ARToolKit | Mailing List Archive |
![]() |
From: | Brian Furtaw <brian@s ......> | Received: | Mar 14, 2001 |
To | FORUM ON AUGMENTED AND VIRTUAL REALITY <ARforum@l ..........> | ||
Subject: | Re: [ARforum] TCM2 Magnetic Compass Issues | ||
FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Build a marketing database and send targeted HTML and text e-mail newsletters to your customers with List Builder. http://www.listbuilder.com ---------------------------------------------------------------------- Faster then glCopyTexImage2D? glCopyTexSubImage2D should have less overhead then glCopyTexImage2D. Use glCopyTexImage2D to create the texture at application startup then use glCopyTexSubImage2D to copy the framebuffer data at run time. If you have an SGI O2 dmBuffers (not part of OpenGL API) require only a pointer switch to move data from framebuffer to texture memory. Perhaps some of these other GuMA style graphics cards have implemented such and interface. Brian Oliver Bimber wrote: > > FORUM ON AUGMENTED AND VIRTUAL REALITY > > --------------------------- ListBot Sponsor -------------------------- > Start Your Own FREE Email List at http://www.listbot.com/links/joinlb > ---------------------------------------------------------------------- > > hi, > does anybody know a fast(er) OpenGL-way of copying the frame-buffer content into > the texture memory (fatser that glCopy..)? > --oliver > > ______________________________________________________________________ > To unsubscribe, write to ARforum-unsubscribe@l .......... -- ----oOOo---- ----oOOo---- ----oOOo---- ----oOOo---- Brian Furtaw (brian@s ......) Graphics Guru Office:(301)572-3293 Fax: (603)250-0370 12200-G Plum Orchard Drive OpenGL/Performer/OpenInventor/ImageVision Silver Spring, Maryland 20904 Volumizer/Optimizer/React/PCI Device Drivers ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... |
From: | Brian Furtaw <brian@s ......> | Received: | Mar 14, 2001 |
To | FORUM ON AUGMENTED AND VIRTUAL REALITY <ARforum@l ..........> | ||
Subject: | Re: [ARforum] TCM2 Magnetic Compass Issues | ||
FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Build a marketing database and send targeted HTML and text e-mail newsletters to your customers with List Builder. http://www.listbuilder.com ---------------------------------------------------------------------- It is fine to copy the change the entire contents of the texture with glCopyTexSubImage2D the important point here is that you are not redefining the texture every frame with glCopyTexImage2D. glCopyTexImage2D allocates a glHandle and texture memory, glCopyTexSubImage2D just replace the previous frames image data with new image data in texture memory. Brian Oliver Bimber wrote: > > FORUM ON AUGMENTED AND VIRTUAL REALITY > > --------------------------- ListBot Sponsor -------------------------- > Build a marketing database and send targeted HTML and text e-mail > newsletters > to your customers with List Builder. > http://www.listbuilder.com > ---------------------------------------------------------------------- > > brian, > thanks! > > > Faster then glCopyTexImage2D? glCopyTexSubImage2D should have less > > overhead then glCopyTexImage2D. > > but i need to copy the entire frame-buffer - not just a portion. > > > Use glCopyTexImage2D to create the > > texture at application startup then use glCopyTexSubImage2D to copy the > > framebuffer data at run time. > > since i use a multi-pass technique, i need to do this every frame (i.e. the > frame-buffer content) changes all the time, and so does the texture. > > > If you have an SGI O2 dmBuffers (not part > > of OpenGL API) require only a pointer switch to move data from > > framebuffer to texture memory. Perhaps some of these other GuMA style > > graphics cards have implemented such and interface. > > does anybody know of other interfaces like this for PC cards? > > --oliver > > > > > Brian > > > > Oliver Bimber wrote: > > > > > > FORUM ON AUGMENTED AND VIRTUAL REALITY > > > > > > --------------------------- ListBot Sponsor -------------------------- > > > Start Your Own FREE Email List at http://www.listbot.com/links/joinlb > > > ---------------------------------------------------------------------- > > > > > > hi, > > > does anybody know a fast(er) OpenGL-way of copying the frame-buffer content into > > > the texture memory (fatser that glCopy..)