developers: cvs update -d point CPFSPD_DIR in Make.env to the CVS version of cpfspd make Regards, Harold ___________________________________ Ing. H.A.W. Schmeitz Research Scientist, Video Processing Systems Philips Research Laboratories High Tech Campus 36 5656 AE Eindhoven, The Netherlands Tel. +31 (0)40 27 46489 har...@ph... |
From: Bram R. <bra...@nx...> - 2008-07-21 15:55:48
|
Dear all, Today I released version 3.0.1 of PV. Main changes: Bugfix release: some issues with infinite loops and screen refresh are now resolved. Download available from: https://sourceforge.net/projects/pfspd/ At NXP we have installed it at our 'savideo' account under utils/pv/release/pv-2.0.1 The 'latest' link is updated. Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... Phone: +31 40 27 25910; Fax: +31 40 27 28504 High Tech Campus 32 (floor 1, office 138) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Clemens W. <cle...@nx...> - 2008-06-09 12:52:03
|
Dear all, A bugfix for PV was checked in today: PV ends up in an infinite loop when reading a file with a component height smaller than 50. This problem is now fixed. Developers, please update your sandboxes. Regards, Clemens Wüst -- Clemens Wüst NXP Semiconductors High Tech Campus 32 (office 131) 5656 AE Eindhoven, The Netherlands tel: +31-40-27 25906 e-mail: cle...@nx... |
From: Bram R. <bra...@nx...> - 2008-05-30 11:16:16
|
Dear all, We are proud to announce new releases of PV and PP. Highlights PV release 3.0.0 (May 30, 2008) This release is almost a complete overhaul of the Pfspd Viewer. Generic component mapping added - the user is free to select any component from multi-component files. Files with many components (i.e. more than 15) are now supported. Menu's reordered for more logical partitioning. All settings can be activated via menus, also showing current settings. Some shortcut keys have changed meaning: - Key R: always cycles between default and full color range - Key 2: is now RGB color method - Keys 2-6 are now mapped on 3-7 (direct selection of color method) - Ctrl-Page-Down/Up are replaced by Shift-Page-Down/Up - Ctrl-Home/End are replaced by Shift-Home/End - ESC now gets out of full-screen mode (windows) New shortcut keys: - M to activate the component mapping dialog - I cycles the signal range value that shows max. saturation - O cycles the (overlay) components - Y cycles the vector accuracy - L cycles the arrow length for vectors - Shift-C for cycling backwards through the color methods - Shift-D for cycling backwards through the de-interlace methods - N/P Next, resp. previous file with lock (keep settings) Changed to new versions of wxWidgets and GTK (Linux) and consequently stopped support for the sgi and sun platforms. The PP release is just a recompile with the new cpfspd release, thereby adding support for pfspd files with a large number of components. Within NXP, the 'latest' links are updated at our 'savideo' account. Location: https://sourceforge.net/project/showfiles.php?group_id=137382 Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 138) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Bram R. <bra...@nx...> - 2008-05-22 11:30:45
|
Harold, others, Again a few updates on PV. We had a recent discussion how to deal with open-with-lock when selected ui components do not exitst in the new file. Previous solution: use first component. Now the default component is chosen. For this, I keep track of the default components in the pfspd module. Further, the pp_options are now in the settings, so copied with duplicating a window. I still have pending bugs: 1) full screen on zoomed image fails https://sourceforge.net/tracker/index.php?func=detail&aid=1964423&group_id=137382&atid=738511 2) resize on nexp/prev file (GTK) https://sourceforge.net/tracker/index.php?func=detail&aid=1963815&group_id=137382&atid=738511 Volunteers to fix? More issues before release? Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... <--- New address and phone since June 2007 --> Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 140) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Harold S. <har...@ph...> - 2008-05-16 13:52:22
|
Hi Bram, others, > Suggested improvement for "Next/prev file with lock": > keep current component mapping; in case selected component does > not exist in the new file, then choose the default one, similarly to the > regular defaults. Yes, this would work. > Is that what you meant? What I meant is to always reset to default mapping whever the user has not changed the mapping. To implement this we need to maintain a flag that indicates whether mapping has been modified. I also noticed that it is not possible anymore to view an RGB file as greyscale; I only see the R component. (or another component that has been mapped) This also applies to the Y+vector and Y+scalar modes. Was the conversion of RGB->Y intentionally removed ? shift-Tab works fine for me in both Linux and Cygwin versions. Still need to compile with MSVC and test. On a related note: make.mtk does not detect my Visual Studio .NET 2003. The installation of Visual Studio 6 is interfering, will need to fix this first. Regards, Harold ___________________________________ Ing. H.A.W. Schmeitz Research Scientist, Video Processing Systems Philips Research Laboratories High Tech Campus 36 5656 AE Eindhoven, The Netherlands Tel. +31 (0)40 27 46489 har...@ph... |
