[[email protected]]

Greetings Earthlings,
currently I'm using one M-Audio Delta
1010LT card with Cubase 3 in a homebrew PC which, briefly, is along
these lines Sempron 2GHz/1Gb RAM/240Gb HDD with only one other card:
AGP video. At Fs 96kHz there is an intermittant "halo" of feint
fizzing sound, like breaking up crossed with mild distortion sort of
sound at a low level when the desired signal is at nominal to peak
level. This disappears when Fs is reduced to 88k2Hz and running at
this rate is problem free day in day out.

I plan to install an additional Delta 1010LT and wordclock lock the
two cards and am of the opinion that as the second card will occupy a
second PCI slot that it will operate fine at Fs 88k2Hz as the card
will be in it's own seperate PCI time-frame. The motherboard is an
Asus, I forget the model number but it was current technology when the
AMD Sempron was purchased and is about three years old.

Personally I have no problem with Fs 44k1Hz but I run at 88k2Hz so
that clients work is archivable and as future-proof as the next
guys ;-)

Experienced PCI card hacks: your comments and advice please. The 96kHz
problem niggles me :-(

TIA Lightclock

Please excuse any smelling pistakes.

[[email protected]]

Maybe a bad crystal for that frequency, causing audible jitter.

Do you mean an actual faulty xtal or do you think that the problem
could be a design flaw and that the "fundamental" frequency (which
will actually be an overtone frequency) and the divider for the sample
clock are causing some sort of interference within the card cct?

I think that you probably mean a faulty xtal, which being run at a
harmonic, could be the culprit. Have you heard of this before?

It would seem that at the worst case I should buy the second
card, try that independantly at Fs 96kHz, and listen carefully.
If that shows the same problem then I will have to swallow my
pride and run at 88k2Hz. This will be just my luck probably
and you 96kHz and 192kHz guys won't mention it again ;-)

Bob Morein
(310) 237-6511

Cheers n beers

Lightclock

[Norbert Hahn[_2_]]

wrote:

Greetings Earthlings,
currently I'm using one M-Audio Delta
1010LT card with Cubase 3 in a homebrew PC which, briefly, is along
these lines Sempron 2GHz/1Gb RAM/240Gb HDD with only one other card:
AGP video. At Fs 96kHz there is an intermittant "halo" of feint
fizzing sound, like breaking up crossed with mild distortion sort of
sound at a low level when the desired signal is at nominal to peak
level.

Does the noise depend on the level of the input signal?
If you have all sliders down on your mixer and you record -
what do you get?
Does it depend on the frequency of the input signal?
Does the noise depend on activity of the computer such as disk
access, moving of the mouse etc.?

Have you looked a the spectrum of the unwanted signal?

This disappears when Fs is reduced to 88k2Hz and running at
this rate is problem free day in day out.

Does "disappear" meen that you simply do not hear it or do you
get a clean recording when you look at the spectrum of the
recording?
Could the noise just get shifted into ultrasonic and thusly
become inaudible?

I would clarify the reason for the problem before buying one more
card. Maybe moving the card to a different PCI slot can cure the
problem - or make it worse.

Norbert

[David Morgan \(MAMS\)]

wrote in message...

Experienced PCI card hacks: your comments and advice please.


Don't waste time and drive space recording at 96K.... problem solved.



--
David Morgan (MAMS)
Morgan Audio Media Service
http://www.m-a-m-s DOT com
Dallas, Texas (214) 662-9901
_____________________________
http://www.januarysound.com

[Steve[_3_]]

wrote in message
...
Greetings Earthlings,
currently I'm using one M-Audio Delta
1010LT card with Cubase 3 in a homebrew PC which, briefly, is along
these lines Sempron 2GHz/1Gb RAM/240Gb HDD with only one other card:
AGP video. At Fs 96kHz there is an intermittant "halo" of feint
fizzing sound, like breaking up crossed with mild distortion sort of
sound at a low level when the desired signal is at nominal to peak
level. This disappears when Fs is reduced to 88k2Hz and running at
this rate is problem free day in day out.

