15 Apr
2014
15 Apr
'14
3:22 p.m.
On Tue, 15 Apr 2014, James Mckernon wrote:
On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens
<len(a)ovenwerks.net> wrote:
I would realy like to stay away from having to use a USB or FW
audio IF. In fact I would like to be able to continue to use my
delta 66 for as long as I can before I spend more money :) The
Thanks for the useful info in your post. Just to be clear on this part: are
you saying you don't want to switch to USB/FW solely because you want to
keep using your delta 66, or because you have some definite preference for
PCIe over USB/FW devices? If the latter, I wonder why?
USB in audio is limited. Getting clear USB ports interupt wise is not
easy. Audio can not be on a hub or share it's usb with anything else, but
many new MB have no mouse or kb port so the USB is already being used for
that much. The real reason though, is latency. With the pci the latency
can be 1/4 what it can be in USB or FW. That is the lowest seeting jack
for USB or FW is 64/2, but I can run the d66 at 16/2 with no problem on a
well tuned system. This does make a difference for live work. I know that
64/2 seems like very good latency (it is) but remember that the card then
adds another ms in each direction as well as the stage distances on top of
that. That is the time it takes the sound to reach my ear after going
through the computer as a processor and then through the air to my ear.
Maybe that is still not worth worrying about... but even with 30 feet of
cord and no digital delay, I can hear the delay from my playing to the
sound reaching my ear.
--
Len Ovens
www.ovenwerks.net
16 Apr
16 Apr
9:19 a.m.
New subject: hardware - Intel CPUs
On 04/15/2014 03:17 AM, Len Ovens wrote:
On Tue, 15 Apr 2014, James Mckernon wrote:
Interesting. What is the difference between speed of sound in air and
the speed of electricity through a cable?
I think the lowest I've had my UCA-202 (USB 1.1) card work at was around
~3msec latency (on a USB2 port) with a Debian RT kernel just fooling
around (no other RT optimizations on this 2.4GHz i7 laptop). I've had
Yoshimi problems when latency gets too low on my stock kernel. And when
I first tried Zyn, it had problems at much higher latencies than Yoshimi
did at the time.
But I haven't done much music stuff with my new laptop.
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens
<len(a)ovenwerks.net> wrote:
I would realy like to stay away from having to use a USB or FW
audio IF. In fact I would like to be able to continue to use my
delta 66 for as long as I can before I spend more money :) The
Thanks for the useful info in your post. Just to be clear on this
part: are
you saying you don't want to switch to USB/FW solely because you want to
keep using your delta 66, or because you have some definite preference
for
PCIe over USB/FW devices? If the latter, I wonder why?
USB in audio is limited. Getting clear USB ports interupt wise is not
easy. Audio can not be on a hub or share it's usb with anything else,
but many new MB have no mouse or kb port so the USB is already being
used for that much. The real reason though, is latency. With the pci the
latency can be 1/4 what it can be in USB or FW. That is the lowest
seeting jack for USB or FW is 64/2, but I can run the d66 at 16/2 with
no problem on a well tuned system. This does make a difference for live
work. I know that 64/2 seems like very good latency (it is) but remember
that the card then adds another ms in each direction as well as the
stage distances on top of that. That is the time it takes the sound to
reach my ear after going through the computer as a processor and then
through the air to my ear. Maybe that is still not worth worrying
about... but even with 30 feet of cord and no digital delay, I can hear
the delay from my playing to the sound reaching my ear. 
10:20 a.m.
New subject: hardware - Intel CPUs
On 16/04/14 17:19, david wrote:
Interesting. What is the difference between speed of sound in air and the speed
of electricity through a cable?
sound is about 3ms per metre (330 m/s) in air
... in a room, talking about sound, there is no significant delay for the
signal in the cable ... from Sydney to Hawaii via copper or fibre yes,
considering phase differences for RGB video signals in a building yes, talking
about timing on a circuit board or a bus while syncing gigahertz clocks yes, but
for sound within a building the speed over copper may as well be infinite.
I was shown a rather nice artwork that used high resolution radar (resolutions
around 1 centimetre I think) to get positions (it was made in a university
robotics department that had such things!). The way those work is very
interesting ... the frequencies are way to high to digitise, so the electronics
has to be all analogue ... the 'circuitry' is basically plumbing ... gold lined
tubes and chambers using resonances and such to measure delays and phase
differences. At those frequencies the speed of light becomes a dominant
consideration.
Simon
10:40 a.m.
New subject: hardware - Intel CPUs
On 16/04/14 18:20, Simon Wise wrote:
On 16/04/14 17:19, david wrote:
and there are stories around of ridiculously expensive delay units for very
short delays (paid for by US military budgets) and australians replacing them
with a much more reliable, and almost free, solution using some switches and a
circuit board with loops of 1, 2, 4, 8, ... etc centimetres each, but it's the
usual way to do that sort of delay.