? > > > --oliver > > > > > > ______________________________________________________________________ > > > To unsubscribe, write to ARforum-unsubscribe@l .......... > > > > -- > > ----oOOo---- ----oOOo---- ----oOOo---- ----oOOo---- > > > > Brian Furtaw (brian@s ......) > > Graphics Guru Office:(301)572-3293 Fax: (603)250-0370 > > 12200-G Plum Orchard Drive OpenGL/Performer/OpenInventor/ImageVision > > Silver Spring, Maryland 20904 Volumizer/Optimizer/React/PCI Device > > Drivers > > > > ______________________________________________________________________ > > To unsubscribe, write to ARforum-unsubscribe@l .......... > > ______________________________________________________________________ > To unsubscribe, write to ARforum-unsubscribe@l .......... -- ----oOOo---- ----oOOo---- ----oOOo---- ----oOOo---- Brian Furtaw (brian@s ......) Graphics Guru Office:(301)572-3293 Fax: (603)250-0370 12200-G Plum Orchard Drive OpenGL/Performer/OpenInventor/ImageVision Silver Spring, Maryland 20904 Volumizer/Optimizer/React/PCI Device Drivers ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... |
From: | "Yohan Baillot" <baillot@a ...............> | Received: | Mar 14, 2001 |
To | "FORUM ON AUGMENTED AND VIRTUAL REALITY" <ARforum@l ..........> | ||
Subject: | RE: [ARforum] TCM2 Magnetic Compass Issues | ||
FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Build a marketing database and send targeted HTML and text e-mail newsletters to your customers with List Builder. http://www.listbuilder.com ---------------------------------------------------------------------- Hello Wayne, unfortunatelly good sensors are the only solution to your problem I think. The IS-300 is about $3000 and it is worth it. But even though we still get good 3-5 degree errors when we are close to car, this is 1/6 of the fov of the display we use and it can be very noticeable. Cheap hardware vision tracking needs to be developed and a lot of people work on now. Yohan _______________________________________________________________________ Yohan BAILLOT Electrical and Computer Engineer Virtual Reality Laboratory, Advanced Information Technology (Code 5580), Naval Research Laboratory, 4555 Overlook Avenue SW, Washington, DC 20375-5337 baillot@a ............... (202) 404 7801 Home : (703) 566 3684 Cell : (407) 222 6359 Fax : (202) 767 1122 Web : www.ait.nrl.navy.mil/people/baillot _______________________________________________________________________ | -----Original Message----- | From: Wayne Piekarski [mailto:wayne@c ..............] | Sent: Wednesday, March 14, 2001 3:27 AM | To: ARforum@l .......... | Subject: [ARforum] TCM2 Magnetic Compass Issues | | | FORUM ON AUGMENTED AND VIRTUAL REALITY | | --------------------------- ListBot Sponsor -------------------------- | Build a marketing database and send targeted HTML and text e-mail | newsletters | to your customers with List Builder. | http://www.listbuilder.com | ---------------------------------------------------------------------- | | Hi everyone! | | I've been using the TCM2 compass from Precision Navigation for | quite a while, | and I was wanting to know if anyone else has used them much for AR | applications. To date, all my work has been using them, but they | aren't all | that fantastic due to their noise and jitter. | | If you read the specs, they talk about errors +- 1 degree or | whatever, but the | output I am seeing is a *lot* larger than this. If you hold the | compass still, | and tap it slightly, or sometimes rotate it a little bit, the | heading will jump | by 20 or more degrees and then return back. An offset of 20 or 40 | degrees is | totally useless, I can get a better value typing the values in on | the keyboard! | | I understand that there might be errors caused by local magnetic | objects, but | if the compass is only being rotated slightly then this should not happen | should it? I have observed large objects causing distortions, but | these tend to | be very smooth as you move closer and closer to the object. | | The compass I have (a TCM2-80) has no fluid filled sensor, it | uses what looks | like solid state ones