Open GoogleCodeExporter opened 8 years ago
Same as the issue reported by taka.atsushi at [Comments]?
Original comment by tom...@gmail.com
on 12 Jan 2011 at 12:26
My issue is similar, but not quite the same.
I am running Win7 x64 and the utility freezes sometimes before and sometimes
after the calibration stage. I modified the code a little bit and it made it
to 1 successful calibration attempt, and then stopped working again.
Any ideas?
Original comment by dothebar...@gmail.com
on 12 Jan 2011 at 6:26
dothebartman, so you mean:
- The image window (not console window) opens up successfully
- The image window is working (refreshing) as expected for a while
- The image window suddenly stops working at some point
- Leaving it for some time, it is still freezing without starting again or
crushing eventually
Am I correct?
If so, does it happen regardless of the motion in the vision of the camera, or
does it seem to be triggered by some motion?
I do not have any ideas off the top of my head.
Original comment by tom...@gmail.com
on 13 Jan 2011 at 7:23
tkabataさん、もう少し情報いただけませんか。
例えばComment
3に書いたような観点でどんなStopの仕方かを教えてください
。
Original comment by tom...@gmail.com
on 13 Jan 2011 at 8:54
You are correct, and the image window freezes regardless of whether or not
there is movement.
I will try some stuff out today and see if I can make any progress.
Let me know if any ideas come to you at some point.
Thanks
Original comment by dothebar...@gmail.com
on 13 Jan 2011 at 9:11
Thanks for your help. I will think around if there are any behaviors that could
make undesirable cumulative effect. (e.g. memory leaks)
Original comment by tom...@gmail.com
on 13 Jan 2011 at 9:49
Hello, i have the same problem than ykabata, the application crashes at henshin
stage = human.
Configuration: Laptop, Win7 Home Edition, Graphic integrated Mobile Intel 965
Express.
Greetings,
Ismael
Original comment by Sharkar...@gmail.com
on 13 Jan 2011 at 11:02
Ismael,
Yesterday I found version 0.1 and 0.1a rashes when your graphic configuration
does not support "shader". The latest version, 0.1b, checks your graphic
capability and reports an error (instead of crashing) if it is the case. Can
you run 0.1b and see whether it reports an error or it still crashes? I assume
graphic chips integrated with laptops typically does not have much
capabilities, so this is likely to be the cause.
Thanks for your help!
Original comment by tom...@gmail.com
on 13 Jan 2011 at 11:54
Well, is it possible that shader is not available through OpenGL although your
GPU supports it? I (not a graphic expert) am getting confused.
Original comment by tom...@gmail.com
on 14 Jan 2011 at 12:48
Thanks for your response! i checked that my graphic card supports shader and i
also installed the lattest version, but it still crashes, i'll add more
information about my configuration in the attached file.
Ismael.
Original comment by Sharkar...@gmail.com
on 14 Jan 2011 at 12:53
Attachments:
Thanks, Ismael. If you got a crash with 0.1b, yes, it is definitely a different
issue from the shader one. Looking back to your issue description, the program
goes at least to the point to display "Hensin Stage = Human", and then crashes
right away. Correct me if I am wrong.
I notice a couple things on your graphic configuration, but I am not sure if
they are related to the issue. (8MB memory is much smaller than my
configuration, and GM965 GPU seems to support only OpenGL 1.5 while I wrote my
program based on a book about OpenGL 3.x) I will check it later.
Original comment by tom...@gmail.com
on 14 Jan 2011 at 1:32
Yes, it is just as you said, it crashes at the point it displays "Hensin Stage
= Human".
I tested the application in another configuration: Win7 Desktop computer, core
2 quad, Nvidia and it worked, but some poses like "emerium beam" don't work,
i'll be posting more information in the next hours in order to give you more
information about working configurations.
Thanks,
Ismael
Original comment by Sharkar...@gmail.com
on 14 Jan 2011 at 4:33
Ismael,
I reproduced a crash with my laptop which is supposed to be the one you and
tkabata encountered.
Also thanks in advance to find and report another issue.
Original comment by tom...@gmail.com
on 14 Jan 2011 at 4:55
I ran the program with Task Manager open, and I observed the memory usage was
skyrocketing and then it crashed eventually. It is weird in a sense that it
only happens on the low-end laptop. There is nothing obviously wrong in my code
either.
Original comment by tom...@gmail.com
on 14 Jan 2011 at 6:35
0.1aの状況
※MMD+Kinectは正しく動作していますのでOpenNI系の問題はない�
��思います。
1.kinect-ultra.exeを起動
2.コマンドコンソールが立ち上がる
3.いくつかメッセージが表示され、その瞬間何かのウィ��
�ドウが一瞬表示され
すぐに消える。
4.コンソールに
Hensin Stage = Human
が表示された数秒後にOSのSTOPダイアログが表示される。
0.1bでテストしたらまたレポートします。
Original comment by tkab...@gmail.com
on 14 Jan 2011 at 8:45
@en Everyone, I nailed down the location of the memory leak that eventually led
the program to a crash, and worked around it. It is weird this memory leak may
happen or not from configuration to configuration, but I hope this to work out
on your configurations. Please download 0.1c and test it.
