headers
Kai Vehmanen | 1 Jun 2005 04:43

Re: ULAW RAW Output

Hi,

On Sun, 29 May 2005, Dan Weber wrote:

> Do you understand that libsndfile is unable to write to fifo buffers?

hmm, that's a good point. So I guess then integrating your library 
(assuming its GPL), or taking the code from libsndfile/libasound/sox/etc 
and creating a new audio-i/o type, would be a better approach. I cannot 
promise when I will have time to look at this, so I leave the choice of 
whether to send me the ulaw/alaw code, to you. Anyone else interested in 
working on this?

I guess even better would be to extend the -f:x,x,x,x audio parameter 
mechanism to support non-PCM sample formats. But this is such a big 
change, that it won't happen in the near future (unless someone else
than me does it).

BTW; can you use sox to do this stuff? Pipe from ecasound -> sox ->
      pipe, or the other way around...?

--

-- 
  http://www.eca.cx
  Audio software for Linux!

-------------------------------------------------------
This SF.Net email is sponsored by Yahoo.
Introducing Yahoo! Search Developer Network - Create apps using Yahoo!
Search APIs Find out how you can build Yahoo! directly into your own
Applications - visit http://developer.yahoo.net/?fr=offad-ysdn-ostg-q22005
(Continue reading)

Permalink | Reply | 0 comments |
headers
Free Lunch | 2 Jun 2005 19:05
Picon

Re: W64 format conversion correctness

Thank you for your response, Kai.

On 5/29/05, Kai Vehmanen <kvehmanen <at> eca.cx> wrote:
> On Fri, 20 May 2005, Free Lunch wrote:
> 
> Before processing, ecasound converts all input data to its internal format
> which is 32bit float atm.

Just to clarify, is that true for cases when recording from alsa with
ecasound?  Does data coming in at 24 bits get converted to 32b and
back to 24b before being written to disk?

> Neither ecasound or libsndfile supports dithering (although it is on
> libsndfile's todo-file and will be utilized by ecasound also when ready),
> so there can be some artifacts from the float->fixed down-conversion.

While dithering would be an improvement, it is a complicated subject
and there are so Many ways to do dithering..

For my audio recording purposes, any conversion is bad.  The streams
must be bit-accurate so that the integrity of the audio chain can be
verified.

> Now, but, but... if you don't perform any processing, just plain copy,
> there shouldn't be any artifacts as the result should be
> fixed(A)->float(B)->fixed(A). It's possible that some rounding/truncation
> errors are created by differences in doing the fixed<->float conversion
> between ecasound and libsndfile. You can work around this by using
> libsndfile for all conversions:
>
(Continue reading)

Permalink | Reply | 1 comment |
headers
Stuart Allie | 3 Jun 2005 02:54
Picon
Favicon

instant gui for ecasound

Hi List,

I had an idea the other day for a quick-and-dirty gui for ecasound. I
used qmidicontrol - a qt-based midi slider box - to control the level
and pan setting for a bunch of chains.

I used the following steps to make this happen:

1. Install a "virtual midi" device. Say this is rawmidi device 1 (0 is
my soundcard).

2. Run qmidicontrol with 8 controls, output on channel 0 (channel 1in
ecasound).

3. Connect qmidicontrol to the virtual midi device (with aconnect, or
alsa-patch-bay for example).

4. Start ecasound with -Md:rawmidi,/dev/snd/midiC1D0 since the virtual
midi device is number 1.

In the chain setup, have chain operators something like this:

-a:kick  -ea:100 -km:1,0,200,1,1 -epp:50 -km:1,0,100,2,1
-a:snare -ea:100 -km:1,0,200,3,1 -epp:50 -km:1,0,100,4,1
-a:toms  -ea:100 -km:1,0,200,5,1 -epp:50 -km:1,0,100,6,1
-a:hihat -ea:100 -km:1,0,200,7,1 -epp:50 -km:1,0,100,8,1

