Re: [Alsa-user] ice1712 recording

[zcx]

On 08/12/16 18:18, James Shatto wrote:

> No, I mean you can record all four channels as input at the same time
> with the same app.  Hence the -c 4 aka 4 channels.  In audacity you
> just select 4 channels and press record.  Once recorded you can break
> them out into 4 mono channels and save each individually (unmixed).
> There's no mixing involved until you configure it to do so.  WAV files
> and other audio formats often contain multiple tracks.  Individual
> unmixed tracks.  They are mixed at the time of playback if configured
> to do so.

Yes that works.  I had to select M Audio Delta 44: ICE1712 Multi(hw:0,0) 
for recording.

Page 4-30 of the ice1712/envy24 data sheet shows that a multi channel 
buffer is used to record "a seamless flow of multiple streams".

So the hardware is capable of multitrack record without using any sound 
servers.

Anyway this seems to work with arecord ...

pcm.D44_cslave {
type dsnoop
ipc_key 223456
slave {
pcm "hw:0,0"
rate48000
period_time 0   # in microseconds
buffer_size 256 # in bytes
period_size 64  # in bytes
channels4
format  S32_LE
}
bindings {
0 0
1 1
2 2
3 3
}
}

pcm.D44_ch12_capture {
typeplug
ttable.0.0 1
ttable.1.1 1
slave.pcm   D44_cslave
}

pcm.D44_ch34_capture {
typeplug
ttable.0.2 1
ttable.1.3 1
slave.pcm   D44_cslave
}

Terminal 1 ...

$ arecord -v -DD44_ch12_capture -r 48000 -c 2 -d 60 -fS32_LE ./test12.wav

Terminal 2 ...

$ arecord -v -DD44_ch34_capture -r 48000 -c 2 -d 60 -fS32_LE ./test34.wav

No XRUNs but the buffer settings could do with some further optimisation.

The key is to use a dsnoop instead of a hw as the slave.

Thanks for your help James.


--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today.http://sdm.link/xeonphi
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ice1712 recording

[Ralf Mardorf]

On Thu, 8 Dec 2016 12:18:47 -0600, James Shatto wrote:
>As well as a few CLI options for the same.  Sox is good for that.

For SpaceFM I wrote a script to split and/or merge stereo wav. Written
on the fly, so most likely not good shell script writing, but it does
what it should do and IIRC ffmpeg was easier to use.

[rocketmouse@archlinux bin]$ ls -Ggl s2*
lrwxrwxrwx 13 Apr 27  2016 s2a -> s2m
lrwxrwxrwx 13 Apr 27  2016 s2d -> s2m
lrwxrwxrwx 13 Apr 27  2016 s2l -> s2m
-rwxr-xr-x 1 1791 Apr 28  2016 s2m
lrwxrwxrwx 13 Apr 27  2016 s2r -> s2m
[rocketmouse@archlinux bin]$ cat s2m
#!/bin/dash

version="2016-04-28"

usage()
{
  cat< mono.wav:
s2m input_file.wav
  stereo.wav > left.wav:  
s2l input_file.wav
  stereo.wav > right.wav:
s2r input_file.wav
  stereo.wav > dual {left,right}.wav:
s2d input_file.wav
  stereo.wav > all {mono,left,right}.wav:
s2a input_file.wav

EOF
  exit $1
}

outfile()
{
  outfile="$(echo "$infile" | sed 's/\(.*\).wav/\1_'$1.$suffix/I)"
  if [ -f "$outfile" ]; then
echo "$outfile already exists"
usage 1
  fi
}

ex2file()
{
  outfile $1
  echo "Export to $outfile"
  case $1 in
mono)
  ffmpeg -i "$infile" -ac 1 "$outfile"
;;
left)
  ffmpeg -i "$infile" -map_channel 0.0.0 "$outfile"
;;
right)
  ffmpeg -i "$infile" -map_channel 0.0.1 "$outfile"
;;
  esac
}

