gmane.comp.audio.audacity.devel

Re: Further problem, obvious in Normalize

Martyn Shaw — 2012-05-17T00:22:02

Well, I had a bit more of a go, and committed it.  Anticipating some 
of the strings will not be up to scratch, but I hope the logic is.

I think I have had enough of 'Normalize' for now.

Thanks for all the input!

TTFN
Martyn

On 07/05/2012 10:35, Peter Sampson wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Email glitches

Vaughan Johnson — 2012-05-12T13:42:34

I switched ISP a few weeks ago. Doing so made it easier to switch to fully hosted vaughan< at >audacityteam.org, so I did that about a week ago. For about an hour I failsafed it via forwards to my gmail account. Somehow that messed up Dreamhost such that it kept forwarding some, didn't send some to the fully hosted pop3, apparently completely lost some (both from and to the fully hosted).

I think it's fixed now but if you were expecting a response from me, please retransmit. 

- Vaughan
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Patch for bug 50 – calculation of remaining space for 24-bit recording

Miroslav Matějů — 2012-05-12T11:13:06

Hello there,
I am sending a patch for the bug 50 in Bugzilla (http://bugzilla.audacityteam.org/show_bug.cgi?id=50). It corrects the remaining recording time available when using 24-bit samples.
* The remaining time calculation was incorrect because there are 32-bit float samples recorded when 24-bit samples selected, and then only 24-bit samples saved. (A one-minute-long test file takes 5 MB for 16-bit samples, 7.5 MB for 24-bit and 10 MB for 32-bit samples.)
* The variable captureFormat is converted by the following code, so the time calculation was not able to see the actual size of samples on disk.
   // Special case: Our 24-bit sample format is different from PortAudio's
   // 3-byte packed format. So just make PortAudio return float samples,
   // since we need float values anyway to apply the gain.
   if (captureFormat == int24Sample)
      captureFormat = floatSample;

* I created a new variable mSaveFormat which keeps the original captureFormat and it is passed to time calculator using the new functi

Can this be committed ?

Steve the Fiddle — 2012-05-11T20:02:59

A proposal page was created on the wiki last July for adding Nyquist
support in Chains.
There are many compelling reasons why this would be a major
enhancement for Audacity.
Even better, there is already a patch that provides this functionality
and it has been tested on Windows Linux and Mac.

http://wiki.audacityteam.org/wiki/Proposal_Nyquist_process_effects_in_Chains

Sadly this enhancement went into a coma on September 11th 2011 and has
not moved since.
Would anyone be interested in reviving it?

Steve

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Patch: fix for incorrect RIFF WAVE headers when exporting to external program

Gale (Audacity Team — 2012-05-11T06:57:55

Exporting a tone to WAV using (external program) and FFmpeg then
looking in a hex editor, I found the ChunkSize content at offset 4 
identical pre-and post-Martyn's commit, and identical to that when
exporting via libsndfile from File > Export. Maybe that's the wrong test,
but anyway...  


I tested in Audacity with your qtaacenc and got a failure
"QTSetComponentProperty error: cannot set input ASBD" pre-patch 
but a good export post-patch. 

Also tested with Neroaac encoder and can confirm that pre-patch 
I get "ERROR: could not parse WAV file" unless I add -ignorelength,
but post-patch I don't need to add that. 

I also tested on a whim if this helped bug 165:
http://bugzilla.audacityteam.org/show_bug.cgi?id=165 

but it doesn't. 



Gale 


--
View this message in context: http://audacity.238276.n2.nabble.com/Patch-fix-for-incorrect-RIFF-WAVE-headers-when-exporting-to-external-program-tp7535906p7549437.html
Sent from the audacity-devel mailing list archive at Nabble.com.

--------------------------------

Re: Patch: fix for incorrect RIFF WAVE headers when exporting to external program

Martyn Shaw — 2012-05-10T22:41:56

Hi

Yes, OK, I had a look at various sites tonight and it certainly 
appears that somebody added it up incorrectly.

I also checked the specs for the other bit, and we seem to have it 
wrong so I committed your changes.  I'll leave it to others to test.