Simon
Interesting. What is the difference between speed of sound in air and the speed
of electricity through a cable?
sound is about 3ms per metre (330 m/s) in air
... in a room, talking about sound, there is no significant delay for the signal
in the cable ... from Sydney to Hawaii via copper or fibre yes, considering
phase differences for RGB video signals in a building yes, talking about timing
on a circuit board or a bus while syncing gigahertz clocks yes, but for sound
within a building the speed over copper may as well be infinite.
I was shown a rather nice artwork that used high resolution radar (resolutions
around 1 centimetre I think) to get positions (it was made in a university
robotics department that had such things!). The way those work is very
interesting ... the frequencies are way to high to digitise, so the electronics
has to be all analogue ... the 'circuitry' is basically plumbing ... gold lined
tubes and chambers using resonances and such to measure delays and phase
differences. At those frequencies the speed of light becomes a dominant
consideration. 
3:20 p.m.
New subject: hardware - Intel CPUs
On Wednesday 16 April 2014 09:09:06 Simon Wise did opine:
On 16/04/14 17:19, david wrote:
We have a gismo thats basically much simpler than all that plumbing, to use
when checking a cable for damage, called a Time Domain Reflectometer. The
pro versions using a tunnel diode switch as a pulse generator, can tell you
theres a bullethole in the line 883.6' out from where you are hooked up.
I've made homemade versions using a pulse generator and a fast oscilloscope
to measure the echo delay, punched some buttons on a good calculator and
then told the tower crew where to open it up and replace a burned up
connector bullet and/or the teflon disk holding it centered in the line.
It got the job done so I figured it was good enough for the girls I go
with. :)
Interesting. What is the difference between speed
of sound in air and
the speed of electricity through a cable?
sound is about 3ms per metre (330 m/s) in air
... in a room, talking about sound, there is no significant delay for
the signal in the cable ... from Sydney to Hawaii via copper or fibre
yes, considering phase differences for RGB video signals in a building
yes, talking about timing on a circuit board or a bus while syncing
gigahertz clocks yes, but for sound within a building the speed over
copper may as well be infinite.
I was shown a rather nice artwork that used high resolution radar
(resolutions around 1 centimetre I think) to get positions (it was made
in a university robotics department that had such things!). The way
those work is very interesting ... the frequencies are way to high to
digitise, so the electronics has to be all analogue ... the 'circuitry'
is basically plumbing ... gold lined tubes and chambers using
resonances and such to measure delays and phase differences. At those
frequencies the speed of light becomes a dominant consideration.
Simon _______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS
17 Apr
17 Apr
3:15 a.m.
New subject: hardware - Intel CPUs
On 16/04/14 23:20, Gene Heskett wrote:
On Wednesday 16 April 2014 09:09:06 Simon Wise did
opine:
> I was shown a rather nice artwork that used high
resolution radar
> (resolutions around 1 centimetre I think) to get positions (it was made
> in a university robotics department that had such things!). The way
> those work is very interesting ... the frequencies are way to high to
> digitise, so the electronics has to be all analogue ... the 'circuitry'
> is basically plumbing ... gold lined tubes and chambers using
> resonances and such to measure delays and phase differences. At those
> frequencies the speed of light becomes a dominant consideration.
We have a gismo thats basically much simpler than all that plumbing, to use
when checking a cable for damage, called a Time Domain Reflectometer. The
pro versions using a tunnel diode switch as a pulse generator, can tell you
theres a bullethole in the line 883.6' out from where you are hooked up.
I've made homemade versions using a pulse generator and a fast oscilloscope
to measure the echo delay, punched some buttons on a good calculator and
then told the tower crew where to open it up and replace a burned up
connector bullet and/or the teflon disk holding it centered in the line.
It got the job done so I figured it was good enough for the girls I go
with. :)
not at all convinced that a fast oscilloscope, a good calculator and the
wet-ware connecting them are simpler than a bit of machined metal with gold
plating, but certainly re-using existing gadgets beats making new ones.
Simon
4:11 a.m.
New subject: hardware - Intel CPUs
On Wednesday 16 April 2014 21:57:21 Simon Wise did opine:
On 16/04/14 23:20, Gene Heskett wrote:
Either method requires that the operator have intimate knowledge of the
characteristics of the transmission line being measured. That is
considerably more important then the cost of the tool, particularly when
the results are adequate for the job at hand.
When you are off the air, the cash machine is broken, and that was at least
2 days quicker than getting a TDR overnighted from Vegas when the rental
check had to be in Vegas before the instrument is pulled off the shelf to
send it. A case of time vs money lost. We actually had to pull it apart
and clean up about 80 feet below the burnup because the soot from the
burnup settles on the disks below it, encouraging the next burnup.
Eventually that 40 YO line did get a full rebuild with Andrews new screwed
together inners. End of problem, and now its off the air as we moved the
tx when we made the digital switchover. So if anyone needs a 450 foot
stack of Stainless G5, with a RCA TF3-AL antenna on top of it, with the
rebuilt 3.125" lines, talk to me. But your crew will have to take it down
& ship it.
On Wednesday 16 April 2014 09:09:06 Simon Wise
did opine:
not at all convinced that a fast oscilloscope, a good calculator and the
wet-ware connecting them are simpler than a bit of machined metal with
gold plating, but certainly re-using existing gadgets beats making new
ones.