From: Bram R. <bra...@nx...> - 2008-05-14 16:38:29
|
Hi all, A short update on PV developments. Over the last days, Clemens and I have done many small modifications. In summary from the readme file: - Bug 1459249: mismatch max window size with dual screen (win32). The max. window size is not limited to the screen size anymore, only to the size of the (zoomed) image. - Bug 1963787: zoomin does not keep center (win32). Now solved, was handled well in canvas using '-' and '=' keys, but not on frame using the numeric '-' and '+' keys. - The Escape key has been given a new functionality. Instead of closing the active window (on all platforms), it is now used to exit fullscreen mode (MSWindows only). - Better separation of frame and canvas, in particular by shifting public variables to private. - Revised the FitCanvas stuff. - Lost of cosmetics to create a consistent code style. In the process, we found a few issues, listed as bugs in Your comments are welcome. Again, some rigorous testing by developers would be appreciated... We think this is about ready for release! Please provide your comments, I intend to create a release in a few days. Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... <--- New address and phone since June 2007 --> Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 140) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Bram R. <bra...@nx...> - 2008-05-08 18:10:16
|
Dear all, In the PV development repository, we have done a significant update. This aims at improving readibility and maintainability of the code. >From the readme file (a very short summary) - Internal code restructuring: all settings are now saved in a separate class Settings, which is implemented in a separate module. - Internal code restructuring to realize a better separation of the frame and canvas modules. Further we have updated some shortcuts to streamline shortcut definition along some simple principles. Again from the readme file: - Shortcut keys Ctrl-Page-Down/Up are replaced by Shift-Page-Down/Up to stick to the general principle that "shift" adapts the meaning of the non-shift operation. - Shortcut keys Ctrl-Home/End are replaced by Shift-Home/End (but still have the same behaviour as plain Home/End: go to the first/last frame). Developers, please update your sandboxes. And eh, ... some serious testing would be appreciated. Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... <--- New address and phone since June 2007 --> Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 140) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Bram R. <bra...@nx...> - 2008-04-28 16:06:56
|
Dear all, Today I created a new release of mtk and cpfspd. Highlights CPFSPD release 1.11.0 - Support for up to 128 components (was 15) - Standard I/O on win32 supported Highlights MTK release 5.3.3 Bugfix: Make.mtk dependency generation for MSVC compiler. At NXP we do not have access to SGI machines anymore. Therefore these releases do not contain pre-compiled sgi executables/libraries anymore. Since this platform is also phased out at Philips Research, it will not hurt too much, otherwise... the source code is available. The software can be downloaded from the public SourceForge server: https://sourceforge.net/projects/pfspd Within NXP, it is also available on our internal Gforge server https://nww.innersource.nxp.com/gf/project/pfspd For NXP research, the "latest" links are updated on our "savideo" group account. Regards, Bram. PS. Developers, please do not forget to update your sandboxes. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... <--- New address and phone since June 2007 --> Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 140) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Parys, H. v. <han...@ph...> - 2008-04-15 09:02:20
|
I wonder if there has ever been thought about color management in the pv viewer. More specifically, I mean taking into account the ICC profile of your monitor. Best regards, Hans Van Parys Senior Design Engineer Digital Video Processing Philips Consumer Lifestyle - Upmarket Displays Pathoekeweg 11, B-8000 BRUGGE Phone: +32 50 455 425 E-mail: han...@ph... |
From: Bram R. <bra...@nx...> - 2008-04-10 16:56:51
|
