Well I certainly do - please add a +1 vote for me.
Use Case: I regularly record internet radio streams for later listening. Often these have dropout interruptions of up to several seconds. I would like to mark thos e with region labels.
Peter.
Can that be done in Nyquist? I don’t believe so, unless we have some separate control file?
Perhaps if we added a help screen, I would be (a little) more persuaded.
Please say why a separate “simple” and “complex” effect is a nonsense. I’ve already pointed out effects where we take this approach, and I think there is also reasonable support for a “simpler” EQ to be added to the current one.
I’m unclear why a grandmother who has never used a computer or audio software before should intuitively understand what a “threshold” might be.
Gale
So you would run Sound/Silence Finder twice - once to region label the dropouts then again to point label the songs?
Gale
Because basically you are having two Anlyzers to do basically the same thing - apart from any other consideration it clutters the menus. I could be persuaded to two analyzers though if an “Adavanced Usage” button and dialog can’t be implemented in the available technology.
If grandma has never used a computer before she will have many steep learning curves before she ever gets to Audacity 
Encountering the term “threshold” if previously unknown to one is surely a cue to go off and RTFM 
Peter.
I would and indeed do:
- label the silences - currently I do this by eye and by hand
- Decide if they are in songs I am interested in in which case ignore - else manually manage a clean click-less removal
- Manually label the songs I am interested in - using Ctrl+M and/or Ctrl+Bto mark beginning-ish and end-i_sh_
- Remove the unmarked stuff
- process the marked stuff
Peter
Or worse, to write to feedback@ or me (amounts to the same thing) berating the software and demanding an immediate refund…
Gale
ROFLMAO 
Except I don’t see the funny side of it so much when It happens 
Gale
I’d just give them a full refund of $0.0
To quote my mother (who the grandmother of my nephews):
“Oh for heavens sake Stephen, why do you have to squeeze the toothpaste from the middle? You push me to the threshold of my patience.”
As Gale wrote, that is not currently possible in Nyquist effects, but I am very keen to see “type 4” Nyquist plug-ins that have less restrictive GUI possibilities. This is on my agenda as my C++ skills improve, but there currently appears to be zero support from any of the developers for anything to do with Nyquist (other than RBD who has not been actively engaged with Audacity for some time).
A “Help Button” for Nyquist effects is something else that I’ve been looking at. It appears to be a prerequisite for a “Nyquist Generate Prompt” (or at least, that was about the only objection to having a “Nyquist Generate Prompt” other than the wall of silence).
In the case of Bass and Treble:
- The effect replaced “Bass Boost” which was critically flawed (clipped at 0 dB), not consistent in its UI with other effects, and limited to one task out of 4 common tasks (boost/cut/bass/treble).
- There is a fully functional and comprehensive “Equalization” tool, which is one of the most complex effects to be shipped with Audacity, having a total of 37 sliders, two “modes”, 8 buttons, 2 radio buttons and a check box. Even for advanced users this effect is way over the top for common simple tasks such as “just reducing the bass a bit”.
In the case of Silence Finder, the effect does not do a “good job” except in a few specific cases where there are relatively constant blocks of sound separated by blocks of click free low level “silence”. This limitation is probably the major reason why most experienced users shun this tool and why many novice users are frustrated by this tool. With the simplicity of only 3 controls I see no way to make significant improvement to this effect - it needs at least one more control to more closely define what is sound and what is silence. There are a variety of ways that “an extra control” could improve the accuracy - it could be a second “threshold” control, or a second “duration” control, or a “tolerance” control… but unless we make big (and probably unfounded) assumptions about how the effect is used there needs to be at least one more control imo.
I’d like to do that, but currently not possible given the limitations of the Audacity/Nyquist GUI.
This is a “chicken and egg” situation - we don’t need a more advanced interface for Nyquist plug-ins because they are just simple one-off effects, but we continually see the limitations of implementing more sophisticated effects in Nyquist due to the limitations of the interface. More sophisticated Nyquist plug-ins make the case for a more sophisticated Audacity/Nyquist GUI, but without a more sophisticated Audacity/Nyquist GUI more sophisticated effects are dismissed out of hand as “too complicated” for novice users. 
(rhetorical) Why don’t you use the Silence Finder for that?
And I think that one of the main difficulties is that they are thinking “how do I label my tracks (songs)?”
Their “tracks” (songs) are not silences. In most cases they want to mark “the start of the song” (which is what “Silence Finder” does in those cases where it works), but “finding silences” would appear to be the exact opposite of what they want to do.