I was shown a rather nice artwork that used high
resolution radar
(resolutions around 1 centimetre I think) to get positions (it was
made in a university robotics department that had such things!). The
way those work is very interesting ... the frequencies are way to
high to digitise, so the electronics has to be all analogue ... the
'circuitry' is basically plumbing ... gold lined tubes and chambers
using resonances and such to measure delays and phase differences.
At those frequencies the speed of light becomes a dominant
consideration.
We have a gismo thats basically much simpler than all that plumbing,
to use when checking a cable for damage, called a Time Domain
Reflectometer. The pro versions using a tunnel diode switch as a
pulse generator, can tell you theres a bullethole in the line 883.6'
out from where you are hooked up. I've made homemade versions using a
pulse generator and a fast oscilloscope to measure the echo delay,
punched some buttons on a good calculator and then told the tower
crew where to open it up and replace a burned up connector bullet
and/or the teflon disk holding it centered in the line. It got the
job done so I figured it was good enough for the girls I go with. :)
Simon
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS
4:39 a.m.
New subject: hardware - Intel CPUs
On 17/04/14 12:11, Gene Heskett wrote:
On Wednesday 16 April 2014 21:57:21 Simon Wise did
opine:
indeed, it is often so
Simon
On 16/04/14 23:20, Gene Heskett wrote:
Either method requires that the operator have intimate knowledge of the
characteristics of the transmission line being measured. That is
considerably more important then the cost of the tool, particularly when
the results are adequate for the job at hand. On Wednesday 16 April 2014 09:09:06 Simon Wise
did opine:
not at all convinced that a fast oscilloscope, a good calculator and the
wet-ware connecting them are simpler than a bit of machined metal with
gold plating, but certainly re-using existing gadgets beats making new
ones.
I was shown a rather nice artwork that used high
resolution radar
(resolutions around 1 centimetre I think) to get positions (it was
made in a university robotics department that had such things!). The
way those work is very interesting ... the frequencies are way to
high to digitise, so the electronics has to be all analogue ... the
'circuitry' is basically plumbing ... gold lined tubes and chambers
using resonances and such to measure delays and phase differences.
At those frequencies the speed of light becomes a dominant
consideration.
We have a gismo thats basically much simpler than all that plumbing,
to use when checking a cable for damage, called a Time Domain
Reflectometer. The pro versions using a tunnel diode switch as a
pulse generator, can tell you theres a bullethole in the line 883.6'
out from where you are hooked up. I've made homemade versions using a
pulse generator and a fast oscilloscope to measure the echo delay,
punched some buttons on a good calculator and then told the tower
crew where to open it up and replace a burned up connector bullet
and/or the teflon disk holding it centered in the line. It got the
job done so I figured it was good enough for the girls I go with. :)
16 Apr
16 Apr
10:57 p.m.
New subject: hardware - Intel CPUs
Speaking about USB audio IFs... It appears that the USB2 devices have
problems with USB3 ports (Like the 1818vsl). See:
https://forums.presonus.com/posts/list/33427.page
I don't know if this has been fixed by Intel since Ivy. It appears to be a
hardware problem with Intel's chipset where the only workaround is to
remove the USB3 driver/kernel module and use all the USB3 ports as USB2
ports.
--
Len Ovens
www.ovenwerks.net
11:04 p.m.
New subject: hardware - Intel CPUs
On 04/16/2014 10:51 PM, Len Ovens wrote:
It appears that the USB2 devices have problems with
USB3 ports (Like the
1818vsl).
So far I have little issues with my USB2 audio interface (not a 1818VSL)
and the xhci driver. I've come across that thread too and it seems the
issue is specific for the 1818VSL.
I should really do some testing how the ehci and xhci drivers compare,
at the moment I'm inclined to think either the USB3 host controllers I
use or the xhci driver performs better with my interface than either
USB2 controllers or the ehci driver.
Jeremy
17 Apr
17 Apr
1:39 a.m.
New subject: hardware - Intel CPUs
On Wed, 16 Apr 2014, Jeremy Jongepier wrote:
On 04/16/2014 10:51 PM, Len Ovens wrote:
It is specific to the Intel Ivy Bridge chipset USB3 implementation. So if
your computer uses a different HW implementation you should be fine. For
example a PCIe card with USB3 ports is ok or a MB that uses different HW.
Intel admits the problem is there but will not fix it and feels that well
written drivers will just ask the USB device for a resend of the missed
packet !!!??? They say that nobody will notice the missing packet
anyway??? I read this to mean "it is in the hardware now and there is no
soft/firmware patch possible." One hopes the 4th gen chips are better
designed. Thats what I am looking to buy. I don't have a USB2 device to
try though.
It appears that the USB2 devices have problems
with USB3 ports (Like the
1818vsl).
So far I have little issues with my USB2 audio interface (not a 1818VSL)
and the xhci driver. I've come across that thread too and it seems the
issue is specific for the 1818VSL. I should really do some testing how the ehci and xhci
drivers compare,
at the moment I'm inclined to think either the USB3 host controllers I
use or the xhci driver performs better with my interface than either
USB2 controllers or the ehci driver.
