rtcontrib output crashes preview, finder

Hi,

I'm working through Axel Jacobs rtcontrib tutorial (on OSX). I'm up to the first use of tregenza.cal to generate sky patches (command below) and finding that the first image, p000.hdr, is corrupted somehow. Preview crashes when I try to open the image and when I even select it in the Finder, the Finder behaves very strangely (I think it's crashing and restarting all in the space of about two seconds). Getinfo reports what looks like a correct header, comparing the headers of p000.hdr and p001.hdr. The files are all the same size. The other images, p001.hdr to p145.hdr all open fine.

$vwrays -ff $vw |rtcontrib @rtc.opt -ffc $(vwrays -d $vw) -V+ -f tregenza.cal -b tbin -o images/patches/p%03d.hdr -m sky_glow testroom_whitesky.oct

Any hints?

John Ford.

Have you tried opening the image in ximage via Terminal.app?

The Finder preview and the Radiance use different libraries to
open/show HDR images.

Regards,
Thomas

···

On Thu, Dec 9, 2010 at 9:07 PM, John Ford <[email protected]> wrote:

Hi,

I'm working through Axel Jacobs rtcontrib tutorial (on OSX). I'm up to the first use of tregenza.cal to generate sky patches (command below) and finding that the first image, p000.hdr, is corrupted somehow. Preview crashes when I try to open the image and when I even select it in the Finder, the Finder behaves very strangely (I think it's crashing and restarting all in the space of about two seconds). Getinfo reports what looks like a correct header, comparing the headers of p000.hdr and p001.hdr. The files are all the same size. The other images, p001.hdr to p145.hdr all open fine.

$vwrays -ff $vw |rtcontrib @rtc.opt -ffc $(vwrays -d $vw) -V+ -f tregenza.cal -b tbin -o images/patches/p%03d.hdr -m sky_glow testroom_whitesky.oct

Any hints?

John Ford.

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

Thomas,

Yes, it opens in ximage. Should tregenza bin 0 be all black?

I'm trying to do the next step which is combining the images into a view of the patches using pcomb. This doesn't seem to work either using images 000-145 or images 001-145 so I don't know if there's some other problem. Terminal just hangs when I send the command, until I ctrl-c it.

I will keep fiddling with it.

John.

JOHN FORD

BLUEBOTTLE
FORTY THREE STAWELL STREET
NORTH MELBOURNE 3051
PHONE 03 9329 0709
FAX 03 9329 1780

WWW.BB3.NET.AU
[email protected]
MOBILE 0411 519 738

···

On 10/12/2010, at 1:28 PM, Thomas Bleicher wrote:

Have you tried opening the image in ximage via Terminal.app?

The Finder preview and the Radiance use different libraries to
open/show HDR images.

Regards,
Thomas

On Thu, Dec 9, 2010 at 9:07 PM, John Ford <[email protected]> wrote:

Hi,

I'm working through Axel Jacobs rtcontrib tutorial (on OSX). I'm up to the first use of tregenza.cal to generate sky patches (command below) and finding that the first image, p000.hdr, is corrupted somehow. Preview crashes when I try to open the image and when I even select it in the Finder, the Finder behaves very strangely (I think it's crashing and restarting all in the space of about two seconds). Getinfo reports what looks like a correct header, comparing the headers of p000.hdr and p001.hdr. The files are all the same size. The other images, p001.hdr to p145.hdr all open fine.

$vwrays -ff $vw |rtcontrib @rtc.opt -ffc $(vwrays -d $vw) -V+ -f tregenza.cal -b tbin -o images/patches/p%03d.hdr -m sky_glow testroom_whitesky.oct

Any hints?

John Ford.

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

Evening John.

I think bin 0 was "reserved" for the ground. Therefore, if you don't have a
ground material, it would be black. In my own tests I also found other bins
to be completely black but I was told that was because of my primitive test
geometry. Doing Axel's tutorial test room seemed to work for me at the time.
I could reproduce the results he got with similar accuracy.