Hi all, Clemens and I have committed quite some changes in PV over the last days. Developers can have a look, experiment with this version and give your feedback. We think the arbitrary component mapping is now functioning well. Be aware that we had to change a few of the keyboard shortcuts. In view of the substantial change, next release will carry version number 3.0.0. We consider the code almost ready for release. Next steps: - rethink the functionality of the arrow vectors and (more importantly) the implemention (seems to relate to specific MVX/MVY components) - rethink the keyboard shortcuts for the grid and arrow vectors (currently inconsistent within the view menu) Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... <--- New address and phone since June 2007 --> Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 140) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Bram R. <bra...@nx...> - 2008-04-04 18:19:11
|
Dear all, Clemens and I made some good progress in implementation of the generic component mapping functionality in PV. Developers can now have a look in the repository and experiment with the current code. For 8 bit and 16 bit components, most functionality is expected to work well. Some work is still in progress for other word widths in the vector and scalar overlay handling. We encouter a few design issues that we would like to discuss with the community. We reported before the following summary from the Readme.txt file: - Component mapping is added. Users can now select which components to use for each of the color display methods. With this functionality files with an arbitrary number of components can now be handled well. - A new color display method "RGB" is added. Further, as part of the addition of the component mapping, the names of the color display methods are updated. The generic principle is as follows: for each of the color modes, the user can select which component to use assording to the following table: color mode component role ("use as") YUV Y, U, V RGB R, G, B Y Y Y+V Y, Vx, Vy V Vx, Vy Y+S Y, S S S The color mode RGB is new compared to the latest relase. Note 1: no RGB->Y matrix anymore In the past, an RGB file was specifically recognised and converted to Y in case of color mode "Y". Now, RGB mode always requires three components. So the functionality to convert RGB to Y is lost. Questions: is this missing functionality considered a serious loss? If yes, what is a good approach to resolve this? An extra color mode? Are we willing to pay the burden of yet an other color mode (burden in the sense of increased UI complexity, ease of use, etc)? Proposal: forget about rgb->y conversion to keep the concept simple and streamlined. Note 2: what to do with the location info? In the current release, all component values are shown in location info. This is not feasible anymore with the many components we now support. With e.g. a 20 component file, this also means that all components must be loaded, yielding a performance penalty. Secondly, some components are treated in a special way: for "MVX" and "MVY" the offset is subtraced, so binary offset representation is changed into 2's complement. Further, the subpixel accuracy is handled resulting in a float representation. With generic component mapping, we cannot select on component name. Hence an other criterium is required. Here are the options: Option 2A: display only the components of the current color mode In this case, the role of each of the components is known, and vx/vy components can be handled accordingly. Furthermore, also U and V can be converted to 2's complement (which is different from the current release, but a clear advantage). Option 2B: display all components used in any of the color modes In this case, the plain value (binary offset) as read from the file is displayed, since one component can have multiple roles. We propose to build option 2A. Note 3: vector subpixel resolution In the current release, this is hardcoded as follows: 8 bit vector components are assumed at 1/4 pixel accuracy; 16 bits vector components "MVX" and "MVY" are assumed at 1/16 pixel accuracy. We propose to add a menu entry for "accuracy" (character 'a') similar to the vector range. Default is 1/4 if both vector components are 8 bits, otherwise 1/16. This vector accuracy is used both in the overlay and in the location info (provided 2A is chosen). BTW. Multiplexed components are handled fully transparantly as separate components in the user interface. Your thoughts are appreciated. Kind regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... <--- New address and phone since June 2007 --> Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 140) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |
From: Clemens W. <cle...@nx...> - 2008-03-19 16:55:03