I plan to install an additional Delta 1010LT and wordclock lock the
two cards and am of the opinion that as the second card will occupy a
second PCI slot that it will operate fine at Fs 88k2Hz as the card
will be in it's own seperate PCI time-frame. The motherboard is an
Asus, I forget the model number but it was current technology when the
AMD Sempron was purchased and is about three years old.

Personally I have no problem with Fs 44k1Hz but I run at 88k2Hz so
that clients work is archivable and as future-proof as the next
guys ;-)

Experienced PCI card hacks: your comments and advice please. The 96kHz
problem niggles me :-(

I have had lots of experience with PCI bus problems before I upgraded to a
PCIe system (Was running an ASUS P5P800SE board w/agp). Here's what I found
over the last couple of years (two years with an M-Audio Delta 1010-Non LT,
1 year with a FW1814, 2 years with a Tascam FW-1804):

1. Your motherboards BIOS should have a setting called "Video Adapter
Priority". Set this to PCI. This will yield priority to your PCI bus.
2. Turn off your graphics acceleration on your video card, disable write
combining.
3. Your on board network card is also PCI. The P5P800 had a gigabit
adapter on the PCI bus. I found that disabling this while doing anything
audio was a big help.
4. Disconnect any high speed USB devices. Your USB controllers are also on
the PCI bus.
5. You may be multipling your problems by adding another card. I would get
your PCI bus problems straightened out before considering adding another
card.

I don't know if your problem is stemming from the PCI bus, but audio
dropouts (heard as split second pops and clicks on raw recordings) will
sound very different (like what you describe) when effects are stacked on
top of them. You may want to further troubleshoot this problem by recording
your 1010LT's sound inputs with nothing connected (ie. recording "silence")
for about 30 minutes and see if there are any pops and/or clicks in your
wave files. These artifacts can be much easier to see when there is nothing
being recorded.

Cheers;
Steve


TIA Lightclock

Please excuse any smelling pistakes.

[Arny Krueger]

wrote in message

Greetings Earthlings,
currently I'm using one
M-Audio Delta 1010LT card with Cubase 3 in a homebrew PC
which, briefly, is along these lines Sempron 2GHz/1Gb
RAM/240Gb HDD with only one other card: AGP video.

Could be bus contention from the AGP card. Running at the higher sample rate
provides enough data rate to break some poor camel's back. The usual
circumvention is to back off on the hardware assists using one of the menus
in display properties.

At Fs
96kHz there is an intermittant "halo" of feint fizzing
sound, like breaking up crossed with mild distortion sort
of sound at a low level when the desired signal is at
nominal to peak level.


Usual question. If you loop the output of the 1010LT to its input, and run
the freeware audio rightmark analysis program, what sort of results do you
get? If anything bad is happening in the card, the rightmark program should
show some kind of noise or distortion artifact in one of its tests.

Of course, who gives a crap about how *any* audio interface runs at 96 KHz
sampling, unless perhaps you are using it for test equipment? There's a ton
of scientific evidence showing when it comes to recording music, if you
can't get it done with 44 or 48 KHz sampling, you can't get it done.


This disappears when Fs is reduced
to 88k2Hz and running at this rate is problem free day in
day out.

Probably a sample defect. Do the right thing - save yourself a ton of grief
and use 44 kHz (target distribution = audio) or 48 kHz (target distribution
= video) sampling.

I plan to install an additional Delta 1010LT and
wordclock lock the two cards and am of the opinion that
as the second card will occupy a second PCI slot that it
will operate fine at Fs 88k2Hz as the card will be in
it's own seperate PCI time-frame.

If you're running multiple audio interfaces in a PC, and their clocks aren't
locked - well that is pretty disasterous.

The 1010LT instructions provide explicit instructions on how to use more
than one at the same time. You might want to follow them. ;-)

The motherboard is an
Asus, I forget the model number but it was current
technology when the AMD Sempron was purchased and is
about three years old.

I've been running a 1010 and two 1010LTs on an old ECS K7S5A with a 2000+
processor for years. I generally record 12-24 tracks for 35 to 90 minutes at
a time in this application.