TTFN
Martyn

On 10/05/2012 09:56, Taihei Monma wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Patch: fix for incorrect RIFF WAVE headerswhen exporting to external program

Taihei Monma — 2012-05-10T08:56:01

Hi Martyn,


This is different issue from the blockAlign problem. lenAfterRiff should be summation of
  sizeof(header.riffType)       [4 bytes]
  sizeof(header.fmtID)          [4 bytes]
  sizeof(header.formatChunkLen) [4 bytes]
  sizeof(header.formatTag)      [2 bytes]
  sizeof(header.channels)       [2 bytes]
  sizeof(header.sampleRate)     [4 bytes]
  sizeof(header.avgBytesPerSec) [4 bytes]
  sizeof(header.blockAlign)     [2 bytes]
  sizeof(header.bitsPerSample)  [2 bytes]
  sizeof(header.dataID)         [4 bytes]
  sizeof(header.dataLen)        [4 bytes]
and sampleBytes. So This should be sampleBytes + 36, not 32.

See the riff wave structure described here for details:
https://ccrma.stanford.edu/courses/422/projects/WaveFormat/
lenAfterRiff is equivalent to ChunkSize in the page.

I don't know how many programs are affected by this problem (this field tends to be ignored as the blockAlign field), but at least it doesn't work with neroAacEnc without -ignorelength switch due to the problem.

Thanks,
Taihei

Re: Can't zoom in on end of selection with Ctrl+Mouse Wheel (svn reviion 11726)

Reid Zak — 2012-05-10T03:38:41

Haha, it is a good rule, however I can think of other audio programs (ok,
just one other comes to mind right now) that do move the user's cursor for
some things (mostly to achieve "Wrap-around", where the user clicks and
moves the cursor off to the left side of some widget and it wraps around
back to the right, effectively letting them slide something infinitely on
one axis)
The software is VirtualDJ ( http://www.virtualdj.com/ ) and you can
download the home version for free.
I think the use case is similar in this instance (at least the way I
envision it), as the program never moves the user's cursor unless the user
is actively using the mouse (I.E. The user is holding control and moving
the scroll-wheel backwards, which simultaneously zooms out and shifts the
cursor a few tens of pixels to the left, and vice versa when zooming back
in)
But I do understand how this might violate some user interface rules, and
understand that it may not be a good idea to implement it.


Thanks!
--Reid


On Wed, May 9, 2012

Re: Can't zoom in on end of selection withCtrl+Mouse Wheel (svn reviion 11726)

Bill Wharrie — 2012-05-10T00:26:29

On 09/05/2012, at 6:18 PM, Martyn Shaw wrote:


Amen!

Re: Patch: fix for incorrect RIFF WAVE headers when exporting to external program

Martyn Shaw — 2012-05-09T23:41:05

Hi Taihei

This fix looks good, and is the same (essentially) as the one on the 
forum post for the blockAlign.

I don't understand the

-   header.lenAfterRiff     = wxUINT32_SWAP_ON_BE(sampleBytes + 32);
+   header.lenAfterRiff     = wxUINT32_SWAP_ON_BE(sampleBytes + 36);

bit, are we 4 bytes out?  If so, why has that ever worked?

Some explanation and I'll commit your patch in a few days, since it 
looks OK at a quick look, unless anybody says otherwise.

TTFN
Martyn

On 09/05/2012 09:26, Taihei Monma wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Can't zoom in on end of selection with Ctrl+Mouse Wheel (svn reviion 11726)

Martyn Shaw — 2012-05-09T22:18:52

Reid

There is a rule
"thou shalt not move the mouse pointer programatically"
in the Apple HIG, since it can confuse the user.  I'm not sure we 
would like different behaviour on different platforms, and some would 
say the HIG are worth following.

See http://docs.wxwidgets.org/stable/wx_wxwindow.html#wxwindowwarppointer

However, if you wanted to make a patch for us to try, I would try it.