Now the first two sliders control the level and pan of the "kick" chain,
and so on. I played with this last night and it worked quite well.
(Continue reading)

Permalink | Reply | 1 comment |
headers
Kai Vehmanen | 9 Jun 2005 23:44

Re: W64 format conversion correctness

Hello,

On Thu, 2 Jun 2005, Free Lunch wrote:
>> Before processing, ecasound converts all input data to its internal format
>> which is 32bit float atm.
> Just to clarify, is that true for cases when recording from alsa with
> ecasound?  Does data coming in at 24 bits get converted to 32b and
> back to 24b before being written to disk?

yes, it applies to all ecasound use-scenarios (data flow inside ecasound
is based on passing around buffers of float samples).

Adding special-case handling for the "fixed-A -> float -> fixed-B" case, 
where A and B are the same, has been considered in the past, but so far 
not actually implemented. The needed change is not a major one, but 
one needs to be _very_ careful so that other uses of ecasound are not 
affected...

>> Neither ecasound or libsndfile supports dithering (although it is on
>> libsndfile's todo-file and will be utilized by ecasound also when ready),
> While dithering would be an improvement, it is a complicated subject
> and there are so Many ways to do dithering..

Yes, and that's why I'd prefer to utilize common implementations: JACK 
implements various dithering algorithms for audio device i/o, libsndfile 
will hopefully do the same for file i/o. Those are the only places where 
we must convert from/to fixed-integer sample presentation.

> For my audio recording purposes, any conversion is bad.  The streams
> must be bit-accurate so that the integrity of the audio chain can be
(Continue reading)

Permalink | Reply | 0 comments |
headers
Louis-David Mitterrand | 10 Jun 2005 16:12

triming silence in sound files

Hello,

I have a bunch of .wav files I have recorded for our (open-source) phone
system's voicemail.

How can I use ecasoud to remove (quasi) silence at the start and end of
these files?

Thanks,

--

-- 
This is Linux Country. On a quiet night you can hear Windows NT reboot.

-------------------------------------------------------
This SF.Net email is sponsored by: NEC IT Guy Games.  How far can you shotput
a projector? How fast can you ride your desk chair down the office luge track?
If you want to score the big prize, get to know the little guy.  
Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20
_______________________________________________
Ecasound-list mailing list
Ecasound-list <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ecasound-list
Permalink | Reply | 4 comments |
headers
Willem van Engen | 11 Jun 2005 15:39
Picon
Favicon

live silence splitting / ringbuffer

Hello,

I'm writing a program that records audio when there is activity at the 
mic, and stops when it's silent again. Every piece of audio gets written 
into a new file. Using pyecasound, I wrote a little gtk gui for that. 
Well, it works, but there's one problem left:
I'm monitoring for sound on the input; after several seconds of sound, I 
want to record it and keep recording until the next silence. It seems to 
me that some sort of buffer is needed.

Is there a ringbuffer in ecasound? If, I might write a program that just 
reads on stdin and just waits some time before writing it back to 
stdout, and use that as output of a chain and input to another. But it 
feels a bit 'hackish' to me; is there another way?

Regards,
- willem

-------------------------------------------------------
This SF.Net email is sponsored by: NEC IT Guy Games.  How far can you shotput
a projector? How fast can you ride your desk chair down the office luge track?
If you want to score the big prize, get to know the little guy.  
Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20
_______________________________________________
Ecasound-list mailing list
Ecasound-list <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ecasound-list
Permalink | Reply | 0 comments |
headers
Kai Vehmanen | 14 Jun 2005 08:58

Re: triming silence in sound files

On Fri, 10 Jun 2005, Louis-David Mitterrand wrote:

> I have a bunch of .wav files I have recorded for our (open-source) phone
> system's voicemail.
[...]
> How can I use ecasoud to remove (quasi) silence at the start and end of
> these files?

ecasound -i foo.wav -o foo-cropped.wav -ge:0.5,0.5,0

See ecasound(1) for more info about the -ge operator. You probably need to 
tune the first two parameters (0.5% example) to get optimal results.

