Any effect plug-in can return audio that is almost not modified - therefore a weak argument. However, “Return Audio” is now renamed to “Apply Action”. This should describe this option in the best manner, I think.
The “Show Analysis” is now at the second place lest Linux and Mac users choose any of the play options (they are at least less seduced to…).
The current version (1.32) has a bug for removing/isolating center. The returned file has not the same length as the original. Use Vocal isolation/removal instead (low-cut at 1 and high cut at 24000).
Version 1.4 will again be faster.
Three minutes are processed in (sensational…) 26 sec. The same over the whole spectrum takes only 36 sec. In general, 1 minute takes 9 to 12 s on my system.
The drawback is that audio with extreme negative correlation has a slightly different weighting. However, this shouldn’t be a problem because the magnitude ratio is very small and the effect will be masked.
I guess that the problem is caused by some quantization error or some bug in Nyquist (already mentioned above).
To be clear, other distortions will have a much greater impact on the result (e.g. data compression).
Some digital artefacts are even less perceptible in the stereo version, compared to the built-in Vocal Removal.
I am currently setting up the help screens. It’s a real pain in the -ahem- neck.
The result would be much more amusing and entertaining if I could pass the job to Koz.
Something like:
Voice Removal:
The Plug-in picks all out of the show it likes, like voice and drums, juggles a little bit with numbers and shuffles the outcome back to you. If you’re not satisfied with the result, give the plug-in author a sound thrashing - this doesn’t improve the quality but is satisfying too.
However, the plug-in has already over 20000 chars/440 lines and I’ve but covered the first 6 points.
I wonder if there was ever such a fuzz about a plug-in that wasn’t intended for shipping with Audacity.
Here’s another sample sound that illustrates about every effect in one go. It is not very subtle - just to ensure everyone notices the difference to the original
we’ll be calling you “MC RJH” - add a stronger beat to it and you could put it on YouTube as a Remix.
:sigh: yes.
Sometimes the documentation can take considerably longer than developing the effect, but the justification is that there’s little point releasing a plug-in to the public if they are not able to use it.
A historical note: When FM synthesis was first developed as a musical instrument it failed to sell, in spite of the huge range of unique sounds, because it was so hard to program and no-one wanted to wade through a huge user manual before they could start using it. Then the Yamaha DX7 came along and became a massive hit because it had lots of built in presets so that it could be used straight out of the box. There are probably still DX7s’ around today that have never been changed from the factory default settings.
I just tried this out and it works perfectly to do exactly what it’s supposed to. This is a total power tool for mashup artists. Awesome, perfect, thanks so much!
Thanks very much for the laurels.
Feel free to propose any effects/modifications that could facilitate your work flow.
Note that the newest version has only one parameter left for each effect (apart from low-cut and high-cut), thus new effects should respect this.
It is here: https://www.dropbox.com/s/tkonxx1njg1lzcu/rjh-stereo-tool.ny
Do we have a plan for releasing a somewhat simplified version of this as a replacement for Vocal Remover, while putting a full “Toolkit” version on Wiki?
The Nyquist Help screens are never going to be sufficient in my opinion. Even a simplified Vocal Remover/Isolator version needs detailed help and examples in the Manual.
The delay feature while not exactly a necessary feature of the Vocal Remover/Isolator seems to be an effective stereo “enhancer”. Is anything like that feature in any other of the Nyquist effects on Forum or Wiki? It seems like it could be part of Channel Mixer, or it could quite intuitively be part of the replacement Vocal Remover if it could be applied with the effect rather than after it.
Well, the Time Shift tool is the actual equivalent, at least if you just shift one channel.
I’ve not seen this option as a stereo enhancer so far. However, you are perfectly right, I guess.
I’ve been pondering about that, thus the late reply.
Delaying one channel produces several effects, depending on the actual delay time.
We all know that panning can be achieved by either different amplitudes or different arriving times. The upper limit for a normal speaker layout is 1.5 ms or about 66 samples.
A greater difference diffuses the stereo image and makes it broader. General Stereo enhancement advice speaks of 5 ms delay as a good reference time.
There are dozens of YouTube video that propose just that.
Higher differences let us hear two different sources–the correlation is gone with the wind.
Interestingnly enough, the delay has a side effect that isn’t very much appreciated in general, namely a comb effect.
