Audacity Forum

The thing that I keep running up against is problems with how to implement presets.

It has been said that we want presets and I agree that they can be very convenient, however the problem is that some presets will by their nature exclude some of the controls.

Examples:

1) Linear Fade.
By definition a linear fade is not "curved".


2) "Equal Power" means: "the power level when cross-fading non-correlated audio remains at unity gain".
This can only apply when

1. crossfading
2. the power gain is unity.

Furthermore, for this to occur without a sudden "jump" at the start/end of the crossfade, the fade must be between 0% and 100% of the original level.
This means that we must chose one of:

1. High/Low point gain do not apply to "Equal Power" presets.
2. We do not use the term "Equal Power"

It has been said that:
Offering "fixed" presets is too restrictive.
Providing controls that only work on some "presets" is too confusing.
Providing controls that work on all "presets" is too complicated.

I'm up to version 29 now and I can't find any satisfactory solution within the limits of the Nyquist plug-in interface that satisfies all of these conditions.
Sorry folks but it looks like the end of the line for a "release" version, which probably also blows out the "Pro Fade Out" effect.

I've not completely given up on this effect. I've just given up trying to keep all the people happy all of the time.

I'm currently testing a new version which embraces many of the points that have been raised in this thread, so many thanks to everyone that has contributed.
The plug-in is basically finished now and I'm just tweaking it for usability. It will be posted in due course, but as a sneak preview for anyone following this thread, the "mid fade" slider is back, but presets are out.

There has been one significant change to the GUI. "Fade In" is now called "Fade Up" and "Fade Out" is called "Fade Down".
This is why:
There is a big crescendo (highlighted) that then drops back down again.
I want to make it drop down a bit further, so I used an "S-Curve" to smoothly lower the level by 6 dB (50% of the original)
I then want the next section to come in with more punch, so I select the beginning of the next section:
And I fade in the next section from 180% to 100% of the original volume with a slightly curved fade shape....
Oops - that's wrong
I did not want to fade "In" (even though I am fading "into" a new section).
I wanted to fade "down" from 180% to 100%:
That's better, now the next section kicks in with a lot more "punch".

My proposal here is that we should go with one of the later (probably the latest) versions of Steve's adjustable fade - and release it.

1) It is much better than anything we have currently.

2) My perception is that Steve has probably reached the limits what is possible in a Nyquist plug-in.

3) If we release this formally then we can garner feedback and then use the initial Nyquist implementation as a test-bed prototype for a later C++ implementation to be undertaken by a developer.

===============================================================

Furthermore: if we do decide to abandon this work and not release it - then I would definitely not want this to be a blocker for the release of Pro-fade-out.

PFO by itself is an extremely valuable tool - far superior to the current linear fade-out (or the Cross-fade-out). Our users deserve to be treated to this simple, easy-to-use tool that produces excellent musical results.

I'm minded to agree with Steve that its is not actually necesssary to balance it with a Pro-Fade-In - but Steve now does have one of those if it were thought absolutely necessary.

Peter

Thanks for your comments (and encouragement) waxcylinder. I share your views about this.
However, I do have one final offering for the pot...

This is the version that I have been using (and will probably stick with). I have added a version number for clarity (now up to version 30 and I think that is probably enough to choose from )

This version does not have everything, but it has a lot more than is currently available. Within the scope of what can be done with a Nyquist plug-in, I think that this version is close to an optimum balance between functionality, simplicity and productive work-flow.

A suggestion as a way of testing usability: set a keyboard short cut to one of the adjustable fades, and when you need to perform a fade effect, use that keyboard shortcut (I'm using F). Use it for long enough to get used to the shortcut, then see what irritates you about the effect.

Well the interface certainly looks understandable ( even to me ) and not too complicated ...

+1 for v30

Peter.

steve wrote:the "mid fade" slider is back, but presets are out.

Thanks, Steve. I think that was probably always likely to be the next move. Obviously there is a major loss here compared to the 14 "presets" we were considering (i.e. not really "presets" but fade type choices). This isn't any more than an off-the cuff comparison - a concern that people will be forced to use the possibly hard to understand mid-fade boost more than needs be. I know you have been trying hard to make "presets" work, and I don't have time to test and consider it just now.

I'm not sure about fade up/down, I find even your example much easier to think of in terms of a fade out. I doubt that's a good change.

I agree it looks pretty "spiffy" and hope that if nothing else the types of control and their order are now settled. I think though if we use the "gain" word for the third control we should use it for the following two controls.


Gale

Gale Andrews wrote:I agree it looks pretty "spiffy"

+1 (nice word Gale ...)

Gale Andrews wrote:I ... hope that if nothing else the types of control and their order are now settled.

+1

Gale Andrews wrote: I think though if we use the "gain" word for the third control we should use it for the following two controls.

I'm minded to agree with this.

Gale Andrews wrote:I'm not sure about fade up/down, I find even your example much easier to think of in terms of a fade out. I doubt that's a good change.

I can see Steve's POV here; the trouble is that without a complete fade to zero (or from zero) then the fade is not really a fade-out (or fade-in).

Peter.

waxcylinder wrote:I can see Steve's POV here; the trouble is that without a complete fade to zero (or from zero) then the fade is not really a fade-out (or fade-in).

The thing that swung it for me was when I was using it and chose "Fade In" when I actually wanted to fade "down". The example that I posted was not theoretical, I actually did that
If I can make that mistake with an effect that I wrote myself, then I'm sure many others who are less familiar with the effect will make that mistake. If someone says "fade it down a bit", then there is no doubt that they want it to go from a higher level to a lower level.

Gale Andrews wrote:Obviously there is a major loss here compared to the 14 "presets" we were considering

Funnily enough it doesn't feel like a major loss when using the effect. If you have an idea of what sort of fade you want it is very quick and easy to achieve it - in some ways easier than trying to work out what the preset names mean.

steve wrote:

Gale Andrews wrote:Obviously there is a major loss here compared to the 14 "presets" we were considering

Funnily enough it doesn't feel like a major loss when using the effect. If you have an idea of what sort of fade you want it is very quick and easy to achieve it - in some ways easier than trying to work out what the preset names mean.

But unless you are aleady a fade experrt of course it is a big loss ... maybe u dunno what type of fade u wanyt... and even if you do, maybe the fade types teach u to pin the shape with the preset label and learn,,, u know gives u examples to start from... maybe this not good exceptr for the developer that wrote it.

yulac wrote:maybe this not good exceptr

What does "exceptr" mean?

yulac wrote:maybe the fade types teach u to pin the shape with the preset label

This is a Nyquist plug-in.
Nyquist plug-ins do not have the ability to save user presets, so unfortunately that option is not on the cards.


This effect has not yet been documented so there aren't any instructions yet, but there will be. The instructions will be simple.
Have you had a go with version 30?
Do you understand what the controls do? If not I'll be happy to describe them - this will help with the documentation.