|
Dear all, Today, we committed the user interface for the component mapping in PV into the development repository. Next action is to add the functionality to this new user interface ;-) We expect this to be finished somewhere next week. Developers are invited to have a look into this new user interface! Of coarse, update your development environment to do so... Excerpt from the Readme.txt file: - Component mapping is added. Users can now select which components to use for each of the color display methods. With this functionality files with an arbitrary number of components can now be handled well. - A new color display method "RGB" is added. Further, as part of the addition of the component mapping, the names of the color display methods are updated. More extended explanation of component mapping in the Readme.txt file: Component mapping The component mapping dialog (shortcut key 'm') allows the choice of arbitrary components to be used in each of the color display methods. E.g. in RGB color display method, components with names X, Y and Z may be selected for use as R, G and B components. E.g. in Y with scalar display method, component names Y and DEPTH may be selected for use as Y and Scalar components. The following table lists the default selection for each of the color display methods. YUV color method -- select components for Y, U, and V: - if there are less than three components available: use first component as Y and as U and as V - else, if components with names "Y", "U", and "V" are available: use component "Y" as Y use component "U" as U use component "V" as V - else: use first component as Y use second component as U use third component as V RGB color method -- select components for R, G, and B: - if there are less than three components available: use first component as R and as G and as B - else, if components with names "R", "G", and "B" are available: use component "R" as R use component "G" as G use component "B" as B - else: use first component as R use second component as G use third component as B Y grayscale method -- select component for Y: - if a component with name "Y" is available: use component "Y" as Y - else: use first component as Y Y with vector method -- select components for Y, Vx, and Vy: - if there are less than three components available: use first component as Y and as Vx and as Vy - else, if components with names "Y", "MVx", and "MVy" are available: use component "Y" as Y use component "MVx" as Vx use component "MVy" as Vy - else: use first component as Y use second component as Vx use third component as Vy Only vector method -- select components for Vx and Vy: - if there are less than two components available: use first component as Vx and as Vy - else, if components with names "MVx" and "MVy" are available: use component "MVx" as Vx use component "MVy" as Vy - else: use first component as Vx use second component as Vy Y with scalar method -- select components for Y and S: - if there are less than two components available, or if no component with name "Y" is available: use first component as Y and as S - else, if there is a non-standard component X available (i.e., not with name "Y", "U", "V", "R", "G", or "B"), and if the name of X does not end with "Vx" or "Vy": use component "Y" as Y use the first component X as S - else, if there is a non-standard component X available (i.e., not with name "Y", "U", "V", "R", "G", or "B"): use component "Y" as Y use the first component X as S - else: use component "Y" as Y use first component as S Only scalar method -- select components for S: - if there is a non-standard component X available (i.e., not with name "Y", "U", "V", "R", "G", or "B"), and if the name of X does not end with "Vx" or "Vy": use the first component X as S - else, if there is a non-standard component X available (i.e., not with name "Y", "U", "V", "R", "G", or "B"): use the first component X as S - else: use first component as S Matching of component names is not case-sensitive. With kind regards, Clemens Wüst and Bram Riemens -- Clemens Wüst NXP Semiconductors High Tech Campus 32 (office 131) 5656 AE Eindhoven, The Netherlands tel: +31-40-27 25906 e-mail: cle...@nx... |
From: Bram R. <bra...@nx...> - 2008-02-29 14:54:10
|
Hi all, Today, we updated the maximum number of components that cpfspd supports. It is changed from 15 to 128. This is available in the development environment. Developers, please update your source tree. Regards, Bram. -- Bram Riemens, Senior Principal NXP Semiconductors / Corporate I&T / Research Email: bra...@nx... <--- New address and phone since June 2007 --> Phone: +31 40 27 25910; Fax: +31 40 27 44639 High Tech Campus 32 (floor 1, office 140) 5656 AE Eindhoven, The Netherlands The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. |