Thanks
Martyn

On 09/05/2012 00:25, Reid Zak wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Patch: fix for incorrect RIFF WAVE headerswhen exporting to external program

Taihei Monma — 2012-05-09T08:26:30

Suggested workaround there is:

header.blockAlign = wxUINT16_SWAP_ON_BE(SAMPLE_SIZE(int16Sample) * header.channels);

This works the same as my patch (SAMPLE_SIZE(int16Sample) returns 2 [bytes]). The point is that the blockAlign field should store size in bytes, not in bits. See the specification from MS:
http://msdn.microsoft.com/en-us/library/ms713497%28v=vs.85%29.aspx


I'm a developer of qtaacenc, a CLI frontend of QuickTime AAC encoder. One of the users reported me that it doesn't work with Audacity, and I found a problem in Audacity. In most cases the blockAlign field is just ignored, so at the moment I cannot find other tools affected by the problem.

Thanks,
Taihei
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. ht

Re: Can't zoom in on end of selection with Ctrl+Mouse Wheel (svn reviion 11726)

Reid Zak — 2012-05-08T23:25:02

Hello!

Good thoughts all!

Here's my two cents:
What if, after the user zooms out so far that the mouse cursor is not over
the point in the waveform that it was when the user started zooming out,
that Audacity just moves the mouse cursor to that original spot in the
waveform where it used to be? (As the user zooms in and out)

Example:
1. Create a 30 second tone
2. Click the waveform at the 12 second mark.
3. Zoom in using Ctrl+Scroll
4. Zoom out until you can see the 4 minute mark on the right of the
timelinee
5. Notice your mouse cursor is not over the original spot you were zooming
to. In the above proposed system, Audacity would instead move the mouse
cursor to where the selection cursor is as the user zooms out. Then, when
the user zooms back in, the mouse cursor would return to where it was
before, to maintain being over the zoom-spot (where the selection cursor
happens to be in this example)

The benefits are:
1. The user could zoom in and out all they want and still be able to zoom
back to the same

Re: Patch: fix for incorrect RIFF WAVE headers when exporting to external program

Gale (Audacity Team — 2012-05-08T22:29:27

Thanks, Taihei. Someone else mentioned this on the Audacity Forum 
a while ago and had a slightly different solution for 

header.blockAlign       = wxUINT16_SWAP_ON_BE(header.bitsPerSample *
header.channels);

Scroll down a bit from this link to see the discussion:
http://forum.audacityteam.org/viewtopic.php?f=15&t=25231#p174669 .

Sorry I can't comment on the technicalities, but what application 
are you using to do the encoding (FFmpeg, libsndfile...), and does that 
make any difference? 



Gale 



--
View this message in context: http://audacity.238276.n2.nabble.com/Patch-fix-for-incorrect-RIFF-WAVE-headers-when-exporting-to-external-program-tp7535906p7540887.html
Sent from the audacity-devel mailing list archive at Nabble.com.

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint s

Re: Can't zoom in on end of selection withCtrl+Mouse Wheel (svn reviion 11726)

Bill Wharrie — 2012-05-08T15:37:43

Steve et al:
I will leave it to you "wheel-zoomers" to sort this out. I haven't used wheel zooming up to now because of the problems in the previous implementation and the fact that on my Mac it is way too easy to zoom out too far; it takes a fine touch on the "Magic Mouse" to do a controlled zoom.

Re: Can't zoom in on end of selection with Ctrl+Mouse Wheel (svn reviion 11726)

Steve the Fiddle — 2012-05-08T10:23:50


I don't think that is viable.
Example, generate a 30 tone, then zoom out and generate another 30
second tone at the 1 minute mark.
This example fails because, from the default "zoom to fit" position,
wheel zoom would appear to be broken (any amount of "zoom out" from
any pointer position would necessitate negative time values).

It is often useful to zoom out beyond the end of the audio (for
example to paste a new audio clip) and it would be inconvenient if the
user was forced to switch from wheel zoom to button zoom as soon as
"zoom to fit" was reached.