@ja
最終的にクラッシュに至るメモリリークがあったので、直��
�ました。構成によって起こる場合と起こらない場合があっ�
��奇妙なのですが、そちらの環境でこの修正が効くことを祈
ります。0.1cをダウンロードしてテストしてみてください。
Cheers!
Original comment by tom...@gmail.com
on 14 Jan 2011 at 9:59
0.1cでテストしましたがやはりだめです。
画面のスナップショットとイベントログをお送りします。
障害が発生しているアプリケーション名:
kinect-ultra.exe、バージョン: 0.0.0.0、タイム スタンプ: 0x4d3018d0
障害が発生しているモジュール名: unknown、バージョン:
0.0.0.0、タイム スタンプ: 0x00000000
例外コード: 0xc0000005
障害オフセット: 0x00000000
障害が発生しているプロセス ID: 0x1640
障害が発生しているアプリケーションの開始時刻:
0x01cbb62e889b2c6c
障害が発生しているアプリケーション パス:
C:\Users\kabayan\Desktop\kinect-ultra-0.1c\kinect-ultra.exe
障害が発生しているモジュール パス: unknown
レポート ID: c95f4bbc-2221-11e0-b145-eca42b108a98
- System
- Provider
[ Name] Application Error
- EventID 1000
[ Qualifiers] 0
Level 2
Task 100
Keywords 0x80000000000000
- TimeCreated
[ SystemTime] 2011-01-17T10:08:53.000000000Z
EventRecordID 18689
Channel Application
Computer kabayan-PC
Security
- EventData
kinect-ultra.exe
0.0.0.0
4d3018d0
unknown
0.0.0.0
00000000
c0000005
00000000
1640
01cbb62e889b2c6c
C:\Users\kabayan\Desktop\kinect-ultra-0.1c\kinect-ultra.exe
unknown
c95f4bbc-2221-11e0-b145-eca42b108a98
Original comment by tkab...@gmail.com
on 17 Jan 2011 at 10:15
Attachments:
tkabataさん、
0.1cだめでしたか、残念。0.1cで解決できた症状は、dothebartman
報告の「ちゃんと(コンソールでない方の)画面は表示されて
、しばらくの間正常に動くが、しばらくたつとメモリを食��
�つぶしてクラッシュしてしまう」というものでした。一方�
��tkabataさんの所の症状は、「画面がちゃんと表示される前��
�クラッシュする」「タスクマネージャで見てメモリがうな�
��のぼりということもない」というまったく別の症状だと思
ってよいでしょうか?
問題が起こる環境が手元にないと問題の特定は難しいので��
�気長に待ってもらう必要があるかもしれません。Henshin
Stageが出た後且つ画面が出る前だとするとクラッシュ箇所は
ある程度特定できるのですが、どのグラフィックス命令が��
�き金になっているかの推測が少々やっかいです。OpenGLのバ�
��ジョンが同じで、しかもGeForce Go
7400より全然低スペックと思われる私のノートPCでは動いて��
�ますので…。そちらに開発環境があればデバッガで追って�
��らうのも大歓迎です。
Original comment by tom...@gmail.com
on 17 Jan 2011 at 12:09
tkabataさん、
クラッシュしてる疑いのある初期化処理のあたりにprintを埋
めまくった版を作りました(添付)。
これをkinect-ultra.exeの代わりに置いて実行し、コンソールに�
��が出たか(最後に出たのは何か)教えてもらえませんか。こ�
��でどのステップでこけているのかさらに絞り込めると思う
のですが。
Original comment by tom...@gmail.com
on 18 Jan 2011 at 9:09
Attachments:
Everyone,
Based on the information Ismael gave me via e-mail, I created another
experimental build as attached kinect-ultra-issue1-with-mipmap-fix.exe. Can you
volunteer to try it and let me know how it worked? (If it did not work, what
was in the console?)
The information from Ismael indicated that the program had crashed in the
constructor of SparkRenderer. After a quick investigation, I noticed that the
process included an OpenGL 3.x API call which might cause a crash on OpenGL
2.x. In this experimental debug build, I made it compatible with OpenGL 2.x and
also added verbose logs around.
Appreciate your help,
Tomoto
Original comment by tom...@gmail.com
on 23 Jan 2011 at 2:40
Attachments:
kinect-ultra-issue1-with-log.exe と kinect-ultra-issue1-with-mipmap-fix.exe
でテストしました。いずれも残念ながらクラッシュします��
�
スクリーンショットとメッセージを添付しておきます。
C:\Users\kabayan\Desktop\うるとらせぶん>kinect-ultra-issue1-with-mipmap-f
ix.exe
kinect-ultra 0.1d
Kinectでウルトラセブン!へようこそ!
実行中には次のキーが使えます。
[ESC] -- 終了
[q][a] -- 3D仮想オブジェクトの奥行を調節する。
アイスラッガー等の奥行がおかしいときに調節してください。
[s] -- 飛んだ時のハメ込み用画像を撮影しなおす。
ハメ込み用画像は起動時に撮影しますが、そのときに人などが
写りこんでしまった場合に使ってください。
初期化が終わるまでしばらくお待ちください…Enjoy!