That is at least good news. I suspect a HW difference in the host off the
top of my head. I think the system pays more attention to them too.
--
Len Ovens
www.ovenwerks.net
10:26 a.m.
New subject: hardware - Intel CPUs
On 04/16/2014 01:33 PM, Len Ovens wrote:
On Wed, 16 Apr 2014, Jeremy Jongepier wrote:
My laptop has a 4-th gen i7 Haswell chip. My USB 1.1 audio card works
fine on any of the USB2 or 3 ports.
My USB2 Canoscan desktop scanner, on the other hand, doesn't work on
either variety of port on this laptop. It works just fine on the USB2
ports on my AMD-based desktop machine (Phenom IIx4 processor and AMD
chipset).
I found some comments saying the Canon wouldn't work on any USB3 ports,
but I also found the closed-source commercial product VueScan which DOES
work with the Canoscan on USB3 ports ... I suppose it's all related to
the phases of the moon or something.
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
On 04/16/2014 10:51 PM, Len Ovens wrote:
It is specific to the Intel Ivy Bridge chipset USB3 implementation. So
if your computer uses a different HW implementation you should be fine.
For example a PCIe card with USB3 ports is ok or a MB that uses
different HW. Intel admits the problem is there but will not fix it and
feels that well written drivers will just ask the USB device for a
resend of the missed packet !!!??? They say that nobody will notice the
missing packet anyway??? I read this to mean "it is in the hardware now
and there is no soft/firmware patch possible." One hopes the 4th gen
chips are better designed. Thats what I am looking to buy. I don't have
a USB2 device to try though. It appears that the USB2 devices have problems
with USB3 ports (Like the
1818vsl).
So far I have little issues with my USB2 audio interface (not a 1818VSL)
and the xhci driver. I've come across that thread too and it seems the
issue is specific for the 1818VSL.
16 Apr
16 Apr
2:54 p.m.
New subject: hardware - Intel CPUs
On Wednesday 16 April 2014 08:44:15 david did opine:
On 04/15/2014 03:17 AM, Len Ovens wrote:
Sound is nominally 720 miles per hour. Rather leisurely IOW.
A perfect cable is C speed, 258 times faster. But cable (coaxial) actually
range in speeds between 66% of C for home usable cables, to around 98% of C
for 9" diameter high power broadcast stuff, C being 186,272 miles per
second in a vacuum. Thats 298,035.2 kilometers per second for the metric
folks here.
On Tue, 15 Apr 2014, James Mckernon wrote:
Interesting. What is the difference between speed of sound in air and
the speed of electricity through a cable? On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens
<len(a)ovenwerks.net> wrote:
I would realy like to stay away from having to use a USB or FW
audio IF. In fact I would like to be able to continue to use my
delta 66 for as long as I can before I spend more money :) The
Thanks for the useful info in your post. Just to be clear on this
part: are
you saying you don't want to switch to USB/FW solely because you want
to keep using your delta 66, or because you have some definite
preference for
PCIe over USB/FW devices? If the latter, I wonder why?
USB in audio is limited. Getting clear USB ports interrupt wise is not
easy. Audio can not be on a hub or share it's usb with anything else,
but many new MB have no mouse or kb port so the USB is already being
used for that much. The real reason though, is latency. With the pci
the latency can be 1/4 what it can be in USB or FW. That is the
lowest seeting jack for USB or FW is 64/2, but I can run the d66 at
16/2 with no problem on a well tuned system. This does make a
difference for live work. I know that 64/2 seems like very good
latency (it is) but remember that the card then adds another ms in
each direction as well as the stage distances on top of that. That is
the time it takes the sound to reach my ear after going through the
computer as a processor and then through the air to my ear. Maybe
that is still not worth worrying about... but even with 30 feet of
cord and no digital delay, I can hear the delay from my playing to
the sound reaching my ear. I think the lowest I've had my UCA-202 (USB 1.1)
card work at was around
~3msec latency (on a USB2 port) with a Debian RT kernel just fooling
around (no other RT optimizations on this 2.4GHz i7 laptop). I've had
Yoshimi problems when latency gets too low on my stock kernel. And when
I first tried Zyn, it had problems at much higher latencies than Yoshimi
did at the time.
But I haven't done much music stuff with my new laptop.
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS
17 Apr
17 Apr
10:53 a.m.
New subject: hardware - Intel CPUs
On 04/16/2014 02:54 AM, Gene Heskett wrote:
On Wednesday 16 April 2014 08:44:15 david did opine:
Then running your sound from stage to backhouse sound board back to
stage and hearing it through headphones would give no latency at all.
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
On 04/15/2014 03:17 AM, Len Ovens wrote:
Sound is nominally 720 miles per hour. Rather leisurely IOW.