case $1 in
  -h|--help)
usage 0
  ;;
esac

infile="$1"
if [ ! -f "$infile" ]; then
  echo "No file $infile"
  usage 1
fi

num_ch=$(exiftool "$infile" | grep -v "File Name" | grep -v "Directory" | grep 
"Num Channels" | cut -d: -f2 | sed s/\ //)
plural=""
if [ "$num_ch" = "" ]; then
  num_ch="No"
else 
  if [ "$num_ch" -ge "2" ]; then
plural="s"
  fi
fi
if [ "$num_ch" != "2" ]; then
  echo "$num_ch channel$plural, not a stereo file"
  usage 1
fi

suffix=$(echo "$infile" | rev | cut -d. -f1 | rev)
if [ "$(echo "$suffix" | tr [:upper:] [:lower:])" != "wav" ]; then
  echo "Suffix is \"$suffix\", must be \"wav\", case sensitivity is not 
required"
  usage 1
fi

case $(basename $0) in
  s2m)
ex2file mono
  ;;
  s2l)
ex2file left
  ;;
  s2r)
ex2file right
  ;;
  s2d)
outfile right
ex2file left
ex2file right
  ;;
  s2a)
outfile left
outfile right
ex2file mono
ex2file left
ex2file right
  ;;
esac

exit



--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today.http://sdm.link/xeonphi
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ice1712 recording

[James Shatto]

> You can mix all 4 inputs down into one stream and then record that, but
do you really mean that you can record to 4 separate application threads
concurrently without mixing?

No, I mean you can record all four channels as input at the same time
with the same app.  Hence the -c 4 aka 4 channels.  In audacity you
just select 4 channels and press record.  Once recorded you can break
them out into 4 mono channels and save each individually (unmixed).
There's no mixing involved until you configure it to do so.  WAV files
and other audio formats often contain multiple tracks.  Individual
unmixed tracks.  They are mixed at the time of playback if configured
to do so.

By default sound is often configured for stereo with one track panned
left and one track panned right.  But it's two tracks, not mixed
(until reproduced and analog-illy mixed in air).  Baring cheap sound
devices that have bleed over between tracks with unintentional mixing.
In audacity you can separate the channels and unpan them for two true
mono tracks.  As well as a few CLI options for the same.  Sox is good
for that.

With unix-isms you can send your output to stdout and pipe to stdin of
another app.  Like tee which can then save a file and redirect the
same output to another file.  I've sometimes gone this route to output
the raw WAV and a compressed MP3 at the same time to two different
storage devices / locations.

I seem to recall an arecord option to output each channel as it's own
mono file.  Which you could tail -f on other terminals to pipe that to
other things.  Seems like --separate-channels is that option.

$ arecord --help

- James

--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today.http://sdm.link/xeonphi
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ice1712 recording

[Ralf Mardorf]

On Mon, 5 Dec 2016 14:27:22 +, zcx wrote:
>On 04/12/16 21:09, Ralf Mardorf wrote:
>> Only one app can grab the device, if you run two instances of the
>> same app, only one instance can grab the device.  
>
>Yes, this appears to be the case.  But do you know why?

A car has got four wheels, why can't you drive each wheel to a
completely different place, at the same time? Why are the four wheels
in sync? It doesn't matter what car you own, by design it's the same for
all cars. A car with four wheels and four seats provides only one seat
for a driver and only one steering wheel.

As already pointed out, one sound server is able to access the device
and several apps could access the sound server. Problem solved.

Regards,
Ralf

-- 
"Warum passieren mir immer Sachen, die sonst nur dämlichen Menschen
passieren?" - Homer Simpson.

-- 
"Pull a Homer -- to succeed despite idiocy." - The Simpsons

--
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ice1712 recording

[zcx]

On 04/12/16 21:09, Ralf Mardorf wrote:
> Only one app can grab the device, if you run two instances of the same
> app, only one instance can grab the device.

Yes, this appears to be the case.  But do you know why?


--
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ice1712 recording

[zcx]

On 04/12/16 21:51, James Shatto wrote:
> It can record from all 4.

You can mix all 4 inputs down into one stream and then record that, but
do you really mean that you can record to 4 separate application threads 
concurrently without mixing?