too, so I can't see how the tapping can cause this | problem. The case is plastic, with a cable going to the PC. There are four | small screws to hold it closed, and the serial connector is | obviously metal as | well. I've played with the calibration, which fixes up some | errors, but it is | still not very good. I run the compass at full 16 Hz speed. Maybe | I haven't | done something that I'm missing? | | I've seen some of the work by Ron Azuma where he talked about calibrating | compasses and using Kalman filters, but this problem I'm having | seems to be a | bit beyond that. The output is all over the place so even a | filter isn't going | to help me here. | | If anyone has any advice, traps that they know of, or suggestions | I would be | very grateful. If I can't get this to work really well, I was thinking of | looking into an IS-300 or something, but they are really expensive and I | already have these TCM2 units so was wanting to use them as best | as possible. | | thanks, | Wayne | | ------------------------------------------------------------------ | ------------- | Wayne Piekarski - PhD Student / Lecturer Phone: | +61-8-8302-3669 | Advanced Computing Research Centre Fax: | +61-8-8302-3381 | University of South Australia Mobile: | 0407-395-889 | Internet: | wayne@c .............. | Research & Development Manager | wayne@s ........... | SE Network Access Pty Ltd http://www.cs.unisa.edu.au/~ciswp ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... |
From: | "Oliver Bimber" <obimber@r .................> | Received: | Mar 14, 2001 |
To | FORUM ON AUGMENTED AND VIRTUAL REALITY <ARforum@l ..........> | ||
Subject: | Re: [ARforum] TCM2 Magnetic Compass Issues | ||
FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Start Your Own FREE Email List at http://www.listbot.com/links/joinlb ---------------------------------------------------------------------- hi, does anybody know a fast(er) OpenGL-way of copying the frame-buffer content into the texture memory (fatser that glCopy..)? --oliver ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... |
From: | "Greg Erdmann" <gerdmann@a .................> | Received: | Mar 14, 2001 |
To | "FORUM ON AUGMENTED AND VIRTUAL REALITY" <ARforum@l ..........> | ||
Subject: | RE: [ARforum] TCM2 Magnetic Compass Issues | ||
FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Build a marketing database and send targeted HTML and text e-mail newsletters to your customers with List Builder. http://www.listbuilder.com ---------------------------------------------------------------------- As an alternative, Ascension's 3DBIRD cost half that amount and is comparable to the IS-300. It too might be worth looking into. See: http://www.ascension-tech.com/products/3dbird/ For more information or contact me directly. Greg Erdmann greg@a ................. -----Original Message----- From: Yohan Baillot [mailto:baillot@a ...............] Sent: Wednesday, March 14, 2001 11:54 AM To: FORUM ON AUGMENTED AND VIRTUAL REALITY Subject: RE: [ARforum] TCM2 Magnetic Compass Issues FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Build a marketing database and send targeted HTML and text e-mail newsletters to your customers with List Builder. http://www.listbuilder.com ---------------------------------------------------------------------- Hello Wayne, unfortunatelly good sensors are the only solution to your problem I think. The IS-300 is about $3000 and it is worth it. But even though we still get good 3-5 degree errors when we are close to car, this is 1/6 of the fov of the display we use and it can be very noticeable. Cheap hardware vision tracking needs to be developed and a lot of people work on now. Yohan _______________________________________________________________________ Yohan BAILLOT Electrical and Computer Engineer Virtual Reality Laboratory, Advanced Information Technology (Code 5580), Naval Research Laboratory, 4555 Overlook Avenue SW, Washington, DC 20375-5337 baillot@a ............... (202) 404 7801 Home : (703) 566 3684 Cell : (407) 222 6359 Fax : (202) 767 1122 Web : www.ait.nrl.navy.mil/people/baillot _______________________________________________________________________ | -----Original Message----- | From: Wayne Piekarski [mailto:wayne@c ..............] | Sent: Wednesday, March 14, 2001 3:27 AM | To: ARforum@l .......... | Subject: [ARforum] TCM2 Magnetic Compass Issues | | | FORUM ON AUGMENTED AND VIRTUAL REALITY | | --------------------------- ListBot Sponsor -------------------------- | Build a marketing database and send targeted HTML and text e-mail | newsletters | to your customers with List Builder. | http://www.listbuilder.com | ---------------------------------------------------------------------- | | Hi everyone! | | I've been using the TCM2 compass from Precision Navigation for | quite a while, | and I was wanting to know if anyone else has used them much for AR | applications. To date, all my work has been using them, but they | aren't all | that fantastic due to their noise and jitter. | | If you read the specs, they talk about errors +- 1 degree or | whatever, but the | output I am seeing is a *lot* larger than this. If you hold the | compass still, | and tap it slightly, or sometimes rotate it a little bit, the | heading will jump | by 20 or more degrees and then return back. An offset of 20 or 40 | degrees is | totally useless, I can get a better value typing the values in on | the keyboard! | | I understand that there might be errors caused by local magnetic | objects, but | if the compass is only being rotated slightly then this should not happen | should it? I have observed large objects causing distortions, but | these tend to | be very smooth as you move closer and closer to the object. | | The compass I have (a TCM2-80) has no fluid filled sensor, it | uses what looks | like solid state ones too, so I can't see how the tapping can cause this | problem. The case is plastic, with a cable going to the PC. There are four | small screws to hold it closed, and the serial connector is | obviously metal as | well. I've played with the calibration, which fixes up some | errors, but it is | still not very good. I run the compass at full 16 Hz speed. Maybe | I haven't | done something that I'm missing? | | I've seen some of the work by Ron Azuma where he talked about calibrating | compasses and using Kalman filters, but this problem I'm having | seems to be a | bit beyond that. The output is all over the place so even a | filter isn't going | to help me here. | | If anyone has any advice, traps that they know of, or suggestions | I would be | very grateful. If I can't get this to work really well, I was thinking of | looking into an IS-300 or something, but they are really expensive and I | already have these TCM2 units so was wanting to use them as best | as possible. | | thanks, | Wayne | | ------------------------------------------------------------------ | ------------- | Wayne Piekarski - PhD Student / Lecturer Phone: | +61-8-8302-3669 | Advanced Computing Research Centre Fax: | +61-8-8302-3381 | University of South Australia Mobile: | 0407-395-889 | Internet: | wayne@c .............. | Research & Development Manager | wayne@s ........... | SE Network Access Pty Ltd http://www.cs.unisa.edu.au/~ciswp ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... |
From: | "Oliver Bimber" <obimber@r .................> | Received: | Mar 14, 2001 |
To | FORUM ON AUGMENTED AND VIRTUAL REALITY <ARforum@l ..........> | ||
Subject: | Re: [ARforum] TCM2 Magnetic Compass Issues | ||
FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Build a marketing database and send targeted HTML and text e-mail newsletters to your customers with List Builder. http://www.listbuilder.com ---------------------------------------------------------------------- brian, thanks! > Faster then glCopyTexImage2D? glCopyTexSubImage2D should have less > overhead then glCopyTexImage2D. but i need to copy the entire frame-buffer - not just a portion. > Use glCopyTexImage2D to create the > texture at application startup then use glCopyTexSubImage2D to copy the > framebuffer data at run time. since i use a multi-pass technique, i need to do this every frame (i.e. the frame-buffer content) changes all the time, and so does the texture. > If you have an SGI O2 dmBuffers (not part > of OpenGL API) require only a pointer switch to move data from > framebuffer to texture memory. Perhaps some of these other GuMA style > graphics cards have implemented such and interface. does anybody know of other interfaces like this for PC cards? --oliver > > Brian > > Oliver Bimber wrote: > > > > FORUM ON AUGMENTED AND VIRTUAL REALITY > > > > --------------------------- ListBot Sponsor -------------------------- > > Start Your Own FREE Email List at http://www.listbot.com/links/joinlb > > ---------------------------------------------------------------------- > > > > hi, > > does anybody know a fast(er) OpenGL-way of copying the frame-buffer content into > > the texture memory (fatser that glCopy..)