Notices

Orecchiette is a simple app designed to record audio data on the N900. Both input and output audio streams can be recorded; so, it can be used to take voice notes, or to record audio from an FM radio app. And, it can combine both streams to record a phone call. And, several video input sources are now also supported.

Orecchiette is open-source software, released under the GPL version 2 (or greater). You can find the source code at Github:

https://github.com/jpietrzak8/Orecchiette

EDIT: Orecchiette 0.3.1 .deb file attached to this post. Also, you can retrieve the latest .deb directly from the package page.

Wow coppy one more useful app !

Is it able to record calls automatic? And it would be great if it can add a record butt with pause resume on call ui

Where is it

We have .deb !

Hmm, it appears that Orecchiette hasn't been made into a package yet. It may take a while; from what I've been reading on the Maemo IRC logs, things have been kind of hectic lately...

Thanks

I'm going to test this the next week.

Great, thanks! However, please do be prepared for bugs; I've only done a minimal amount of testing on the app so far.

There is some confussions, as OGG is indeed a container, but containers with vorbis audio inside have .ogg file extension, and are called "ogg" or "OGG" literally everywhere. Anyway, it's brillant codec by Xiph.org (the same organization, who gave us FLAC), and leading format of royalty-free, patent-free audio codecs. Furthermore, from technical point of view, it's a whole eternity better codec than mp3, and despite being "lossy" (as opposed to FLAC), it received so-called "transparent" codec mark - i.e., in plentora of blind tests, people were not able to distinguish it from lossless version of same recording (hit/miss ratio comparable to flipping a coin). All of that, while maintaining awesome ratio of decreasing file size.

Add to this fact, that it have pretty neat implementation for Maemo, and it seems like codec of choice for recording FM Radio, etc Of course, for conversations (call recording), speex (also by xipg.org) sounds even better.
---

Feature request - above obvious auto-recording of calls (SIP and GSM alike), which you're already considering for sure, it would be nice to have a (remembered) advanced configuration for codecs in use. IMO, most feasible way would be a text field, where user would put a string of text containing command-line parameters for codec, just passed transparently to used encoder (in addition to normal things defined by our application, like save path, etc).

This would save a 43456765434566 number of different option fields in GUI for every possible codec, while still keeping benefits of using GUI for generic options. It's worth to add, that this is behaviour used in most FOSS advanced recording/ripping/encoding programs even in desktop world - like Exact Audio Copy. For example, this is my "additional command line parameters" field in EAC, for creating best possible quality/size ratio (keeping being "transparently compressed" .ogg files of my music CDs collection:

...which doesn't interfere, but adds too, a Q=6 and tag-related settings from GUI. Would be pretty timefconsuming to completely GUI'dise all advanced parameters for all codecs available in Orecch', and keep them updating as new adv. features are added to encoders.

/Estel

Well, I'm not ready to start packaging audio data into containers just yet. But I'll go ahead and see if I can add Vorbis to the list of encodings for Orecchiette.

Well, that would make sense, if I were using the command line to construct the GStreamer pipeline. However, by constructing and managing the plugins directly, Orecchiette has much better control over them. (Thus the ease with which it can pause, stop, or otherwise modify running pipelines -- of course, until the autobuilder comes back, you'll just have to trust me on this one.)

And honestly, I figure if a user knows enough about GStreamer plugins to specify a slew of advanced options, they're probably able to construct a custom pipeline for themselves... But in any case, I'll take a look at EAC and see if I can steal any of their ideas.

My complaint against recaller (other than it sometimes "forgets" to stop recording, which is easily resolved by starting and stopping another call, e.g. to check my remaining credit) is that there is a HUGE disproportion between the volume of my own voice and the other party's. Has this been/can this be addressed in orecchiette?

In fact, it has -- Recaller is using "source.hw0" to record voice data directly from the microphone source. However, after doing some research, I found that there are a number of different "source"s available; I'm currently using "source.voice", a wrapper over the hw0 input which does seem to have the input volume at the same level as the output (which I'm getting from "sink.hw0.monitor"). I haven't done a lot of testing yet (to be honest, I've only recorded one phone call), but it seems to work quite well.

I haven't yet been able to test against a bluetooth headset yet; I'm still using "source.hw1" to retrieve that data, so it'd probably need some work to increase the volume. (Or maybe "source.voice" will automatically switch to hw1?)

Anyway, lots more work to be done on this app.