A perfect cable is C speed, 258 times faster. But cable (coaxial) actually
range in speeds between 66% of C for home usable cables, to around 98% of C
for 9" diameter high power broadcast stuff, C being 186,272 miles per
second in a vacuum. Thats 298,035.2 kilometers per second for the metric
folks here. On Tue, 15 Apr 2014, James Mckernon wrote:
Interesting. What is the difference between speed of sound in air and
the speed of electricity through a cable? On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens
<len(a)ovenwerks.net> wrote:
I would realy like to stay away from having to use a USB or FW
audio IF. In fact I would like to be able to continue to use my
delta 66 for as long as I can before I spend more money :) The
Thanks for the useful info in your post. Just to be clear on this
part: are
you saying you don't want to switch to USB/FW solely because you want
to keep using your delta 66, or because you have some definite
preference for
PCIe over USB/FW devices? If the latter, I wonder why?
USB in audio is limited. Getting clear USB ports interrupt wise is not
easy. Audio can not be on a hub or share it's usb with anything else,
but many new MB have no mouse or kb port so the USB is already being
used for that much. The real reason though, is latency. With the pci
the latency can be 1/4 what it can be in USB or FW. That is the
lowest seeting jack for USB or FW is 64/2, but I can run the d66 at
16/2 with no problem on a well tuned system. This does make a
difference for live work. I know that 64/2 seems like very good
latency (it is) but remember that the card then adds another ms in
each direction as well as the stage distances on top of that. That is
the time it takes the sound to reach my ear after going through the
computer as a processor and then through the air to my ear. Maybe
that is still not worth worrying about... but even with 30 feet of
cord and no digital delay, I can hear the delay from my playing to
the sound reaching my ear.
12:13 p.m.
New subject: hardware - Intel CPUs
On Thursday 17 April 2014 06:09:56 david did opine:
Sound is nominally 720 miles per hour. Rather leisurely IOW.
A perfect cable is C speed, 258 times faster. But cable (coaxial)
actually range in speeds between 66% of C for home usable cables, to
around 98% of C for 9" diameter high power broadcast stuff, C being
186,272 miles per second in a vacuum. Thats 298,035.2 kilometers per
second for the metric folks here.
Then running your sound from stage to backhouse sound board back to
stage and hearing it through headphones would give no latency at all.
For an analogue board, small fraction of a millisecond, for a digital
board, anybodies guess. A/D and D/A's are essentially pretty quick, but I'd
still put most digital boards above a millisecond.
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS
On 04/16/2014 02:54 AM, Gene Heskett wrote:
> On Wednesday 16 April 2014 08:44:15 david did opine:
>> On 04/15/2014 03:17 AM, Len Ovens wrote:
>>> On Tue, 15 Apr 2014, James Mckernon wrote:
>>>> On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens <len(a)ovenwerks.net>
wrote:
> I
would realy like to stay away from having to use a USB or
> FW audio IF. In fact I would like to be able to continue to
> use my delta 66 for as long as I can before I spend more
> money :) The
>
> Thanks for the useful info in your post. Just to be clear on this
> part: are
> you saying you don't want to switch to USB/FW solely because you
> want to keep using your delta 66, or because you have some
> definite preference for
> PCIe over USB/FW devices? If the latter, I wonder why?
USB in audio is limited. Getting clear USB ports interrupt wise is
not easy. Audio can not be on a hub or share it's usb with anything
else, but many new MB have no mouse or kb port so the USB is
already being used for that much. The real reason though, is
latency. With the pci the latency can be 1/4 what it can be in USB
or FW. That is the lowest seeting jack for USB or FW is 64/2, but I
can run the d66 at 16/2 with no problem on a well tuned system.
This does make a difference for live work. I know that 64/2 seems
like very good latency (it is) but remember that the card then adds
another ms in each direction as well as the stage distances on top
of that. That is the time it takes the sound to reach my ear after
going through the computer as a processor and then through the air
to my ear. Maybe that is still not worth worrying about... but even
with 30 feet of cord and no digital delay, I can hear the delay
from my playing to the sound reaching my ear.
Interesting. What is the difference between speed of sound in air and
the speed of electricity through a cable? 
12:35 p.m.
New subject: hardware - Intel CPUs
On Thu, Apr 17, 2014 at 06:13:03AM -0400, Gene Heskett wrote:
For an analogue board, small fraction of a
millisecond, for a digital
board, anybodies guess. A/D and D/A's are essentially pretty quick, but I'd
still put most digital boards above a millisecond.
It can be less. This is one of the reasons for using 96 kHz.
If you let the anti-aliasing filters start to roll off well
below the Nyquist rate they can be very short. The latency
is essentially inversely proportional to the width of the
transition band of the filters.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
10:41 p.m.
New subject: hardware - Intel CPUs
On 04/17/2014 12:13 AM, Gene Heskett wrote:
On Thursday 17 April 2014 06:09:56 david did opine:
Then running your sound from stage to backhouse sound board back to
stage and hearing it through headphones would give no latency at all.
For an analogue board, small fraction of a millisecond, for a digital
board, anybodies guess. A/D and D/A's are essentially pretty quick, but I'd
still put most digital boards above a millisecond.
That's interesting. We're looking into replacing the analog board with a
digital one.
