Opened 11 years ago

Closed 9 years ago

Last modified 9 years ago

#3405 closed enhancement (fixed)

ALSA input for analog capture

Reported by: anonymous Owned by: Janne Grunau
Priority: minor Milestone: 0.22
Component: mythtv Version: head
Severity: medium Keywords:
Cc: Ticket locked: no

Description

Currently, mythtv only supports OSS for analog capture. Since ALSA is the main audio system for Linux (2.6.0+), mythtv should alsa support ALSA capture.

Attachments (8)

mythtv_alsa.patch (28.9 KB) - added by amistry@… 11 years ago.
Patch for alsa capture
mythtv_alsa-2.patch (29.0 KB) - added by amistry@… 11 years ago.
Further adhere to coding standards
3405-v1.patch (33.0 KB) - added by danielk 11 years ago.
Cleans up error handling, moves from libmythtv to libmyth so mythphone can use it evenually.
3405-v2.patch (33.0 KB) - added by danielk 11 years ago.
This just updates the patch for the current trunk. It still needs to handle error conditions better.
alsa_input_3405-v3-trunk-18043.patch.gz (8.3 KB) - added by Alan Calvert <cal@…> 10 years ago.
3405-v4.patch (41.0 KB) - added by Alan Calvert <cal@…> 10 years ago.
@ 18856
3405-v5.patch.gz (8.7 KB) - added by Alan Calvert <cal@…> 10 years ago.
@ 19944
3405-v6.patch.gz (8.5 KB) - added by Alan Calvert <cal@…> 9 years ago.
refresh @ 20846, lightly tested.

Download all attachments as: .zip

Change History (25)

Changed 11 years ago by amistry@…

Attachment: mythtv_alsa.patch added

Patch for alsa capture

Changed 11 years ago by amistry@…

Attachment: mythtv_alsa-2.patch added

Further adhere to coding standards

comment:1 Changed 11 years ago by Stuart Auchterlonie

Milestone: unknown0.21
Version: unknownhead

OSSAudioInput::OpenDevice?() is broken.

If the device open fails it throws a verbose error but then continues on it's merry way.

It repeatedly checks if (0 <= retval) and while it is reusing retval it just looks plain wrong.

Stuart

comment:2 Changed 11 years ago by amistry@…

Stuart: You're right, it is broken, but only with respect to logging info. The point of the repeated checking is to give the function a single return point at the end of the function. This way, I only have to do any potential cleanup at one point, at the end. All the 'if (0 <= retval)' is there to make sure further code only executes if everything up until that point is OK. If it's not, it'll be skipped and a return code of < 0 (usually -1) will signify an error. You'll notice all the functions I've wrote are like that inclusing the alsa input. If for some reason this is unacceptable, I can rework the patch to put returns after each error instead of just at the end.

comment:3 Changed 11 years ago by danielk

Owner: changed from Isaac Richards to danielk
Type: patchenhancement

Changed 11 years ago by danielk

Attachment: 3405-v1.patch added

Cleans up error handling, moves from libmythtv to libmyth so mythphone can use it evenually.

comment:4 Changed 11 years ago by danielk

Milestone: 0.210.22

Changed 11 years ago by danielk

Attachment: 3405-v2.patch added

This just updates the patch for the current trunk. It still needs to handle error conditions better.

comment:5 Changed 10 years ago by Alan Calvert <cal@…>

Presenting patch alsa_input_3405-v3-trunk-18043:
I've use this patch on three systems for over a year, so I'm interested in it's welfare. This little effort started as a refresh to current trunk, but it sort of morphed beyond that. Perhaps I've addressed some of danielk's concerns over error handling. Subjectively, I believe it holds synch better than anything else I've tried. Objectively, it seems to work ok, and does pretty much all the things it needs to do to perform respectably. Enjoy.

Changed 10 years ago by Alan Calvert <cal@…>

Changed 10 years ago by Alan Calvert <cal@…>

Attachment: 3405-v4.patch added

@ 18856

comment:6 Changed 10 years ago by Alan Calvert <cal@…>

v3 was pretty rough, v4 is a bit nicer.

comment:7 Changed 10 years ago by Dibblah

Status: newassigned

comment:8 Changed 10 years ago by Janne Grunau

Owner: changed from danielk to Janne Grunau
Status: assignedaccepted

Changed 10 years ago by Alan Calvert <cal@…>

Attachment: 3405-v5.patch.gz added

@ 19944

comment:9 Changed 10 years ago by Alan Calvert <cal@…>

3405-v5 refresh to 19944, some tweaks & tidies, and reinstate alsa device selection (it fell off).

comment:10 in reply to:  description Changed 9 years ago by anonymous

Any update on this?

Kernels built without OSS support are becoming more common (like in Fedora 11) so it's an actual problem now.

comment:11 Changed 9 years ago by anonymous

And software-based analogue capture cards are becoming less common, antiquated tech ...

Actually that's an understatement, they were surpassed by hardware capture cards years ago.

comment:12 Changed 9 years ago by anonymous

Maybe so, but I have such a card and I'm not afraid to use it. ;)

I'm compiling svn+patch now, to test it.

comment:13 Changed 9 years ago by anonymous

Been using the patch for a while now and it seems to work.

Changed 9 years ago by Alan Calvert <cal@…>

Attachment: 3405-v6.patch.gz added

refresh @ 20846, lightly tested.

comment:14 Changed 9 years ago by Janne Grunau

Resolution: fixed
Status: acceptedclosed

(In [21448]) adds alsa audio input to analog cards. Fixes #3405

factors the audio device handling code out of NuppelVideoRecorder? into a AudioInput? class with OSS and new ALSA implementation.

based on patches by: Alan Calvert, Anand K. Mistry and Daniel Kristjansson

comment:15 Changed 9 years ago by Janne Grunau

(In [21457]) fix windows build after [21448], Refs #3405

comment:16 Changed 9 years ago by Janne Grunau

(In [21471]) link libmythtv against alsa/oss libs for [21448]. Refs #3405

comment:17 Changed 9 years ago by Janne Grunau

(In [21526]) fix strerror handling in audioinputoss verbose messages. Refs #3405

Note: See TracTickets for help on using tickets.