> I suspect that adding -c 4 would overcome the issue.  The -f dat has a
> default of -f S16_LE -c 2 -r 48000.  So you might want to change that
> parameter as well.

Terminal 1 ...
$ arecord -DD44_ch12_capture -f S24_LE -c 4 -r 96000 test12.wav
Recording WAVE 'test12.wav' : Signed 24 bit Little Endian, Rate 96000 
Hz, Channels 4

Terminal 2 ...
$ arecord -DD44_ch34_capture -f S24_LE -c 4 -r 96000 test34.wav
arecord: main:722: audio open error: Device or resource busy

Terminal 1 ...
^C

Terminal 2 ...
$ arecord -DD44_ch34_capture -f S24_LE -c 4 -r 96000 test34.wav
Recording WAVE 'test34.wav' : Signed 24 bit Little Endian, Rate 96000 
Hz, Channels 4

Terminal 1 ...
$ arecord -DD44_ch12_capture -f S24_LE -c 4 -r 96000 test12.wav
arecord: main:722: audio open error: Device or resource busy

i.e. the same behaviour as before.  It appears that one thread is 
blocking the other.

> I have that card, although I haven't booted that old PCI system in a
> while.

Please try it and see if you can get it to work with multiple concurrent 
input streams.

> It's a nice card, a shame
> most newer things don't have PCI in favor of PCIe.

This core2duo machine has a gigabyte motherboard with a single PCI slot.

The application also has two by two channel separate outputs.  It is 
easy to separate those to Channel 1 & 2 and Channel 3 & 4 in .asoundrc 
using the dshare plugin.

I have tried to separate the two by two channel separate inputs using 
two instanced of dsnoop, but it does not work.  You get the same 
blocking problem that is exhibited by arecord.

> Assuming that it didn't get index 0 in /proc/asound/cards.

The Delta 44 is at index 0.

zcx


--
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ice1712 recording

[James Shatto]

It can record from all 4.  Although many applications only care about
left / channel 1 (defaults).  I tend to run pulseaudio over jackd
setup.  As that was the only way to have your mic be an input other
than 1 for apps like skype.  Since channel 1 and 2 are typically
stereo output.

$ man arecord

-c, --channels=#
  the number of channels.  The default is one channel. .

I suspect that adding -c 4 would overcome the issue.  The -f dat has a
default of -f S16_LE -c 2 -r 48000.  So you might want to change that
parameter as well.

$ arecord -Dhw:1 -f S24_LE -t wav -c 4 -r 96000 output.wav

I have that card, although I haven't booted that old PCI system in a
while.  What I recall of that card, that should be the highest
sampling available for it.  24 bit, 96kHz, 4 channels (input).  In
theory it has a 10 channel output mixer.  It's a nice card, a shame
most newer things don't have PCI in favor of PCIe.

You might want to simplify your .asoundrc, it's more likely to get in
the way than help these days.

#-
defaults.ctl.card 1
defaults.pcm.card 1
defaults.pcm.device 0
#-

Assuming that it didn't get index 0 in /proc/asound/cards.  But did
get 1.  In alsa speak that's equivalent to -Dhw:1,0 .  Although you
might want to omit the ,0 since that's typcially playback, not
capture, so -Dhw:1

- James


On 12/4/16, Ralf Mardorf  wrote:
> On Sun, 4 Dec 2016 20:18:24 +, zcx wrote:
>>I have a Delta 44 sound card here that uses the ice1712 chipset.
>>
>>Am I right in thinking that although the card has 4 mono inputs, it
>>can only capture one stream at a time?  arecord seems to think so...
>
> Only one app can grab the device, if you run two instances of the same
> app, only one instance can grab the device.
>
> If several apps should be able to use the device at the same time, you
> need a workaround, e.g. dmix or e.g. a sound server, such as e.g. jackd.
>
> --
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> ___
> Alsa-user mailing list
> Alsa-user@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/alsa-user
>

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ice1712 recording

[Ralf Mardorf]

On Sun, 4 Dec 2016 20:18:24 +, zcx wrote:
>I have a Delta 44 sound card here that uses the ice1712 chipset.
>
>Am I right in thinking that although the card has 4 mono inputs, it
>can only capture one stream at a time?  arecord seems to think so...