OpenGL version = 2.1.0
Shader language version = 1.20 NVIDIA via Cg compiler
Henshin Stage = HUMAN
LOG: initProjection()
LOG: g_voxelObjectMapper = new VoxelObjectMapper(&g_depthGen)
LOG: g_sparkRenderer = new SparkRenderer(&g_renderingCtx)
LOG: glTexParameteri(GL_TEXTURE_2D, GL_GENERATE_MIPMAP, GL_TRUE)
LOG: g_flatImageRenderer = new ImageRenderer(&g_renderingCtx, &g_imageGen)
C:\Users\kabayan\Desktop\うるとらせぶん>
よろしくお願いします。
Original comment by tkab...@gmail.com
on 26 Jan 2011 at 3:50
Attachments:
tkabataさん、報告どうもありがとうございます。Ismaelからも�
��ったく同じ報告をもらいました。少しずつ先には進んでお
り、突っかかるところはテクスチャ初期化だろうと目星は��
�いているので、あと少しで解決できると踏んでいます。(た
だしGPUのテクスチャ用メモリが足りないみたいな話だった��
�合には結局動かないという可能性もなきにしもあらずです�
��…そうでないことを祈ります。)
Original comment by tom...@gmail.com
on 26 Jan 2011 at 4:04
Hi All, I took a look at the source code around the crash point (ImageRenderer
initialization), but did not see any obvious issues so far. It may take some
time. Thanks!
Original comment by tom...@gmail.com
on 27 Jan 2011 at 9:22
[deleted comment]
[deleted comment]
Based on the information from nuwand...@hotmail.com posted as Issue 8, I am
almost certain that this happened because your platform (GPU and graphics
driver) did not support a feature required by this program (called VAOs). I
modified the program to check for this feature at start up and exit with a
proper error message (instead of crash) if it is not supported. This
modification will be included in the next release package.
I am sorry that I could not make this program work on your configuration
because it turned out an insufficiency of the platform feature.
Original comment by tom...@gmail.com
on 5 Mar 2011 at 9:04
Issue 8 has been merged into this issue.
Original comment by tom...@gmail.com
on 5 Mar 2011 at 9:10
Issue 10 has been merged into this issue.
Original comment by tom...@gmail.com
on 5 Mar 2011 at 9:19
[quote]
Comment 1 by project member tom...@gmail.com, Today (51 minutes ago)
I merge this issue into Issue #1. Possibly your GPU does not support a required
feature called VAOs.
[/quote]
What can we do about this ? other than replacing the GPU...
Original comment by ilkerakt...@gmail.com
on 5 Mar 2011 at 10:13
Well...
(1) Replace the GPU
(2) Wait for the manufacture to release a new driver that supports the required
feature
(3) Wait for someone to volunteer to migrate my program onto the old OpenGL API
Your screenshot says your OpenGL version is 1.5, which is way too old. I wonder
why your "latest" driver only supports this old version although Mobility
Radeon 9700 should support 2.0 according to ATI's specification chart. I do not
think this program could work on 1.5 anyway even if we avoided the usage of
VAOs.
Original comment by tom...@gmail.com
on 6 Mar 2011 at 6:27
Ok; I've found another version (newer) of GPU driver.
Now it says:
Open GL version: 2.0.5819 WinXP Release
Shader language version: 1.10
Henshin Stage =HUMAN
and then crashes.
So, now my OpenGL version meets the requirements. Correct ?
Why does it crash ?
usage of VAOs ?
what can I do ?
Thanks.
Original comment by ilkerakt...@gmail.com
on 6 Mar 2011 at 9:43
Most probably, your GPU and the driver still do not support VAO feature which
is required by this program. I am sorry, but your options are still the same.
"Star" this issue and wait, then you will be notified via an e-mail when there
are any developments.
Thanks for your interest.
Original comment by tom...@gmail.com
on 6 Mar 2011 at 7:48
What is this VAO and why do we need it for your app ?
I'm asking because I use this notebook for many purposes (including games) and
I've never had a problem with my display adapter until your application.
Original comment by ilkerakt...@gmail.com
on 6 Mar 2011 at 8:06
This program is utilizing some of the cutting-edge features such as VAOs
supported only by relatively new GPUs for a better performance and ease of
programming. Commercial products (including games) are usually designed to work
without these features (for extra development and performance costs) in order
to get as many customers as possible. Meanwhile, this program is a
non-commercial exploration into the new technologies, so we do not hesitate to
use these new features even if the "market" becomes smaller. Also, since this
project is Open Source, anyone can contribute to modify the program for older
GPUs if he or she wants to make it happen.
Original comment by tom...@gmail.com
on 6 Mar 2011 at 8:24
understood.
Then we (who have older GPUs) are unlucky.
Original comment by ilkerakt...@gmail.com
on 6 Mar 2011 at 8:57
I am sorry about that. Just star this issue and wait for good news you may
receive.
Original comment by tom...@gmail.com
on 6 Mar 2011 at 9:29
Original issue reported on code.google.com by
tkab...@gmail.com
on 10 Jan 2011 at 1:00