--
  http://www.eca.cx
  Audio software for Linux!

-------------------------------------------------------
This SF.Net email is sponsored by: NEC IT Guy Games.  How far can you shotput
a projector? How fast can you ride your desk chair down the office luge track?
If you want to score the big prize, get to know the little guy.  
Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20
_______________________________________________
Ecasound-list mailing list
Ecasound-list <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ecasound-list
Permalink | Reply | 3 comments |
headers
Louis-David Mitterrand | 14 Jun 2005 09:13

Re: triming silence in sound files

On Tue, Jun 14, 2005 at 09:58:53AM +0300, Kai Vehmanen wrote:
> On Fri, 10 Jun 2005, Louis-David Mitterrand wrote:
> 
> >I have a bunch of .wav files I have recorded for our (open-source) phone
> >system's voicemail.
> [...]
> >How can I use ecasoud to remove (quasi) silence at the start and end of
> >these files?
> 
> ecasound -i foo.wav -o foo-cropped.wav -ge:0.5,0.5,0
> 
> See ecasound(1) for more info about the -ge operator. You probably need to 
> tune the first two parameters (0.5% example) to get optimal results.

Yes, after looking through the examples I found that operator, and
indeed it does what I am looking for. However if a file has some silence
embedded (a sentence with several words) it will be truncated at the
first silence, which is expected behavior. So I had to use "-ge:1,0" and
forego end trimming.

To effectively trim the end silence, one could maybe process the file in
reverse with that same operator? Or some other trick?

Thanks,

--

-- 
If at first you don't succeed, redefine success.

-------------------------------------------------------
This SF.Net email is sponsored by: NEC IT Guy Games.  How far can you shotput
(Continue reading)

Permalink | Reply | 2 comments |
headers
Benoit DELAGARDE | 15 Jun 2005 14:33

ewf file don't work?

Hi all.

I'm using ecasound (v2.4.1) on a debian (recentlly sable) sarge.

I want to insert a short sound (a BIP) in a wav file.
After reading documentation, the only way i fonund to do that is to use
the ewf format. If there is another way, can any one explain me?

But I enconter problem and I don't understant why.
My file mark1.ewf:

-------mark1.ewf--------
source = mark.wav
offset = 10
start-position = 0.0
------------------------

My file mark2.ewf
-------mark2.ewf--------
source = mark.wav
offset = 20
start-position = 0.0
------------------------

"ecasound -i mark1.ewf -o /dev/dsp" give me a sound after 10s and the
quit
"ecasound -i mark2.ewf -o /dev/dsp" just quit after 20s

Does any know why?
(Continue reading)

Permalink | Reply | 1 comment |
headers
Stuart Allie | 16 Jun 2005 03:48
Picon
Favicon

controllers always added to last chain-op?

Hi All,

The ecasound man page says that "All controllers are attached to the
selected (=usually the last specified effect/controller) effect",
however, that doesn't seem to be the case:

Here's the transcript from a sample session.

************************************************************************
********
*                 ecasound v2.4.1 (C) 1997-2005 Kai Vehmanen

************************************************************************
********
- Session created
----------------------------------------------------------
- Chainsetup created (cmdline)
---------------------------------------------
(libkvutils) kvu_rtcaps: warning! POSIX_MEMLOCK not supported
(eca-session) Note! Unable to create a valid chainsetup from the
command-line 
... arguments.
ecasound ('h' for help)> c-add fred
(eca-control-objects) Added chains: fred.
ecasound ('h' for help)> cop-add -ea:123
ecasound ('h' for help)> cop-add -etf:50
ecasound ('h' for help)> cop-select 1
ecasound ('h' for help)> cop-selected
1
ecasound ('h' for help)> ctrl-add -km:1,0,100,1,1
(Continue reading)

Permalink | Reply | 1 comment |

Gmane