Only one app can grab the device, if you run two instances of the same
app, only one instance can grab the device.

If several apps should be able to use the device at the same time, you
need a workaround, e.g. dmix or e.g. a sound server, such as e.g. jackd.

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

[Alsa-user] ice1712 recording

[zcx]

I have a Delta 44 sound card here that uses the ice1712 chipset.

Am I right in thinking that although the card has 4 mono inputs, it can 
only capture one stream at a time?  arecord seems to think so...

$ arecord -l
 List of CAPTURE Hardware Devices 
card 0: M44 [M Audio Delta 44], device 0: ICE1712 multi [ICE1712 multi]
   Subdevices: 1/1
   Subdevice #0: subdevice #0

I have an alsa application that needs to capture two audio signals 
concurrently.  Both are to be sampled at 48000 sps.  The two resulting 
data streams must be kept completely separate for the application to work.

I am trying to avoid using another sound card.

So I tried this in .asoundrc...

pcm.ice1712 {
typehw
card0
device  0
channels12
}

pcm.D44_ch12_capture {
typeplug
ttable.0.0 1
ttable.1.1 1
slave.pcm   ice1712
}

pcm.D44_ch34_capture {
typeplug
ttable.0.2 1
ttable.1.3 1
slave.pcm   ice1712
}

If I open a terminal both the capture PCMs work fine on their own...

  $ arecord -f dat -r 48000 -DD44_ch12_capture test12.wav
Recording WAVE 'test12.wav' : Signed 16 bit Little Endian, Rate 48000 
Hz, Stereo
^C

$ arecord -f dat -r 48000 -DD44_ch34_capture test34.wav
Recording WAVE 'test34.wav' : Signed 16 bit Little Endian, Rate 48000 
Hz, Stereo
^C

But if I then open another terminal and try to use the capture PCMs 
concurrently I get

$ arecord -f dat -r 48000 -DD44_ch34_capture test34.wav
arecord: main:722: audio open error: Device or resource busy

I suspect this is a limitation of the ice1712 hardware.

Can anyone shed any light on this please?

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Alsa-user mailing list
Alsa-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ICE1712: Recording from line-in with Audacity

[Klaus-Peter Schrage]

Frank Barknecht wrote:
This is from the Wiki at http://audacityteam.org/wiki/index.pl?LinuxIssues 

   Audacity can use ALSA natively if you compile it to use
  PortAudio v19 instead of the default, v18. Note that v19 is
  still evolving and not as well tested, but it is the only
  way that Audacity supports ALSA. Many users have reported
  good experiences with it. To use it, reconfigure Audacity
  like this: 

  ./configure --with-portaudio=v19 --without-portmixer


Now, is there any tricky mapping of channels, probably with .asoundrc, 
so that the line-in signal shows up in ch. 1 and 2?


Not it pure OSS probably, but you could use libaoss to let
Audacity-OSS talk to ALSA devices without knowing it does so. 

But actually I record radio a lot with ecasound, which has the
advantage, that it can be put into a cronjob and thus can even record,
when I'm away/asleep.
Thanks, Frank, for your suggestions. I didn't dive into ecasound yet, 
and I don't catch the logics behind aoss, so I tried to compile 
audacity-1.2 with the portaudio-v19 option:

Hard stuff, I got the same compile errors mentioned on the wiki page, 
and I didn't succeed until I copied an original portaudio snapshot from 
their homepage INTO the audacity source.

Now I have two i/o devices in audacity: /dev/dsp again and TerraTec 
DMX6Fire, which, until now, seems to be unusable.

I'll keep on trying ...

Klaus

---
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470alloc_id=3638op=click
___
Alsa-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ICE1712: Recording from line-in with Audacity

[Frank Barknecht]

Hallo,
Klaus-Peter Schrage hat gesagt: // Klaus-Peter Schrage wrote:

 Thanks, Frank, for your suggestions. I didn't dive into ecasound yet, 
 and I don't catch the logics behind aoss, so I tried to compile 
 audacity-1.2 with the portaudio-v19 option:
[...] 
 Now I have two i/o devices in audacity: /dev/dsp again and TerraTec 
 DMX6Fire, which, until now, seems to be unusable.