For instance, we can use the effect as a simple hum remover:
generate a chirp tone
generate a square wave at 50 Hz.
mix those tracks down to a new track.
make a stereo track out of it (duplicate, make stereo)
(there are 2 mono and one stereo tracks in the project)
Now, call my plug-in and delay by 882 samples.
The hum is not present on the left side, but we hear clearly the notches produced by the comb.
In general, comb is a bad thing in mono sounds because it sounds metallic. However, it is unavoidable for two or more speaker systems and gives a spacial touch.
A pseudo stereo effect uses this phenomenon. There are usually more than just one delay involved though.
I’ve included this option because engineers just use such effects to enhance their records. Therefore, the purpose would be to counteract this stereo enhancement–quite the opposite of your suggestion.
However, as I’ve written 20 posts before, it should actually be fully automatic, i.e. the audio should be shifted such that the similarity is greatest (explicitly for the range of the human voice).
The inverse effect would of course be possible as well. The ultimate goal must be to remove vocals from a mono sound as well or to transfer a mono sound into a stereo version with the usual distribution (drums and vocals center-panned).
In this light, the delay option seems to be a rather weak stereo or removal enhancer. It is of course possible to control the enhancement by delaying and afterwards using “Show Analysis” to go for e.g. 50 % correlation.
I am presumably the only one that would go to such length just to gain some audio quality.
In conclusion, I wouldn’t include this option in the SE version.
No.
I’m a little bit experimenting with a narrower “Center Isolation” and I can’t figure out how to integrate it smoothly.
It should ideally be hidden from the user. However, a narrow center does not well remove vocals (when inverted and played in sync with the original)
As I’ve mentioned to Gale, the skeleton effect (to ship with Audacity)
will not have such a advanced setting.
It will probably only have a single multi choice control with
Remove Vocals
Isolate Vocals
Help
The low- and highpass filter could take the values from the spectral selection.
However, I don’t know yet how feasible and accessible this will be.
I also want to integrate a perceptual filter to skip silent (FFT-) frames. This might speed up the performance for some material.
I’m not wanting to rush you
If there is a more recent fully working version than that in the first post of this topic, then I think it would be useful to add that to the first post and mark it as “updated version”.
You could post the most recent version in a new post, then just add a link to it at the bottom of the first post in the topic. Then if / when there are newer version, update that link. I’m not sure how VI accessible this forum is, but if that is difficult to do, you could just post the most recent working version and I can add a link to the first post for you.
Hello, I downloaded this plugin because the demo (make someone happy) impressed me.
I also downloaded the song clip and tried to remove vocals on the unedited part (first 15 seconds).
However, after using the default settings nothing happens. The track still has the vocals.
I tried changing the lower and upper cut values but I couldn’t get the vocals to dissapear, not even a little bit.
How are you supposed to use this plugin so that the vocals get removed as well as in the example?
Thank you
Ok, so you should perhaps try the slim edition.
It works only with the newest Audacity version (2.1.0 or higher)
The name has changed to “Vocal Reduction and Isolation”, thus it will appear at the end of the effect list.
The Help page can be found at: http://manual.audacityteam.org/o/man/vocal_reduction_and_isolation.html vocalrediso.ny (10.8 KB)
I would expect it to produce the same result (None).
Have you tried the analysis feature of the plug-in?
You can also click on debug, instead of OK.
There might be a hidden bug somewhere, but it is more likely that it does not come from the plug-in itself.
I’m sorry, I somehow totally missed the notification email. I’ll try this version first thing tomorrow and tell you if it works or not.
BTW, is this forum the only way to get the “slim” version currently?
As things stand at the moment:
The effect will be included in the next Audacity version 2.1.1.
It should be available within the newest alpha builds (June, 28th and newer) and the upcoming release candidate (will be announced throughout the forum).
The official release date is the 6th July 2015, perhaps a week delay if not everything goes smoothly.
Stay tuned.
Robert
Note: The documentation page has slightly altered due to a GUI change and corresponds not 100 percent to the given plug-in version.
The functionality is still the same.
The downloaded version will just be overwritten by the shipped version if you are going to upgrade to 2.1.1.
I assume we are still talking about Dinah Washington’s
I’ve recorded an excerpt and the effect works perfectly well, both with removal and isolation.
Try it with this:
All you have to do is importing it, select all (Ctrl+a or similar), call the plug-in and hit enter.
The settings should be “Remove Vocals”, Strength 1.0, low-cut 120, high-cut 9000 or alike.
we’ll be calling you “MC RJH” - add a stronger beat to it and you could put it on YouTube as a Remix.