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
On 04/16/2014 02:54 AM, Gene Heskett wrote:
> On Wednesday 16 April 2014 08:44:15 david did opine:
>> On 04/15/2014 03:17 AM, Len Ovens wrote:
>>> On Tue, 15 Apr 2014, James Mckernon wrote:
>>>> On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens <len(a)ovenwerks.net>
wrote:
>> I would realy like to stay away from
having to use a USB or
>> FW audio IF. In fact I would like to be able to continue to
>> use my delta 66 for as long as I can before I spend more
>> money :) The
>>
>> Thanks for the useful info in your post. Just to be clear on this
>> part: are
>> you saying you don't want to switch to USB/FW solely because you
>> want to keep using your delta 66, or because you have some
>> definite preference for
>> PCIe over USB/FW devices? If the latter, I wonder why?
>
> USB in audio is limited. Getting clear USB ports interrupt wise is
> not easy. Audio can not be on a hub or share it's usb with anything
> else, but many new MB have no mouse or kb port so the USB is
> already being used for that much. The real reason though, is
> latency. With the pci the latency can be 1/4 what it can be in USB
> or FW. That is the lowest seeting jack for USB or FW is 64/2, but I
> can run the d66 at 16/2 with no problem on a well tuned system.
> This does make a difference for live work. I know that 64/2 seems
> like very good latency (it is) but remember that the card then adds
> another ms in each direction as well as the stage distances on top
> of that. That is the time it takes the sound to reach my ear after
> going through the computer as a processor and then through the air
> to my ear. Maybe that is still not worth worrying about... but even
> with 30 feet of cord and no digital delay, I can hear the delay
> from my playing to the sound reaching my ear.
Interesting. What is the difference between speed of sound in air and
the speed of electricity through a cable?
Sound is nominally 720 miles per hour. Rather leisurely IOW.
A perfect cable is C speed, 258 times faster. But cable (coaxial)
actually range in speeds between 66% of C for home usable cables, to
around 98% of C for 9" diameter high power broadcast stuff, C being
186,272 miles per second in a vacuum. Thats 298,035.2 kilometers per
second for the metric folks here.
18 Apr
18 Apr
12:37 a.m.
New subject: hardware - Intel CPUs
On Thursday 17 April 2014 18:18:05 david did opine:
On 04/17/2014 12:13 AM, Gene Heskett wrote:
We replaced an old 12 stereo channel 48 input logitek analog board,
couldn't keep ahead of replacing the CDS channel switches in it because it
was full of 5532's and even with added fans, their life in that heat was
about a year, with a 24 input Macky(SP?). Digital, programmable motorized
gain pots, very low noise. But we had to shut the studio monitors off, the
delay had our talent stumbling over their own shoelaces trying to read the
teleprompter.
I didn't have a way to measure it, other than a dual trace scope and hand
claps which said 16 to 17 ms, but I'd have guessed around 20. Knowing what
to expect, I still couldn't read the opening paragraph of a tv technology
article without stumbling lots of times if the monitor was loud enough to
sound like an echo. The TD can still talk in their ear, but no backfeed
goes out that circuit, its killed when the mics go live.
We have an even newer board now, 48 stereo wide SMPTE out, and I believe
its considerable faster than the Macky was, but I don't think the studio
monitors are used for anything but talkback during commercials yet. I've
been retired since mid 2002, and haven't really tracked that stuff 100%
since.
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS
On Thursday 17 April 2014 06:09:56 david did
opine:
For an analogue board, small fraction of a millisecond, for a digital
board, anybodies guess. A/D and D/A's are essentially pretty quick,
but I'd still put most digital boards above a millisecond.
That's interesting. We're looking into replacing the analog board with a
digital one. On 04/16/2014 02:54 AM, Gene Heskett wrote:
> On Wednesday 16 April 2014 08:44:15 david did opine:
>> On 04/15/2014 03:17 AM, Len Ovens wrote:
>>> On Tue, 15 Apr 2014, James Mckernon wrote:
>>>> On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens <len(a)ovenwerks.net>
wrote:
>>> I would realy like to stay away from having to use a USB
>>> or FW audio IF. In fact I would like to be able to
>>> continue to use my delta 66 for as long as I can before
>>> I spend more money :) The
>>>
>>> Thanks for the useful info in your post. Just to be clear on this
>>> part: are
>>> you saying you don't want to switch to USB/FW solely because you
>>> want to keep using your delta 66, or because you have some
>>> definite preference for
>>> PCIe over USB/FW devices? If the latter, I wonder why?
>>
>> USB in audio is limited. Getting clear USB ports interrupt wise is
>> not easy. Audio can not be on a hub or share it's usb with
>> anything else, but many new MB have no mouse or kb port so the
>> USB is already being used for that much. The real reason though,
>> is latency. With the pci the latency can be 1/4 what it can be in
>> USB or FW. That is the lowest seeting jack for USB or FW is 64/2,
>> but I can run the d66 at 16/2 with no problem on a well tuned
>> system. This does make a difference for live work. I know that
>> 64/2 seems like very good latency (it is) but remember that the
>> card then adds another ms in each direction as well as the stage
>> distances on top of that. That is the time it takes the sound to
>> reach my ear after going through the computer as a processor and
>> then through the air to my ear. Maybe that is still not worth
>> worrying about... but even with 30 feet of cord and no digital
>> delay, I can hear the delay from my playing to the sound reaching
>> my ear.