Hmm, I see trouble coming. I will need to try this as well in the next
time, but won't be able to do so this week. I'd be interested in your
progress though.

Regarding ecasound, I attached a little shell script that we use for
about four years now at my work (DLF) to record various radio shows
for later use like transcribing interviews and stuff. You'd call it
from a cronjob like this: 

# Forschung Aktuell aufnehmen:
31 16 * * 1-5   /usr/local/bin/auto_aufnahme.sh forschung   1740 mp3

This will record for about half an hour (1740 seconds) into a
subdirectory called forschung directly to format mp3 every weekday
starting at 16:31 o'clock. 

The script is using various ecasound options to ensure 
click-free recordings even with a non-patched kernel by
double-buffering and realtime mode. 

As I said, this is working like that for years now.

ciao
-- 
 Frank Barknecht   _ __footils.org__


auto-aufnahme.sh
Description: Bourne shell script

Re: [Alsa-user] ICE1712: Recording from line-in with Audacity

[Frank Barknecht]

Hallo,
Andreas von Heydwolff hat gesagt: // Andreas von Heydwolff wrote:

 I can recommend also KHdRecord at http://software.jodda.de/record.html. 
 The timer is configurable from within the program.

Did you write a postcard? Quoting the site: 
  
  Most people don't like this program. I have downloads of 100 source
  packages per week, but I got only 3 postcards at all. So what is the
  reason, that you don't like it?

I like this attitude... ;) 

I didn't knew about KHdRecord, but I see one problem. It seems to only
use OSS as well, which can be a problem especially on ICE1712 chips.

For me, nothing beats ecasound in regard to reliably and quick
recording, if you can live with the command line.  And even one cannot
live with a command line there's help in the form of tkeca:
http://tkeca.sourceforge.net/

ciao
-- 
 Frank Barknecht   _ __footils.org__


---
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470alloc_id=3638op=click
___
Alsa-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/alsa-user

Re: [Alsa-user] ICE1712: Recording from line-in with Audacity

[Andreas von Heydwolff]

Frank Barknecht wrote:
Hallo,
Andreas von Heydwolff hat gesagt: // Andreas von Heydwolff wrote:

I can recommend also KHdRecord at http://software.jodda.de/record.html. 
The timer is configurable from within the program.


Did you write a postcard?
yes I did :-)

I like this attitude... ;) 
me too.

I didn't knew about KHdRecord, but I see one problem. It seems to only
use OSS as well, which can be a problem especially on ICE1712 chips.
I am using it with a Delta 44 card, the results are satisfying for me.

For me, nothing beats ecasound in regard to reliably and quick
recording, if you can live with the command line.  And even one cannot
live with a command line there's help in the form of tkeca:
http://tkeca.sourceforge.net/
Never used ecasound yet, will also give it a try. Thanks for the link 
and shellscript, will look into it.

---
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470alloc_id=3638op=click
___
Alsa-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/alsa-user

[Alsa-user] ICE1712: Recording from line-in with Audacity

[Klaus-Peter Schrage]

I have a Terratec DMX6Fire 2496 card, and I'd like to record radio 
broadcasts through line-in which means input channels 3  4 on my card.
Using arecord, I can type for example
	arecord -D line_in -d 20 -f cd tst.wav
with line_in defined by the following clip in .asoundrc:

#adcdac 3 and 4
pcm.line_in {
type plug
ttable.0.2 1
ttable.1.3 1
slave.pcm ice1712
}
I'd like to record through Audacity, which I find very convenient, but 
it operates in OSS emulation mode and the only i/o device accessable 
there is /dev/dsp. To capture line-in I have to open 4 channels with the 
line-in signal going to ch. 3 and 4, ch 1 and remaing silent.

Now, is there any tricky mapping of channels, probably with .asoundrc, 
so that the line-in signal shows up in ch. 1 and 2?

TIA
Klaus
---
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470alloc_id=3638op=click
___
Alsa-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/alsa-user