17 Apr
2014
17 Apr
'14
9:59 p.m.
On Wed, 16 Apr 2014, david wrote:
cat /proc/interrupts will tell you that quick enough. For some reason a
lot of MB share irq16 with 3 or 4 things... worse one of those tings is
often one of the PCI(e) slots. No need really as most modern MB have
access to 48 plus irqs. My old board has 24 but 2-7 are unused (for
hysterical reasons?) and 10 and 13 seem to be skipped too. I think 20 goes
to my internal audio which I have turned off (AC97) so it doesn't show.
I have noticed that on any of the MB I have checked or seen irq maps for
that the internal audio always has it's own irq and quite often it is the
highest one which in most systems these days has the most priority. I
found this out because I had the card I was using for midi in the higher
of the two slots and my audio below. I had trouble with xruns on the
audio, but when I put the audio on the higher irq I had no more trouble.
I know that in theory that shouldn't happen because there are two part to
the irq drivers, a stub to answer the IRQ and save enough info to work and
the other part that the os prioritizes and does all the work. So the os
should be able to prioritize by the module name. I just know what I have
found works best.
I have heard the words "in a modern system" too many times. I think any
system can do better audio if it is tuned/tweaked.
--
Len Ovens
www.ovenwerks.net
On 04/16/2014 03:17 AM, Len Ovens wrote:
I used to play and try to adjust sound while practicing, so I had 50 feet
or so of cable. By the time I got to the sound board it was getting hard
to play.
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.
You're a few orders of magnitude better than me! I don't think I'd notice
that much difference. I use headphone monitor.
I find monitoring with headphones when playing bass is hard. I use a small
amp tilted back about 10 feet ahead of me... so 13-14 feet to my ear I
guess. Bass needs some space to work. My amp just gives me enough for me
and the board can add whatever they want to the mains. I go a bit heavy on
the high end for my monitor and the mains can roll that off for house
sound.
That's my understanding, too. Things like staying
away from USB mice and
keyboards, making sure the USB audio is the only device on that bus, yah?
On that irq... same for pci(e) really even prioritizing shared irqs, lone
irqs are still best.
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.
My desktop has 2 PCIe slots; the other 4 are PCI. It also has 4 USB ports +
2 more USB connector points on the mobo, 4 SATA connectors, an EIDE
connector, a floppy connector, plus the built-in audio and video and
Ethernet. I haven't checked to see what's sharing interrupts and what isn't.
I wouldn't be surprised if there aren't a bunch of shared interrupts! 
18 Apr
18 Apr
9:08 a.m.
New subject: hardware - Intel CPUs
On 04/17/2014 09:53 AM, Len Ovens wrote:
cat /proc/interrupts will tell you that quick enough. For some reason a
lot of MB share irq16 with 3 or 4 things... worse one of those tings is
often one of the PCI(e) slots. No need really as most modern MB have
access to 48 plus irqs. My old board has 24 but 2-7 are unused (for
hysterical reasons?) and 10 and 13 seem to be skipped too. I think 20
goes to my internal audio which I have turned off (AC97) so it doesn't
show.
I think mine has a clump on IRQ16. I thought leaving behind the old IBM
AT bus architecture was supposed to free us from IRQ problems?
On Wed, 16 Apr 2014, david wrote:
I think I would, too. When I was playing bass on stage ages ago, I don't
think I got more than 20 feet from my amp. I wasn't a singer, so no
need. The one point where we needed cable was when the lead guitarist
and I would finish one song by somersaulting across the stage while playing.
My church band's former music director used to wirelessly hook his
guitar into the system, then walk around while we rehearsed so he could
hear how things sounded in the auditorium. That was in the days when we
could get the place whenever we wanted to set everything up for
rehearsal. Our sound tech has gone around with a soundmeter in our new
location and fiddled with levels. And if one of us has to miss practice
(our rule is you don't get to play that Sunday), we try to sit various
places and afterwards talk about how things sounded.
On 04/16/2014 03:17 AM, Len Ovens wrote:
I used to play and try to adjust sound while practicing, so I had 50
feet or so of cable. By the time I got to the sound board it was getting
hard to play. 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.
You're a few orders of magnitude better than me! I don't think I'd
notice that much difference. 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.
My desktop has 2 PCIe slots; the other 4 are PCI. It also has 4 USB
ports + 2 more USB connector points on the mobo, 4 SATA connectors, an
EIDE connector, a floppy connector, plus the built-in audio and video
and Ethernet. I haven't checked to see what's sharing interrupts and
what isn't. I wouldn't be surprised if there aren't a bunch of shared
interrupts! I have noticed that on any of the MB I have checked or
seen irq maps for
that the internal audio always has it's own irq and quite often it is
the highest one which in most systems these days has the most priority.
I found this out because I had the card I was using for midi in the
higher of the two slots and my audio below. I had trouble with xruns on
the audio, but when I put the audio on the higher irq I had no more
trouble.
I've discovered on my desktop that if I disable the onboard audio (which
I don't use at all), ALSA doesn't load or start. But that motherboard on
that is starting to have problems - sometimes it will boot up with a
blank screen for a moment or so, then the BIOS display appears with a
red warning that the "last overclocking setting failed". Which is odd
because I have never and am not now overclocking it. I guess if it
finally goes out I'll replace it with an Intel board. I like the Haswell
architecture, it gives my laptop great performance while maintaining
battery life (about 3.5hrs with my 2.4GHz i7). The previous model used
slower Intel chips to squeeze out that battery life. Now if only video
support for the HD4600 chip gets better - playing MP4s or other video
formats results in frozen image except for a band or two that change.
Ubuntu 13.10 can play video on it (if you install a bunch of libraries
from outside the Ubuntu reality), so maybe it's just taking some time
for Ubuntu's X tweaks to get back into the general X codebase.