About pcomb: It may take quite a long time to process all the images. You
can try and split the one long pcomb command into a few smaller pcombs and
then combine the resulting images. If you suspect a single image to be
corrupt just exclude that from the process and see if the pcomb command
finishes.

If you are familiar with bash scripting you can run a first loop and combine
each images separately with a known good image. This should at least show
you if there are images that pcomb can't handle.

Regards,
Thomas

···

On Thu, Dec 9, 2010 at 10:53 PM, John Ford <[email protected]> wrote:

Thomas,

Yes, it opens in ximage. Should tregenza bin 0 be all black?

I'm trying to do the next step which is combining the images into a view of
the patches using pcomb. This doesn't seem to work either using images
000-145 or images 001-145 so I don't know if there's some other problem.
Terminal just hangs when I send the command, until I ctrl-c it.

I will keep fiddling with it.

John.

JOHN FORD

BLUEBOTTLE
FORTY THREE STAWELL STREET
NORTH MELBOURNE 3051
PHONE 03 9329 0709
FAX 03 9329 1780

WWW.BB3.NET.AU <http://www.bb3.net.au/>
[email protected] <[email protected]>
MOBILE 0411 519 738

On 10/12/2010, at 1:28 PM, Thomas Bleicher wrote:

Have you tried opening the image in ximage via Terminal.app?

The Finder preview and the Radiance use different libraries to
open/show HDR images.

Regards,
Thomas

On Thu, Dec 9, 2010 at 9:07 PM, John Ford <[email protected]> wrote:

Hi,

I'm working through Axel Jacobs rtcontrib tutorial (on OSX). I'm up to the
first use of tregenza.cal to generate sky patches (command below) and
finding that the first image, p000.hdr, is corrupted somehow. Preview
crashes when I try to open the image and when I even select it in the
Finder, the Finder behaves very strangely (I think it's crashing and
restarting all in the space of about two seconds). Getinfo reports what
looks like a correct header, comparing the headers of p000.hdr and p001.hdr.
The files are all the same size. The other images, p001.hdr to p145.hdr all
open fine.

$vwrays -ff $vw |rtcontrib @rtc.opt -ffc $(vwrays -d $vw) -V+ -f
tregenza.cal -b tbin -o images/patches/p%03d.hdr -m sky_glow
testroom_whitesky.oct

Any hints?

John Ford.

_______________________________________________

Radiance-general mailing list

[email protected]

http://www.radiance-online.org/mailman/listinfo/radiance-general

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

Thomas,

Got it working.

In the meantime discovered that date in OSX doesn't return nanoseconds, despite Unix date returning nanoseconds. Also, I don't think 'recno' works on OSX.

John.

···

On 10/12/2010, at 3:12 PM, Thomas Bleicher wrote:

Evening John.

I think bin 0 was "reserved" for the ground. Therefore, if you don't have a ground material, it would be black. In my own tests I also found other bins to be completely black but I was told that was because of my primitive test geometry. Doing Axel's tutorial test room seemed to work for me at the time. I could reproduce the results he got with similar accuracy.

About pcomb: It may take quite a long time to process all the images. You can try and split the one long pcomb command into a few smaller pcombs and then combine the resulting images. If you suspect a single image to be corrupt just exclude that from the process and see if the pcomb command finishes.

If you are familiar with bash scripting you can run a first loop and combine each images separately with a known good image. This should at least show you if there are images that pcomb can't handle.

Regards,
Thomas

On Thu, Dec 9, 2010 at 10:53 PM, John Ford <[email protected]> wrote:
Thomas,

Yes, it opens in ximage. Should tregenza bin 0 be all black?

I'm trying to do the next step which is combining the images into a view of the patches using pcomb. This doesn't seem to work either using images 000-145 or images 001-145 so I don't know if there's some other problem. Terminal just hangs when I send the command, until I ctrl-c it.

I will keep fiddling with it.

John.