>
> Interesting. What is the difference between speed of sound in air
> and the speed of electricity through a cable?
Sound is nominally 720 miles per hour. Rather leisurely IOW.
A perfect cable is C speed, 258 times faster. But cable (coaxial)
actually range in speeds between 66% of C for home usable cables, to
around 98% of C for 9" diameter high power broadcast stuff, C being
186,272 miles per second in a vacuum. Thats 298,035.2 kilometers
per second for the metric folks here.
Then running your sound from stage to backhouse sound board back to
stage and hearing it through headphones would give no latency at all.
8:41 a.m.
New subject: hardware - Intel CPUs
On 04/17/2014 12:37 PM, Gene Heskett wrote:
On Thursday 17 April 2014 18:18:05 david did opine:
How long a delay is there between an organ note and the return echo in a
large cathedral? Much more than the millisecond scale, maybe?
Ages ago, we were at a Day on the Green concert in Oakland Coliseum, a
baseball field in California. Featured 5 big name rock bands. Stage was
set up around and behind home plate. We were on sitting on the grass of
the field, behind us was the 2-3 tiers of concrete seating.
One of the bands that was playing was Heart; at the time, a key element
of their sound was how clean it was. They were the 4th band to play.
They got up and played, and their sound was awful because of the
enormous echo coming off the concrete behind us. It was almost as loud
as they were, but heavily delayed. They went through their show and were
probably relieved. The first three bands were just straight loud rock
bands, they probably didn't even notice the echo, or thought it was cool
because it made things louder. Heart was followed by The Eagles (this
was part of their Hotel California tour) with Joe Walsh at the time.
They very skillfully integrated the echo into their performance, timing
things so the echo sounded like it was part of the music.
On 04/17/2014 12:13 AM, Gene Heskett wrote:
We replaced an old 12 stereo channel 48 input logitek analog board,
couldn't keep ahead of replacing the CDS channel switches in it because it
was full of 5532's and even with added fans, their life in that heat was
about a year, with a 24 input Macky(SP?). Digital, programmable motorized
gain pots, very low noise. But we had to shut the studio monitors off, the
delay had our talent stumbling over their own shoelaces trying to read the
teleprompter.
I didn't have a way to measure it, other than a dual trace scope and hand
claps which said 16 to 17 ms, but I'd have guessed around 20. On Thursday 17 April 2014 06:09:56 david did
opine:
That's interesting. We're looking into replacing the analog board with a
digital one. On 04/16/2014 02:54 AM, Gene Heskett wrote:
> On Wednesday 16 April 2014 08:44:15 david did opine:
>> On 04/15/2014 03:17 AM, Len Ovens wrote:
>>> On Tue, 15 Apr 2014, James Mckernon wrote:
>>>> On Mon, Apr 14, 2014 at 2:06 AM, Len Ovens <len(a)ovenwerks.net>
wrote:
>>>> I would realy like to
stay away from having to use a USB
>>>> or FW audio IF. In fact I would like to be able to
>>>> continue to use my delta 66 for as long as I can before
>>>> I spend more money :) The
>>>>
>>>> Thanks for the useful info in your post. Just to be clear on this
>>>> part: are
>>>> you saying you don't want to switch to USB/FW solely because you
>>>> want to keep using your delta 66, or because you have some
>>>> definite preference for
>>>> PCIe over USB/FW devices? If the latter, I wonder why?
>>>
>>> USB in audio is limited. Getting clear USB ports interrupt wise is
>>> not easy. Audio can not be on a hub or share it's usb with
>>> anything else, but many new MB have no mouse or kb port so the
>>> USB is already being used for that much. The real reason though,
>>> is latency. With the pci the latency can be 1/4 what it can be in
>>> USB or FW. That is the lowest seeting jack for USB or FW is 64/2,
>>> but I can run the d66 at 16/2 with no problem on a well tuned
>>> system. This does make a difference for live work. I know that
>>> 64/2 seems like very good latency (it is) but remember that the
>>> card then adds another ms in each direction as well as the stage
>>> distances on top of that. That is the time it takes the sound to
>>> reach my ear after going through the computer as a processor and
>>> then through the air to my ear. Maybe that is still not worth
>>> worrying about... but even with 30 feet of cord and no digital
>>> delay, I can hear the delay from my playing to the sound reaching
>>> my ear.
>>
>> Interesting. What is the difference between speed of sound in air
>> and the speed of electricity through a cable?
>
> Sound is nominally 720 miles per hour. Rather leisurely IOW.
> A perfect cable is C speed, 258 times faster. But cable (coaxial)
> actually range in speeds between 66% of C for home usable cables, to
> around 98% of C for 9" diameter high power broadcast stuff, C being
> 186,272 miles per second in a vacuum. Thats 298,035.2 kilometers
> per second for the metric folks here.
Then running your sound from stage to backhouse sound board back to
stage and hearing it through headphones would give no latency at all.