I know that in theory that shouldn't happen
because there are two part
to the irq drivers, a stub to answer the IRQ and save enough info to
work and the other part that the os prioritizes and does all the work.
So the os should be able to prioritize by the module name. I just know
what I have found works best.
I have heard the words "in a modern system" too many times. I think any
system can do better audio if it is tuned/tweaked.
Sure! I even had decent audio working on my old effectsbox laptop, which
was an old Toshiba with a 2.8GHz Celeron and only 768MB of memory. It
was old before it landed in my collection! (Audio quality was fine.
Processor, on the other hand, was not really up to the job!)
Toshiba laptops used to be really easy to use with Linux; Toshiba used
"bog standard hardware" (as a friend of mine called it) and everything
had full, mature Linux support. Don't know about current Toshibas.
They've probably been pulled further into the Windows orbit since then.
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
9:52 a.m.
New subject: hardware - Intel CPUs
david wrote:
Toshiba laptops used to be really easy to use with
Linux; Toshiba used "bog
standard hardware" (as a friend of mine called it) and everything had full,
mature Linux support. Don't know about current Toshibas. They've probably
been pulled further into the Windows orbit since then.
Ack! I had a terrible experience with a Toshiba laptop from
maybe five years ago, a Satellite. The fan control didn't
work at all under Linux. None of the toshiba-specific kernel
modules helped. I needed for the fan to turn on during the
power-on sequence, and then, after boot, needed to turn down
the CPU frequency to the lowest setting, or it would
overheat. It was always iffy, and I had to prop up the
laptop on top of some improvised standoffs so that the
airflow would be sufficient.
When I finally got a new laptop, I ended up running a
vaccuum cleaner with the hose again against the outflow vent
for a couple hours while I made a final copy of the
important hard disk partitions.
I had forgotten this horrible trauma until you reminded me.
Thanks :-)
Joel
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
--
Joel Roth
9:57 p.m.
New subject: hardware - Intel CPUs
On 04/17/2014 09:52 PM, Joel Roth wrote:
david wrote:
Your Toshiba is much newer than any of the three we had here (have only
one left, my wife's)! We've never had any problems with fan or other
thermal control on any of them, except the very oldest. When it turned
12, it started overheating. The fan would run like crazy with little
effect. Once the machine shut down when it got too hot.
I probably could have fixed it by opening it up and cleaning it out
again, but the new laptop was arriving and I decided I no longer needed
two laptops. Sitting it on a spare laptop cooler kept it working long
enough. (We live in rather warm area of Hawaii without A/C.)
But through it all, Linux supported all the Toshiba laptop features
without any complaints for hoop-jumping.
--
David W. Jones
gnome(a)hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com
Toshiba laptops used to be really easy to use
with Linux; Toshiba used "bog
standard hardware" (as a friend of mine called it) and everything had full,
mature Linux support. Don't know about current Toshibas. They've probably
been pulled further into the Windows orbit since then.
Ack! I had a terrible experience with a Toshiba laptop from
maybe five years ago, a Satellite. The fan control didn't
work at all under Linux. None of the toshiba-specific kernel
modules helped. I needed for the fan to turn on during the
power-on sequence, and then, after boot, needed to turn down
the CPU frequency to the lowest setting, or it would
overheat. It was always iffy, and I had to prop up the
laptop on top of some improvised standoffs so that the
airflow would be sufficient.
When I finally got a new laptop, I ended up running a
vaccuum cleaner with the hose again against the outflow vent
for a couple hours while I made a final copy of the
important hard disk partitions.
I had forgotten this horrible trauma until you reminded me.
Thanks :-)
9:01 p.m.
New subject: hardware - Intel CPUs
On Thu, 17 Apr 2014, david wrote:
I've discovered on my desktop that if I disable
the onboard audio (which I
don't use at all), ALSA doesn't load or start.
My multi-tracking audio card is pci so the system is seeing it before alsa
loads, that may not be the case with USB audio. In fact with a PCI sound
card, the bios may have to deal with it before the kernel ever sees the
system, so it is the same as an internal card. I use two cards, I have a
delta 66 for tracking and an ensoniq (pre soundblaster) card that is only
used for it's MIDI interface. I wish I could just load the midi part of
the ens1370 driver, but I guess the audio quality is (for what it is)
usable if I really want 8 ins (I only have 8 mic pres)... but I have yet
to use more than 2 inputs at once anyway.
Jackdbus starts with session, PA has all devices turned off so any audio
goes through jack. I think I will hack together something in the systray
for changing jack's latency... maybe turn off pa-jack as latency goes
down... turn off some of the system services while I am at it, deal with
ondemand too.
--
Len Ovens
www.ovenwerks.net
9:27 p.m.
New subject: hardware - Intel CPUs
On Fri, 18 Apr 2014 11:55:40 -0700 (PDT)
Len Ovens <len(a)ovenwerks.net> wrote:
On Thu, 17 Apr 2014, david wrote:
This was my experience too. PCI card and you can disable the motherboard
sound. USB card and you can't.
--
Will J Godfrey
http://www.musically.me.uk
Say you have a poem and I have a tune.
Exchange them and we can both have a poem, a tune, and a song.
I've discovered on my desktop that if I
disable the onboard audio (which I
don't use at all), ALSA doesn't load or start.
My multi-tracking audio card is pci so the system is seeing it before alsa
loads, that may not be the case with USB audio. In fact with a PCI sound
card, the bios may have to deal with it before the kernel ever sees the
system, so it is the same as an internal card.
4068
days inactive
4069
days old
linux-audio-user@lists.linuxaudio.org
5 comments
4 participants
participants (4)
-
david -
Joel Roth -
Len Ovens -
Will Godfrey