Personally I have no problem with Fs 44k1Hz but I run at
88k2Hz so that clients work is archivable and as
future-proof as the next guys ;-)

Fear of the dark, methinks.

Experienced PCI card hacks: your comments and advice
please. The 96kHz problem niggles me :-(

1010LTs are dirt cheap. Try another.

[Arny Krueger]

"Soundhaspriority" wrote in message

wrote in message
...

Maybe a bad crystal for that frequency, causing audible
jitter.

Terribly unlikely. The usual sources of clock jitter are noise or power
supply artifacts creeping into the clock circuits. Usually, this is guarded
against with considerable passion, but a bad electrolytic in the wrong place
could make it happen.

Do you mean an actual faulty xtal or do you think that
the problem could be a design flaw and that the
"fundamental" frequency (which will actually be an
overtone frequency) and the divider for the sample clock
are causing some sort of interference within the card
cct?

I've definately seen the power supply bug cause jitter. It shows up like a
sore thumb in a Rightmark analysis of the card. There will be visible
sidebands around the test signal in any of the FFT analysis. They usually
show as a thickening with normal frequency scaling, but the Rightmark
(freeware) program provides a scale magnification facility that shows the
whole mess quite clearly. The sidebands will be at +/- 60 Hz from the
fundamental. The displacement may be multiplied if you're looking at a
harmonic. But the harmonics in most cards are so low that the jitter is down
in the noise.

I think that you probably mean a faulty xtal, which
being run at a harmonic, could be the culprit. Have you
heard of this before?
Crystals are still troublesome parts. They age, they
develop unwanted modes, and inspite of the small size,
they are shock sensitive.

I've never seen an audio interface that worked at all that had problems that
were tracable to its crystal clock. Of coruse I've only worked with a few
hundred audio interfaces and the next one might be the exception. ;-)

I don't think it's possible to derive 88.2 and 96 from
the same clock with professional accuracy. Consumer stuff
has been known to use such tricks. So a problem only on
96 points to a crystal.
It would seem that at the worst case I should buy the
second card, try that independantly at Fs 96kHz, and listen
carefully. If that shows the same problem then I will have to
swallow my pride and run at 88k2Hz. This will be just my luck
probably and you 96kHz and 192kHz guys won't mention it again ;-)


I record at 88, because it simplifies the resampling.

Most people record at 44.1 and then you don't have to resample at all. Whoda
thunk? ;-)

[Mark]

Maybe a bad crystal for that frequency, causing audible
jitter.

Terribly unlikely. The usual sources of clock jitter are noise or power
supply artifacts creeping into the clock circuits.


Agreed. Usually the same "physical crystal" is used for all rates.
The divider ratio inside the chip is changed for different rates.

Mark

[Mike Rivers]

On Mar 24, 11:08 am, "Soundhaspriority" wrote:

WRONG. I'm looking at the card, and it has two crystals.

Are you looking at the SCHEMATIC of the card, or just the card? They
may not both be used for clocking audio data in and out of the
converters.

[Julien BH]

Me thinks the problem is the whole "AMD Semptron" in the OP's first
message.
Get a real computer.

[Arny Krueger]

"Soundhaspriority" wrote in message

"Mike Rivers" wrote in message
...
On Mar 24, 11:08 am, "Soundhaspriority"
wrote:
WRONG. I'm looking at the card, and it has two
crystals.

Are you looking at the SCHEMATIC of the card, or just
the card? They may not both be used for clocking audio
data in and out of the converters.

One of the crystals is marked: 24.576
The other crystal is marked: 22.5792

24.576 / 22.5792 = 1.08843537415
48 / 44.1 = 1.08843537415

What a coincidence! Could there be any relationship
between the two crystals and the sampling rate? Of course
not! The second crystal is for people on this group to
argue about.

So, if the problem is the 48/96 KHz crystal, then the board should also be
broken at 48 kHz.

If it works at both 44 and 88, then we know that the divider is probably
good.

[Mark]

On Mar 24, 12:24*pm, "Soundhaspriority" wrote:
"Mike Rivers" wrote in message

...

