3 ms·

I'm very glad to see arstechnica using the blurbusters tool for testing monitor rate! It was initially a Win32 exe that I created 5 years ago [1] with a simple

I'm very glad to see arstechnica using the blurbusters tool for testing monitor rate! It was initially a Win32 exe that I created 5 years ago [1] with a simple idea to display very distinct frames, back when people were using all kinds of silly less accurate methods like timers to do these things. Blurbusters did a great job of making the tool more accessible to more people (visit website vs. download and run an unknown exe). They were even respectful enough to ask for my permission before adding it their website (which I of course was happy to give permission to, and not that I would mind too much if they didn't, it's a very simple idea after all). [1] http://hardforum.com/showthread.php?t=1423433 http://hardforum.com/showthread.php?t=1423433

tofof 12y ago
Except that they didn't do so! They just took a screenshot, the contents of which include the instructions that a screenshot is meaningless.
- shurcooL 12y ago
  Hah, I just noticed that those 2 really nice photos at the bottom are a comment from blurbusters, not a part of the article. One step at a time, hehe.
  - nacs 12y ago
    Well its kind of part of the article -- it's a promoted comment (and the only one).
pinumbernumber 12y ago
"UNSUPPORTED: VSYNC is not available on the Linux platform". What feature exactly is supported on Windows and OSX but not Linux that provides correct vsync?
- jrockway 12y ago
  Linux supports a variety of compositing managers (programs that receive pixels from the application and are responsible for actually making the video hardware draw them), and these each have various settings that control how they interact with blanking. I have mine set to sync to vblank and so Chrome animations seem to sync to vblank too. Changing the settings on the fly, though, is implementation dependent, and whatever implements the feature for this website probably doesn't want to bother. It does work fine on ChromeOS, so it's a "generic Linux" thing, not some specific problem with the Linux kernel itself.
jmpeax 12y ago
I don't quite understand how it works. My guess is that it doesn't actually perform any test itself, you're the one that is deciding whether the refresh rate is actually 60Hz. The big green "VALID" doesn't say you're monitor is at 60Hz, but that it is drawing at 60Hz to the frame buffer, and you can now observe the motions to decide for yourself whether it is smooth enough that the refresh is running at 60Hz?
- shurcooL 12y ago
  That's exactly right. See the comment at the bottom of the article with two photos, it describes it in detail (confirming what you said).
- ojbyrne 12y ago
  The comment at the bottom adds the key piece of information - you need a camera that you can set the shutter speed to 1/60th of a second.
stinos 12y ago
I'm very glad to see arstechnica using the blurbusters tool for testing monitor rate! You should be, it's a good trick and at least uses external hardware which makes it more trustworthy than trying to measure what the system is doing from within the system. (e.g. 'it can be confirmed with apps that measure your refresh rate') Usually that's fine, but still a 60Hz refresh doesn't mean that whatever program is rendering onto the screen is capable of getting 60 frames per second effectively being put out by the screen without dropping any frames. Which is ok for most everyday usage, but e.g. for neurosciences doing research of the visual system that is hardly sufficient. You really want to be 100% sure that whatever you tell a program to display, it does this correctly frame by frame. And the only way to make sure that happened is to use external, continuously recording, hardware (photodiode or high speed cameras).

tofof 12y ago

Except that they didn't do so! They just took a screenshot, the contents of which include the instructions that a screenshot is meaningless.

shurcooL 12y ago
Hah, I just noticed that those 2 really nice photos at the bottom are a comment from blurbusters, not a part of the article. One step at a time, hehe.
- nacs 12y ago
  Well its kind of part of the article -- it's a promoted comment (and the only one).

pinumbernumber 12y ago

"UNSUPPORTED: VSYNC is not available on the Linux platform". What feature exactly is supported on Windows and OSX but not Linux that provides correct vsync?

jrockway 12y ago
Linux supports a variety of compositing managers (programs that receive pixels from the application and are responsible for actually making the video hardware draw them), and these each have various settings that control how they interact with blanking. I have mine set to sync to vblank and so Chrome animations seem to sync to vblank too. Changing the settings on the fly, though, is implementation dependent, and whatever implements the feature for this website probably doesn't want to bother. It does work fine on ChromeOS, so it's a "generic Linux" thing, not some specific problem with the Linux kernel itself.

jmpeax 12y ago

I don't quite understand how it works. My guess is that it doesn't actually perform any test itself, you're the one that is deciding whether the refresh rate is actually 60Hz. The big green "VALID" doesn't say you're monitor is at 60Hz, but that it is drawing at 60Hz to the frame buffer, and you can now observe the motions to decide for yourself whether it is smooth enough that the refresh is running at 60Hz?

shurcooL 12y ago
That's exactly right. See the comment at the bottom of the article with two photos, it describes it in detail (confirming what you said).
ojbyrne 12y ago
The comment at the bottom adds the key piece of information - you need a camera that you can set the shutter speed to 1/60th of a second.

stinos 12y ago

I'm very glad to see arstechnica using the blurbusters tool for testing monitor rate! You should be, it's a good trick and at least uses external hardware which makes it more trustworthy than trying to measure what the system is doing from within the system. (e.g. 'it can be confirmed with apps that measure your refresh rate') Usually that's fine, but still a 60Hz refresh doesn't mean that whatever program is rendering onto the screen is capable of getting 60 frames per second effectively being put out by the screen without dropping any frames. Which is ok for most everyday usage, but e.g. for neurosciences doing research of the visual system that is hardly sufficient. You really want to be 100% sure that whatever you tell a program to display, it does this correctly frame by frame. And the only way to make sure that happened is to use external, continuously recording, hardware (photodiode or high speed cameras).