Vista and Windows 7.
Leif Asbrink
After having found out that Windows is useless with PCI soundcards
on modern motherboards I made some tests on a not so modern computer.
A Pentium III at 650 MHz. As it turns out, Windows Vista and Windows 7
have a very CPU-hungry process, audiodg, that uses up nearly all
the CPU time on this computer. The tests were without second FFT
and without any output (no signal clicked.)
http://www.sm5bsz.com/linuxdsp/run/snd-compat/d44-cpuload.htm
(referenced from)
http://www.sm5bsz.com/linuxdsp/run/snd-compat/snd-compat.htm
The conclusion is that Vista and Windows 7 are useless for Linrad on a
Pentium III.
The sad news is that even a Pentium IV is not fast enough for use
of Linrad with the WSE converters under Vista or Windows 7. Occasional
loss of input data causes a time skew between the two RF channels that
would degrade noise blanking and make the polarisation graph non-valid.
I do not think Linrad will work well with MAP65 under Vista/7 on any
computer because occasional loss of data would bring incorrect data
into the decoder. It seems to me that Vista has bad real-time performance.
I have tested Winrad and it seems the ASIO drivers send data directly
from the hardware to Winrad bypassing the audiodg processes. Finding
out how to incorporate the ASIO API in the mingw package and to use
it properly in Linrad seems to be a too major task. The single channel
ASIO driver in Winrad does not work properly under Vista while the WMME
driver works perfectly fine on my Pentium IV (but with a high CPU
load.) I do not know whether the ASIO problem is Vista itself or something
in the Winrad code that is incorrect for Vista. (I installed Winrad
by running the installer in XP compatible mode.)
Under Windows XP both the WMME and the ASIO input routines of
Winrad run without errors at a low CPU load on the Pentium IV.
It even runs perfectly well on the Pentium III.
My conclusion is the following: Do not use Linrad with soundcards
on computers running Windows Vista or Windows 7. Errors are easily
seen if a single very strong signal is present at the input.
Errors also manifest themselves as a somewhat too low sampling speed.
The good news is that Perseus, SDR-14 and SDR-IQ do not go through
the Vista sound system so they should work fine. They also do not
use the PCI bus so they run well on modern computers. I do not
yet know Whether Vista creates problems with the digital receivers
on slow computers, but I will have a look at it. I have a feeling
that the Vista latency might be too long to be compatible with the
USB interrupt routines on a 650 MHz processor.
73
Leif / SM5BSZ
rad...@attglobal.net
I am always impressed at the amount of Windows informattion you
uncover and share with us, all this from a guy who says he knows
notheing about windows.. Thank you.
What is the path of least resistance for those of us who use Windows? It
sounds like Windows XP on a Modern Computer. Are all Windows XPs
acceptable?
73,
John K1VBM
---- Original Message ----
Leif Asbrink
> What is the path of least resistance for those of us who use Windows? It
> sounds like Windows XP on a Modern Computer.
That depends on the hardware.
If you have a direct conversion radio with two RF channels connected
to a PCI soundcard there is nothing from Microsoft that works on a
modern computer. You would have to find an old computer and you would
have to run XP on it. Or perhaps Windows 2000. I actually think 2000
is better if you want to run many other programs at the same time but
I have not made any serious comparison.
If you have a direct conversion radio with a single RF channel and use
the builtin audio or perhaps a PCI Express card you can use any
version of Windows (32 as well as 64 bit) on a modern computer. (But
I have only tested a single one, the D5400XS.)
If you have just a single CPU core I think you should use XP or 2000
even if the processor is fast.
I do not yet know enough about how Perseus and SDR-14 work on different
machines. Modern Windows may add extra delay due to some needless
audio processing on the output (equalizers, echo compensation,
channel delay...) but I would not think there would be any problems on
the input side.
> Are all Windows XPs acceptable?
I have a limited budget for buying Microsoft products. The only XP
I tested is XP Professional 32 as well as 64 bit. The 64 bit version
does not work with any of my PCI cards. (I have no 64 bit system that
does not have a PCI Express slot.) But Windows 2000 is fine so I see no
reason why any 32 bit XP should have a problem.
ALL MODERN LINUX DISTRIBUTIONS. 32 BIT AS WELL AS 64 BIT WORK FINE WITH
LINRAD ON ALL COMPUTERS TOGETHER WITH ALL SOUNDCARDS.
73
Leif / SM5BSZ
Dominique DEHAYS
ALL MODERN LINUX DISTRIBUTIONS. 32 BIT AS WELL AS 64 BIT WORK FINE WITH
LINRAD ON ALL COMPUTERS TOGETHER WITH ALL SOUNDCARDS.