On Mar 24, 11:08 am, "Soundhaspriority" wrote:

WRONG. *I'm looking at the card, and it has two crystals.

Are you looking at the SCHEMATIC of the card, or just the card? They
may not both be used for clocking audio data in and out of the
converters.

One of the crystals is marked: 24.576
The other crystal is marked: 22.5792

24.576 / 22.5792 = 1.08843537415
48 / 44.1 = 1.08843537415

What a coincidence! Could there be any relationship between the two crystals
and the sampling rate? Of course not! *The second crystal is for people on
this group to argue about.

Find something else more important.

Bob Morein
(310) 237-6511

then they did it the hard way for some reason...

or they were out of 28.224MHz crystals that day. :-).

Mark

[jwvm]

On Mar 24, 12:42 pm, Julien BH wrote:
Me thinks the problem is the whole "AMD Semptron" in the OP's first
message.
Get a real computer.

Really? Semprons are not supercomputers but they are not terribly slow
by recent historical standards. Successful recordings has been made on
considerably sower machines. Assuming that the motherboard can read
the data from the soundcard quickly enough over the PCI bus, its not
obvious at all what effect the processor would have. How efficiently
effects could be done is certainly a function of the CPU but that is
not an issue for acquisition.

[Arny Krueger]

"Julien BH" wrote in message


Me thinks the problem is the whole "AMD Semptron" in the
OP's first message.

Semprons are pretty cheap, but compared to the previous generation chips
(usually only a year or two old) they are often pretty competitive.

Thing is that the cost savings on a Sempron over one of the low end dual
cores is not all that much any more. Why try to save $40 on a chip that you
have to plug into a minium $50 motherboard with $50 worth of RAM that is
talking to a $80 hard drive, a $35 DVD writer and running a $100 operating
system?

Get a real computer.

Which would be?

[Julien BH]

On Mar 24, 3:35 pm, "Arny Krueger" wrote:
"Julien BH" wrote in message



Me thinks the problem is the whole "AMD Semptron" in the
OP's first message.

Semprons are pretty cheap, but compared to the previous generation chips
(usually only a year or two old) they are often pretty competitive.

Thing is that the cost savings on a Sempron over one of the low end dual
cores is not all that much any more. Why try to save $40 on a chip that you
have to plug into a minium $50 motherboard with $50 worth of RAM that is
talking to a $80 hard drive, a $35 DVD writer and running a $100 operating
system?

Get a real computer.

Which would be?

Core 2 Duo all the way duuuude ))

[Julien BH]

On Mar 25, 8:16 am, "Arny Krueger" wrote:
"Julien BH" wrote in message





On Mar 24, 3:35 pm, "Arny Krueger"
wrote:
"Julien BH" wrote in message



Me thinks the problem is the whole "AMD Semptron" in the
OP's first message.

Semprons are pretty cheap, but compared to the previous
generation chips (usually only a year or two old) they
are often pretty competitive.

Thing is that the cost savings on a Sempron over one of
the low end dual cores is not all that much any more.
Why try to save $40 on a chip that you have to plug into
a minium $50 motherboard with $50 worth of RAM that is
talking to a $80 hard drive, a $35 DVD writer and
running a $100 operating system?

Get a real computer.

Which would be?

Core 2 Duo all the way duuuude ))

Well, whatever floats your boat. My video editing box has a A64 6000+ dual
core and it gets the job done. ;-)

The Athlon is not bad at all. That's not what I was saying. It's just
that I for one would skip the Celeron (Intel) or the Semptron (AMD)
for a serious (or not so serious) computer used for recordings.

[Arny Krueger]

"Julien BH" wrote in message


The Athlon is not bad at all. That's not what I was
saying. It's just that I for one would skip the Celeron
(Intel) or the Semptron (AMD) for a serious (or not so
serious) computer used for recordings.

It's hard to get any of us old timers that recorded umpty-dump tracks with
say a Pentium 2. I still remember recording 24 tracks with a 666 MHz
Pentium 2. Of course that was audio recorded clean - no record-time EFX.