TimeDX is designed to allow you
to test all components of DMDX before using it in order to make sure that
the whole lot works individually and in combination if you don't want to use
DMDX in it's Auto mode. This is doubly true
for the various video modes that DMDX could use, their refresh interval must
be examined (unless DMDX's Auto mode is used) and optionally tuned by you in order that the retrace can be
correctly anticipated -- for each different video mode used.
When using DMDX or TimeDX it is best to kill off all other
tasks not essential to the running of the machine, they are only going to
steal CPU cycles from the tasks that matter. One thing I've seen a
couple of times with Vista is that the first time you run a test that uses a
new video mode the results are just all over the place, run that test again and
they're fine. Suspect what's happening is similar to what started
happening back with XP where the first time a new video mode is used DirectX is
busily writing new registry keys but where XP was fairly quick doing this Vista
The steps in using TimeDX are as follows:
- Step 0.
Close all other programs before
running TimeDX or DMDX. A long time ago I discovered that a major source
of disruption in the Refresh Rate test was due to MS Word being open as I was updating
this documentation, I was ready to give up and go panhandling for a living
until I nailed down just what the culprit was.
- Step 1.
The first time you use TimeDX
you should check that the Millisecond Timer is functioning. You should also do this if you managed
to crash TimeDX or DMDX, it appears that the system can take a while to repair
the millisecond callback (you might have to restart TimeDX if it isn't functioning).
- Step 2.
Check that the Video Mode you want to use works. If you are using a TFT display (or LCD flat
panel) then you will want to make sure you are using the display's native
resolution. See the Video Mode
- Step 3.
Initially you should use the
Rate test at least once to ascertain
how well your machine can automatically determine the retrace interval, it
doesn't need to be done for every video mode you select (although it doesn't
hurt). If you are using an LCD flat panel display this test needs
to be paid extra attention as it can be used to ascertain the flat panel's
native capabilities which is absolutely critical for any tachistoscopic
- Step 4.
Next you are going to jump ahead
to the Advanced tests.
Again, at least once you should check that the millisecond callback latencies
aren't disturbed while there is a bunch of Video activity, you certainly don't have to do this for every video
- Step 5.
This one's a big test, another
Advanced test, you
are going to fire up the Time Video Mode Test and determine the values that will let DMDX synchronize
itself with the vertical retrace.
For each video mode you might use you will have to perform Step 2 and Step 5
(unless of course you're using DMDX's Auto mode). The values
you determine will be written to the registry when you press the Save Last Used
values in Registry button where other routines can retrieve
them when they wish to synchronize with the vertical retrace -- which I figure
is eons ahead of having to always wait several seconds as video modes are
changed for the automatic routines to determine the characteristics of the
new video mode. The down side is that if you ever change video cards you
will have to re-time all the video modes.
For now ignore the Sound Buffers, one thing at a time.
The important thing to remember here is that the results must be saved to the registry and that
until you do this DMDX will not run, instead it will be complaining about
registry keys missing.
- Step 6.
You're done if you're not
presenting audio stimuli or using non standard input devices (this can
include international keyboards though) or a parallel IO device.
- Step 7.
Now you can
play with the Sound stuff. Pick some .WAV files and play them. During
this and the next few sound related test if you notice less than stellar
performance from your sound card you may want to choose a different Direct
Sound driver, some drivers are
less than excellent (and, no, I don't know why more than driver one can appear).
- Step 8.
Next you want
to go back to the Advanced test and
tell it to play a few buffers as well, see if the parameters need more tuning,
it is normal for the error rate to got up a bit. The Tachistoscopic Acid Test
has been dropped from TimeDX 5.
- Step 9.
Use the Sound Latency test even if you don't have a oscilloscope as this
test really cuts the duds out. Of the three sound cards I had decided were
OK two failed (one miserably) just playing the audio portion of the test,
I didn't need an oscilloscope.
- Step 10.
optionally use the Input test to see how the various input devices you might
want to use perform. The names of both input devices and the buttons on
them will probably be needed for DMDX as the
item file will more than likely wind up specifying which input device to use and which buttons are the correct, incorrect and request keys.
- Step 11.
If you are
going to using a PIO then you will want to set up it's address with in the
PIO Test test, and
if you are going to be using another machine to monitor DMDX then you will want to set up the network address
with the Network Monitor test.