AdB
> The single channel
> ASIO driver in Winrad does not work properly under Vista while the WMME
> driver works perfectly fine on my Pentium IV (but with a high CPU
> load.) I do not know whether the ASIO problem is Vista itself or something
> in the Winrad code that is incorrect for Vista. (I installed Winrad
> by running the installer in XP compatible mode.)
>
could you please elaborate a little on this ? As I told you on another
group, I am about to
bring to life a quad core Intel CPU, on a mixed PCI - PCIExpress
motherboard, and one
of the op-sys I will install on it, with a multi boot manager, is Windows 7.
So I would like to know which problems you have encountered with Winrad
and Windows 7.
I did a short test a few weeks ago, when I installed the Release
Candidate of Windows 7
on an old Asus laptop, with a Pentium M (Mobile) clocked at 1.4 GHz, and
Winrad ran
without error, without using the XP compatibility mode. But I did not
test the ASIO drivers,
as they do not exist for the chip set of that old laptop.
I will report my findings on the new PC, when completed, installed and
running.
73 Alberto I2PHD
Stan, W1LE
I appreciate the update.
You have provided important answers that have been haunting me for months.
> My conclusion is the following: Do not use Linrad with soundcards
> on computers running Windows Vista or Windows 7. Errors are easily
> seen if a single very strong signal is present at the input.
> Errors also manifest themselves as a somewhat too low sampling speed.
>
>
So, Windows will not work properly with PCI or PCIe soundcards or the
mother board sound.
How about the firewire FA-66 (4 channel) and the USB EMU 0202 ?
Looks like I need to dust off my Linux hat for the WSE and Delta-44
combination.
Thanks Stan, W1LE Cape Cod FN41sr
stan...@aol.com
This might be a related issue? On 1 of my Linrad systems I am having audio drop out and DA sync error. The system is Perseus and Winows XP service pack 3. The computer is an Intel Quad core. The sound card is Creative Sound Blaster PCI.
Thanks Stan
Leif Asbrink
> could you please elaborate a little on this ? As I told you on another
> group, I am about to
> bring to life a quad core Intel CPU, on a mixed PCI - PCIExpress
> motherboard, and one
> of the op-sys I will install on it, with a multi boot manager, is Windows 7.
>
> So I would like to know which problems you have encountered with Winrad
> and Windows 7.
The problem is with Windows Vista. Presumably it is with Windows 7 as well,
I did not try.
The test is with an audio generator into one channel. I see two carriers
because there is not I and Q with a 90 degree shift. With the WMME
drivers all is OK on the Pentium 4, but with the ASIO driver I see
transcients that indicate disconitnuities in the input stream.
> I did a short test a few weeks ago, when I installed the Release
> Candidate of Windows 7
> on an old Asus laptop, with a Pentium M (Mobile) clocked at 1.4 GHz, and
> Winrad ran
> without error, without using the XP compatibility mode. But I did not
> test the ASIO drivers,
> as they do not exist for the chip set of that old laptop.
>
> I will report my findings on the new PC, when completed, installed and
> running.
I am awaiting your report with great interest:-)
73
Leif
Leif Asbrink
> So, Windows will not work properly with PCI or PCIe soundcards or the
> mother board sound.
What I have found is that motherboards that have both PCI and PCIe
slots do not work properly with PCI cards with any version of Windows.
I do not know whether PCIe cards work, but the motherboard device
works well at 96 kHz on the D5400XS (which does not work with
the Delta44.)
> How about the firewire FA-66 (4 channel) and the USB EMU 0202 ?
I would guess that the problem is with PCI only, but I do not know.
> Looks like I need to dust off my Linux hat for the WSE and Delta-44
> combination.
Or you could use XP on an older PC without PCIe slots:-)
73
Leif / SM5BSZ
Leif Asbrink
> This might be a related issue? On 1 of my Linrad systems I am
> having audio drop out and DA sync error. The system is Perseus
> and Winows XP service pack 3. The computer is an Intel Quad core.
> The sound card is Creative Sound Blaster PCI.
Probably not. With the Perseus you can easily run out of CPU time,
particularly when sampling at 2 MHz.
Use 'T' and look at the timing for the different threads. Use
Linrad-03.07 which has more reliable thread timing than previous
versions. If some thread runs near 100% it may not deliver data
in time and that would cause drop outs and DA sync errors.
Some day I will add error messages to warn about bad parameter
settings when thread times become too large........
73
Leif / SM5BSZ
rad...@attglobal.net
Many thanks,
John
Subject: [Linrad] Re: Vista and Windows 7.
WITH
LINRAD ON ALL COMPUTERS TOGETHER WITH ALL SOUNDCARDS.
Leif / SM5BSZ