I don’t agree that “not asking” is the same as “not wanting”. People only ask for things that they think should exist but appear to be missing. Our documentation for exporting multiple songs gives precedence to using point labels (presumably to tie in with using “Silence Finder”, but I’d expect that in most cases users would want a consistent “space” at the start and end of each song, which is better achieved by labelling the songs with region labels.
I agree that there are probably not many uses for “region labels in Silence Finder”, because most uses of Silence Finder are for labelling “sounds” (ie the start of the song). In this context using a region label for “silences” is not useful, but for labelling songs (sounds) it is highly relevant. It seems bananas to me that the main tool for marking songs (sounds) is “Silence” finder.
It’s not only “something in the code”, it has real impact on using the effect.
Here I have two tracks with about 7 seconds of silence between them. Why has Silence Finder (with default settings) put one label in the middle of the silence and one at the end?
Because I want to avoid false-positives. [rhetorical] Good reason?
Actually I suspect it’s down to the fact that we only had point labels in 1.2.x and most of us documenters got that method embedded in our workflows and mind-sets from that early usage 
Peter.
Yes, but as I am sure you understand, they just purchased the USB gizmo as a means to an end, and Audacity came with it.
Aside from the common misconception that Audacity makes the gizmo, Audacity appears (and probably is) useless to them unless they are able to use it to save their recorded tracks to a file.
The perceived “over-complication” of Audacity was exactly why Ion/Numark stopped distributing Audacity in favour of a “simple” program targeted at a dumbed down perception of what most users would want to do with their recording. But, it kept more of the Ion/Numark customers happy.
Gale
Precisely why I don’t think that tarting up the interface of Silence Finder is sufficient. It needs a rethink.
From a documentation point of view, how difficult would it be:
- IF we had a “sound finder” that worked reliably at marking songs with region labels, to update the documentation to recommend region labels?
- Do you see advantages / disadvantages in doing so (other than the time and effort in writing it)?
Yes, because that is the most common user case. Both “Silence Finder” and “Sound Finder” both detect “Sounds”. Neither of them detect silences 
Sound Finder detects sounds and places region labels relative to the start/end detected sounds.
Silence Finder detects sounds and places point labels relative to the start of detected sounds.
“Silence Finder” is like telling someone that is waiting for a bus: “It is now 5 minutes before the end of the period of no buses”.
Should we even be trying to “compete” for Ion/Numark customers?
How much do Ion/Numark contribute to Audacity?
Perhaps better that we leave their more technically challenged users with “EZ Vinyl Tape Converter” (or whatever they currently ship) and those that want more from their editing software can upgrade to Audacity.
On the subject of “two versions of Sound/Silence Finder”:
- Would you say that splitting tape/vinyl recordings into separate tracks is the main use for Sound/Silence Finder?
- If so, how about a specialised tool for that one specific job, and a general Sound/Silence finder tool for all other sound/silence finding jobs?
The advantage of this is that we could then make accurate predictions about the type of audio being used in the “simple” tape/vinyl track splitter effect.
- The “sounds” will usually be about 3 to 4 minutes duration, and always more than a few seconds.
- The “silences” between sounds will usually be about 4 seconds.
- Tape recordings will usually have fairly constant “pink” noise between sounds.
- Vinyl recordings will usually have crackles and rumble between sounds.
- There may be a long period of “silence” at the start/end of the recording (which could perhaps be used to automatically set the “threshold” level).
I recall you wanted the Help button to go to the online released Manual if there was no installed Manual, but I thought user should have the opportunity to actually download the Manual, as they do in the same scenario when they press Help > Manual in the menu.
That apart, I support a Help button, though I think for challenged users the difference in usefulness between a Help button and a Nyquist Help screen may be quite small.
My point is not why we need a simple EQ, but why do you need to keep Bass and Treble so “simple” while making Sound/Silence Finder so “complex” (without a “simple” alternative)?
The main problem I have with novices (sometimes several times a week) is that they find it very hard to grasp how to use the Silence Finder, to the extent it is hard to even get them to click OK and see how well or how badly it even does at default settings. Try and show them Sound Finder, and hardly any of them can cope with it. 
Of course it would help them enormously if the Silence Finder detection was more accurate, but how close do you feel we could get to user achieving (say) 90% accurate detection, without having to change the default settings, given a typical vinyl recording?
What do the experienced users writing here who shun Silence Finder and Sound Finder do instead?
I’ve no problem in principle with one extra control, maybe even two extra depending on their presentation.
