Smooth Audio Fades?
 

This seems like such a simple basic action, and yet I cannot get FCP to do this correctly.

After being dissatisfied with the way an audio clip faded out when using an audio transition, I decided to try opening my audio clip in the viewer and fading out using key frames. I set the first point where I want the fade to begin, then where I want the fade to end I drag the point down to zero. When I lowered the audio at the last frame, the level line makes an abrupt drop at the end. I was expecting it to make a straight linear transition which could then be smoothed out with a curve. Instead the audio starts to descend smoothly and then drops off a cliff at the last second. I tried to bend the line the other way with the pen tool to make a graceful transition out, but to no avail. The pen tool is not allowing me to curve the line in this way. How can I make the audio fade/curve smoothly out? Even just a straight linear transition would be preferable to whats it's doing. I remember years ago being able to do this in imovie, and it handled it correctly. What am I doing wrong here?

Well, not that I'm a good one to take advice from because I tend to mess with something and if I find a way to make it work, and it's not something I have to do very much, I just limp through it the way I can get to work. That said, the way I do this is to use more key frames and make the curve straight that way. At least you get more control on how it fades with more key frames.

You technically can't have a linear audio fade. Why? Because audio is logarithmic. Any editing software that displays it as linear is lying to you. I've never had an issue with an audio fade being too abrupt. Do you have an example to show us?

Audio is neither logarithmic nor linear, but our ears are (to some degree) sensible to relative levels, which can be expressed by a logarithmic scale. The whole psychoacoustic truth is much more complex, but that's another story.
An editing tool doesn't lie, if it visualizes audio linear or logarithmic, but if its axis' labels don't match the used visualization.
Final Cuts Audio Transitions aren't made for fade-ins/outs, but for cross-fades. If you want a cross-fade, which is monotonic, symmetric and doesn't get louder in the middle, it has to be linear. So it's perfectly ok, if they are linear.
But to get back to the question: If an fade-out/in shall be perceived linearly, it has to be exponential. But it seems to me, that Final Cut only supports linear fades (even with keyframes, also no ease in/out with volume-keyframes possible). Possible workarounds I can figure out besides the one Denise described:
-if you need exponential fades only infrequent: Use a capable external audio editing tool to "bake" these fades into audio-files, which you then drop onto the Final Cut timeline
-if you need exponential fades often or want more powerful audio processing in general: Do the whole audio processing in an external application like Logic

I appreciate the responses.

The odd thing is, there is a fade out earlier in the video that I did using an audio transition cross fade, and it sounded fine to me. The particular audio clip I'm trying to fade now, doesn't though. I don't know why. So I figured that I would try to finesse it with key frames. I couldn't understand why FCP was not allowing me to create a bezier curve the way you can with a video transition or effect. I figured that there must be something I'm doing wrong, a step or setting I'm missing. But it appears that bezier curves simply cannot be used to smoothly ease in and out with audio. I'm kind of shocked and disappointed that this "professional" software will not give me the kind of control that I expected from it. I guess I'll have to try a "work around". The thing is... I bought this software so that I wouldn't have to do any more weird work arounds.

Send the file to Soundtrack which has a number of fade options or do what I do in the occasional situations where the nature of the audio prevents the stock audio crossfade transition from sounding OK.

Layer the audio tracks to separate tracks and either use Denise's key frame method or use the standard audio transition but change the lengths of the fades until you get the crossfade you want. Sometimes I unlink the audio from the video track and extend it 1/2 a second or so to get a good cross dissolve. I'm sure this will work for you unless there's something unusual with this particular bit of audio that you are having problems with.

The reason all NLE programs have OMF exports is so the sound can be sent to a professional audio program as needed. That's a professional method. Every documentary and concert DVD that I have worked on has gone this route and the results are truly worth it. And I let an experienced audio professional do the work, I don't pretend to have more than an adequate ability to manipulate audio.

Thanks William. I will do as you suggest.

I had posted this question on Apple's website also and a responder suggested a method that worked brilliantly. So I decided to share this information here for the benefit of anyone who has been following this thread.

This method, like Denise's, uses keyframes; only you don't have to place all those plot points manually. FCP will do it for you. Here is what the other responder told me:

"Have you used the Audio Mixer? Go to Tools > Audio Mixer. You get slider controls for each audio track, plus one for the master level. If you change these while your timeline is playing, it will change the level of the clip at the current time. It won't make keyframes, it will just change the audio of the entire clip.

To create keyframes, just go to your User Preferences, go to the Editing tab, and on the right column is an option that says "Record Audio Keyframes." Check the box next to this, then choose "All" from the dropdown menu.

Now, go back to your Audio Mixer, play your sequence, and start moving the sliders. As soon as you stop playing the sequence, it will create keyframes that mimic exactly what you did with the slider controls.

If you don't like what you did, move the playhead to a point before the keyframes begin, hit play, and it will replace the previous keyframes."

I did exactly this, and it worked perfectly. FCP plotted a string of keyframe points that was a map of what I did on the slider. I was then able to further fine tune it by manually adjusting any awkward points as a result of my unsteadiness. The result was a smooth ramping down of the audio clip that faded out perfectly.