JOHN FORD

BLUEBOTTLE
FORTY THREE STAWELL STREET
NORTH MELBOURNE 3051
PHONE 03 9329 0709
FAX 03 9329 1780

WWW.BB3.NET.AU
[email protected]
MOBILE 0411 519 738

On 10/12/2010, at 1:28 PM, Thomas Bleicher wrote:

Have you tried opening the image in ximage via Terminal.app?

The Finder preview and the Radiance use different libraries to
open/show HDR images.

Regards,
Thomas

On Thu, Dec 9, 2010 at 9:07 PM, John Ford <[email protected]> wrote:

Hi,

I'm working through Axel Jacobs rtcontrib tutorial (on OSX). I'm up to the first use of tregenza.cal to generate sky patches (command below) and finding that the first image, p000.hdr, is corrupted somehow. Preview crashes when I try to open the image and when I even select it in the Finder, the Finder behaves very strangely (I think it's crashing and restarting all in the space of about two seconds). Getinfo reports what looks like a correct header, comparing the headers of p000.hdr and p001.hdr. The files are all the same size. The other images, p001.hdr to p145.hdr all open fine.

$vwrays -ff $vw |rtcontrib @rtc.opt -ffc $(vwrays -d $vw) -V+ -f tregenza.cal -b tbin -o images/patches/p%03d.hdr -m sky_glow testroom_whitesky.oct

Any hints?

John Ford.

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

Hi John,

The "recno" variable should work in rcalc, if that's what you're talking about. I use it all the time. The Finder isn't very bright when it comes to HDR images -- long headers may confuse it.

Use the new dctimestep rather than pcomb to combine images. If you don't have the new version of dctimestep, use the -h option to pcomb to avoid ridiculous header lengths.

Cheers,
-Greg

···

From: John Ford <[email protected]>
Date: December 9, 2010 10:04:50 PM PST

Thomas,

Got it working.

In the meantime discovered that date in OSX doesn't return nanoseconds, despite Unix date returning nanoseconds. Also, I don't think 'recno' works on OSX.

John.

Hi John,

I find that completely black hdr images consistently crash preview. It will also crash finder if you use finder in one of the modes that show a preview of the image like cover flow or the columns.

Also, I've never been able to get Axel's command for random color images to work on a mac the way it is written in the tutorial. When I try it rcalc waits patiently for input that doesn't exist. But you can't add -n because then recno is invalid. If you give it simple input using cnt (ie. cnt 1 | rcalc ) it doesn't work because all the values are seeded with the same number (recno is 1 every time). Instead I used the $RANDOM variable to seed rcalc's rand function like so:

pcomb $(for i in {000..145}; do \
    echo "-c $(echo $RANDOM | rcalc -e '$1=rand($1);$2=rand($1+.3);$3=rand($1+.7)') \
    images/patches/p$(printf "%03d" $i).hdr"; done)> images/sky_colourcomb.hdr

Admittedly this solution is not elegant. If it's elegance we're after we could use the new capability in dctimestep (which wasn't available when Axel wrote his tutorial):
cnt 145 | rcalc -e '$1=rand(recno);$2=rand(recno+.3);$3=rand(recno+.7)' | dctimestep images/patches/p%03d.hdr > images/sky_colourcomb.hdr

Best,
Andy

···

On Dec 9, 2010, at 11:24 PM, Greg Ward wrote:

Hi John,

The "recno" variable should work in rcalc, if that's what you're talking about. I use it all the time. The Finder isn't very bright when it comes to HDR images -- long headers may confuse it.

Use the new dctimestep rather than pcomb to combine images. If you don't have the new version of dctimestep, use the -h option to pcomb to avoid ridiculous header lengths.

Cheers,
-Greg

From: John Ford <[email protected]>
Date: December 9, 2010 10:04:50 PM PST

Thomas,

Got it working.

In the meantime discovered that date in OSX doesn't return nanoseconds, despite Unix date returning nanoseconds. Also, I don't think 'recno' works on OSX.

John.

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general