? > > --oliver > > > > ______________________________________________________________________ > > To unsubscribe, write to ARforum-unsubscribe@l .......... > > -- > ----oOOo---- ----oOOo---- ----oOOo---- ----oOOo---- > > Brian Furtaw (brian@s ......) > Graphics Guru Office:(301)572-3293 Fax: (603)250-0370 > 12200-G Plum Orchard Drive OpenGL/Performer/OpenInventor/ImageVision > Silver Spring, Maryland 20904 Volumizer/Optimizer/React/PCI Device > Drivers > > ______________________________________________________________________ > To unsubscribe, write to ARforum-unsubscribe@l .......... ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... |
From: | "Wayne Piekarski" <wayne@c ..............> | Received: | Mar 14, 2001 |
To | ARforum@l .......... | ||
Subject: | [ARforum] TCM2 Magnetic Compass Issues | ||
FORUM ON AUGMENTED AND VIRTUAL REALITY --------------------------- ListBot Sponsor -------------------------- Build a marketing database and send targeted HTML and text e-mail newsletters to your customers with List Builder. http://www.listbuilder.com ---------------------------------------------------------------------- Hi everyone! I've been using the TCM2 compass from Precision Navigation for quite a while, and I was wanting to know if anyone else has used them much for AR applications. To date, all my work has been using them, but they aren't all that fantastic due to their noise and jitter. If you read the specs, they talk about errors +- 1 degree or whatever, but the output I am seeing is a *lot* larger than this. If you hold the compass still, and tap it slightly, or sometimes rotate it a little bit, the heading will jump by 20 or more degrees and then return back. An offset of 20 or 40 degrees is totally useless, I can get a better value typing the values in on the keyboard! I understand that there might be errors caused by local magnetic objects, but if the compass is only being rotated slightly then this should not happen should it? I have observed large objects causing distortions, but these tend to be very smooth as you move closer and closer to the object. The compass I have (a TCM2-80) has no fluid filled sensor, it uses what looks like solid state ones too, so I can't see how the tapping can cause this problem. The case is plastic, with a cable going to the PC. There are four small screws to hold it closed, and the serial connector is obviously metal as well. I've played with the calibration, which fixes up some errors, but it is still not very good. I run the compass at full 16 Hz speed. Maybe I haven't done something that I'm missing? I've seen some of the work by Ron Azuma where he talked about calibrating compasses and using Kalman filters, but this problem I'm having seems to be a bit beyond that. The output is all over the place so even a filter isn't going to help me here. If anyone has any advice, traps that they know of, or suggestions I would be very grateful. If I can't get this to work really well, I was thinking of looking into an IS-300 or something, but they are really expensive and I already have these TCM2 units so was wanting to use them as best as possible. thanks, Wayne ------------------------------------------------------------------------------- Wayne Piekarski - PhD Student / Lecturer Phone: +61-8-8302-3669 Advanced Computing Research Centre Fax: +61-8-8302-3381 University of South Australia Mobile: 0407-395-889 Internet: wayne@c .............. Research & Development Manager wayne@s ........... SE Network Access Pty Ltd http://www.cs.unisa.edu.au/~ciswp ______________________________________________________________________ To unsubscribe, write to ARforum-unsubscribe@l .......... |