The desire to conflate sound detection with silence detection in the interface is one of my concerns, not least that is an extra control.
Too complicated does not matter for e.g. Sample Data Export. Silence Finder is a completely different “market”, notwithstanding experienced users deserve a powerful tool too. This is what still makes me think two tools are desired for the job.
What I “dismissed out of hand” was a jump from a simple (albeit flawed) Silence Finder to a complex multi-control tool for power users, with removal of GUI help. I do come round more to thinking that a succinct word or two for a control can be beneficial even for novices, but I think in GUI help needs to be there first.
In practice this does not appear to be a problem. They get that they (seem to be) putting a label in in the silence between the songs, and that putting the label a little before the end of the silence is a sensible enough place to put it. They are thus “splitting on silence”. They don’t have to worry about a label having two edges.
As Peter says this is partly historical, but also works well from a practical standpoint. Long experience dealing with the challenged makes me think that point labels are easiest documented and understood. Actually I am slightly surprised at this, but I have no doubt it’s true for the people I have to deal with.
One might think so, but the concept of region labels seems tough to get across to some people. They can understand 0.3s before the start of the song, even if the silence at the end of the song varies.
Because there is a click there. I’ve no problem with improving that.
I’ve no problem with that in itself, but I would still caution that there are many out there who will struggle badly with finding sounds instead of silence.
As far as I can see you have not said why you are (technically) opposed to a simple, accurate Silence Finder (let it actually find silence if you prefer) and a complex, accurate Sound (and Silence) Finder with bells and whistles.
See above for the serious concerns I have about this.
Then as already pointed out there is the problem of people trying to use your new tool finding sounds with region labels when they are looking at other documentation trying to find silences. That is not meant to be a block to progress and improvement, but it is a real concern in the case of this effect.
Please, that really is a code thing. It is not how users are presented with Silence Finder. Alex Brown wrote Silence Finder that finds sounds for a good reason, I think.
Again, there is no reason not to write a Silence Finder that actually does that in code, if you want to.
Gale
In my opinion Ion/Numark, as one of the first to popularise USB turntables and cassette decks with Audacity included, pushed Audacity into the limelight to a very great extent.
That has largely happened, and is not the problem. The problem is all the other bundlers who got on the bandwagon that Ion/Numark started and who largely bundle Audacity.
I believe it’s the main use for Silence Finder. I think the users of Sound Finder have diverse uses for it, but only a small proportion of Sound Finder users use it for splitting vinyl/tape.
This seems a perfectly legitimate alternative approach to me on the face of it, but I still suspect it needs to be presented to the user as “splitting on silences”, because this is what will be most easily understood.
Can this idea cope with the “click in the silence” problem you pointed out?
Gale
The advantage of using point labels for me is that I find it easier to manage the inter-track silences/fades and control their length and position.
Peter.
Well if there are no track lengths given with the LP then I use Ctrl+M as I am recording to place an approximate marker which I can adjust later. Actually sometimes I zoom in and slide the label to the correct position and then type in the label text while Audacity carries on with the recording (but that’s advanced use - and not to be recommended methinks).
Or if the LP has track length listings I will set up a label track with the track positions marked as specified (note: these are not always accurate). I them zoom and slide at the track ends while Audacity is recording to more accurately place the label and type its text while Audacity continues recording (similarly not for the faint-hearted).
Sometimes I just let it record unattended, zoom to fit and then place the labels by eye.
In all cases the label position will need fine tuning when the inter-track gaps silences/fades/duration are dealt with.
The hardest ones to deal with are live albums that just flow - but then Silence/Sound Finder would struggle with that too.
Peter.
Are there any proofs for that or is it just a statement from Ion/Newmark?
It seems to me that the strategies of the different manufacturers diverge in this regard.
Some can’t have enough gadgets built-in in their software packages.
Of course, this is most likely the case when the target customers are expected to be young - like those who purchase DJ-mixers.
I have the suspicion that the statistics often generate a causal connection where previously was none. Furthermore, those enterprises seek to form their clientele based on a subjectively perceived rightness of some doubtful surveys.
Just playing with an idea here.
How about if something like this could be made to work?
There’s really no way to deal with that automatically, other than with track durations from the sleeve notes.
Which is a far cry from the automation that many users want.
I agree that if you want hands-on, per track control over inter-track silences and fades, then point labels are the way to go.
If you want a “standard” 1.5 second lead in, 2.5 second lead out and want it automated, then region labels must be the better option? (so as to tie in with "Export Multiple based on labels).