I agree that this is a little disconcerting but I suspect there would
be other disconcerting (possibly worse) features of any solution.
Somewhere along the line it is necessary to "break the rules". I think
that the only "logically correct" solution is the one proposed by
Martyn (show negative time values) but this breaks the Audacity rule
of "active time" (viewable/playable) time starts at zero. Although I
have not (yet) tried Martyn's proposal I suspec

Re: Can't zoom in on end of selection withCtrl+Mouse Wheel (svn reviion 11726)

Bill Wharrie — 2012-05-07T21:26:51

On 07/05/2012, at 5:05 PM, Gale (Audacity Team) wrote:


I'd prefer that zooming out stop when the graph is entirely to the left of the mouse pointer position. That avoids most situations where the graph can collapse to a tiny region at the start of the track (unless the user is zooming out e.g. at the 30-second mark on a one hour track).


The big advantage I see in Martyn's solution is that the mouse pointer remains stuck at whatever time is was hovering over. The disadvantage when zooming out is that this would eventually necessitate having the Timeline start at a negative value. Would it be possible to stop zooming out one step before you need to display negative time?

The most disconcerting part of the current implementation is that you can't get back where you came from after you zoom out too far. The instinct is to leave the mouse pointer where it is and zoom back in, but you eventually end up with the end of the graph under your mouse pointer.

Re: Can't zoom in on end of selection with Ctrl+Mouse Wheel (svn reviion 11726)

Gale (Audacity Team — 2012-05-07T21:05:29

Nonetheless there is an inconsistency here that is a bit confusing. You 
can zoom in where the mouse pointer is in white space before the audio
starts, but something else happens when you zoom in where the pointer
is after the end of the audio. Also if you zoom out where the pointer 
is after the end of the audio, the behaviour is different to when you 
zoom in. 

Is some fix not possible where we can do what Martyn wants, but at 
least have scrolling in the opposite direction get you the graph back? 

Thanks for the improvements so far, though since most zooming I do
is actually to the centre of the selection, I liked using wheel instead 
of buttons for that.  CTRL + SHIFT + Wheel?   




Gale 







--
View this message in context: http://audacity.238276.n2.nabble.com/Can-t-zoom-in-on-end-of-selection-with-Ctrl-Mouse-Wheel-svn-reviion-11726-tp7528411p7536944.html
Sent from the audacity-devel mailing list archive at Nabble.com.

------------------------------------------------------------------------------

Re: Can't zoom in on end of selection with Ctrl+Mouse Wheel (svn reviion 11726)

Steve the Fiddle — 2012-05-07T17:24:05

Thanks James. Testing with Revision: 11734 I'm happy again :)

I don't know if what we have now is ultimately "the best" behaviour
but I agree that what we have now is a lot better than 2.0 and
resolves the usability issues that I raised with revision 11726.
Personally I'm happy with how Ctrl+Mouse Wheel now works, but if
anyone feels that it is worth trying other possibilities then I'll be
happy to give them a test drive.

Steve


On 7 May 2012 12:51, James Crook <crookj< at >indigo.ie> wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Patch: fix for incorrect RIFF WAVE headers whenexporting to external program

Taihei Monma — 2012-05-07T15:55:09

Hi folks,

When exporting to external programs, Audacity generates a wave file including incorrect header values. Here is a patch to fix the problem.

Regards,
Taihei

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________
audacity-devel mailing list
audacity-devel< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/audacity-devel

Re: Can't zoom in on end of selection with Ctrl+Mouse Wheel (svn reviion 11726)

James Crook — 2012-05-07T11:51:23

Martyn,

I disagree with you.

If you allow scrolling to time before zero, inexperienced users will 
much more easily lose the graph.  Losing the graph is the problem.  An 
experienced user would think to use zoom-to-fit but a newbie would not.  
A user might zoom in beyond the end of the audio, move the mouse 
slightly right and zoom out again, and be way off to the right.  Zoomed 
out a long way the graph can be too tiny to notice if it's only a minute 
or so of audio.  If you think allowing time before zero is an easy fix, 
try it, and see what the user experience is like.

I'm not going to let you be a backseat driver on this.  My fix is better 
than what we had.  If you think your proposed solution is better still, 
then you can disable or completely remove my code in TrackPanel that reads:

       double audioEndTime = mTracks->GetEndTime();

       // When zooming in in empty space, it's easy to 'lose' the waveform.
       // This prevents it.
       // IF zooming in
       if( steps > 0)
       {