For an analogue board, small fraction of a millisecond, for a digital
board, anybodies guess. A/D and D/A's are essentially pretty quick,
but I'd still put most digital boards above a millisecond. Knowing what
to expect, I still couldn't read the opening paragraph of a tv technology
article without stumbling lots of times if the monitor was loud enough to
sound like an echo. The TD can still talk in their ear, but no backfeed
goes out that circuit, its killed when the mics go live.
Hmm, I've never tried reading on TV. Worked on-air in a radio station,
but that was all headphones, and much smaller physical spaces.
We have an even newer board now, 48 stereo wide SMPTE
out, and I believe
its considerable faster than the Macky was, but I don't think the studio
monitors are used for anything but talkback during commercials yet. I've
been retired since mid 2002, and haven't really tracked that stuff 100%
since.
Good to leave the work life behind and put your energy into your own
life! I transitioned from IT/IS work through systems analyst/business
analyst to what I do now, which is manage content and work with
enterprise content management systems from the business side. Since we
dropped using Windows in our house about 10 years ago, I'm way behind on
a lot of Windows things!
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
17 Apr
17 Apr
10:02 p.m.
New subject: hardware - Intel CPUs
On Wed, 16 Apr 2014, david wrote:
Then running your sound from stage to backhouse sound
board back to stage
and hearing it through headphones would give no latency at all.
Nothing I can hear. This is very important becasue if there was noticable
delay then stage sound would be ahead of the mains... might give some
interesting filtering from being out of phase.
--
Len Ovens
www.ovenwerks.net
9:58 p.m.
New subject: hardware - Intel CPUs
On 04/16/2014 02:54 PM, Gene Heskett wrote:
Sound is nominally 720 miles per hour. Rather leisurely IOW.
A perfect cable is C speed, 258 times faster. But cable (coaxial) actually
range in speeds between 66% of C for home usable cables, to around 98% of C
for 9" diameter high power broadcast stuff, C being 186,272 miles per
second in a vacuum. Thats 298,035.2 kilometers per second for the metric
folks here.
Just a tiny comment to the speed ratio: There is a missing factor
seconds/hour. Light is travelling in the magnitude of one million faster
than sound in air.
/ Hans
18 Apr
18 Apr
12:16 a.m.
New subject: hardware - Intel CPUs
On Thursday 17 April 2014 18:12:12 Hans Wilmers did opine:
On 04/16/2014 02:54 PM, Gene Heskett wrote:
Gack! I hate it when I miss the obvious stuffs. So 258 or 259 *3600 for
the speed ratio in round figures. Round figures light is 930,600 times
faster than sound.
Definitely my bad. Open reply, insert foot fresh from mouth. Duh!
Sound is nominally 720 miles per hour. Rather
leisurely IOW.
A perfect cable is C speed, 258 times faster. But cable (coaxial)
actually range in speeds between 66% of C for home usable cables, to
around 98% of C for 9" diameter high power broadcast stuff, C being
186,272 miles per second in a vacuum. Thats 298,035.2 kilometers per
second for the metric folks here.
Just a tiny comment to the speed ratio: There is a missing factor
seconds/hour. Light is travelling in the magnitude of one million faster
than sound in air.
/ Hans
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS
16 Apr
16 Apr
3:23 p.m.
New subject: hardware - Intel CPUs
On Tue, 15 Apr 2014, david wrote:
Interesting. What is the difference between speed of
sound in air and the
speed of electricity through a cable?
I was talking about through air. I can only play as fast as I can hear and
at thirty feet or so from the sound source, my instrument sounds delayed
from the rest of the band because I hear their sound that much later and
play my part that much later. I don't normally play at that distance, but
ten feet is pretty standard.
I think the lowest I've had my UCA-202 (USB 1.1)
card work at was around
~3msec latency (on a USB2 port) with a Debian RT kernel just fooling around
(no other RT optimizations on this 2.4GHz i7 laptop). I've had Yoshimi
problems when latency gets too low on my stock kernel. And when I first
tried Zyn, it had problems at much higher latencies than Yoshimi did at the
time.
Is that measured round trip? Or one way through the jack part? What sample
rate? (I'm using 48k) I can get qjackctl so it says I have .6ms latency.
But I think that is one way, 1.2ms ... then add ~1 ms each way just to get
through the sound card monitor mixing and maybe a bit more and I have
already three and a half ms. Then I am some distance away from my monitor
too. Perhaps with a newer faster CPU I could run at 96k and work with a FW
device and get the same results. USB is very handy for many things but for
sound it takes some playing around to get a clean low latency set up.
However, it looks like I can still get lots of MB with PCI slots in them.
I will probaly do that. Hopefully with three PCI slots I can get one that
is irq clean.
--
Len Ovens
www.ovenwerks.net
4012
days inactive
4015
days old
linux-audio-user@lists.linuxaudio.org
22 comments
7 participants
participants (7)
-
david -
Fons Adriaensen -
Gene Heskett -
Hans Wilmers -
Jeremy Jongepier -
Len Ovens -
Simon Wise