What Goes On
What Goes On: The Beatles Anomalies List
mailto:junk-noclick@signal11.org.uk

Extras - Out Of Phase Stereo Methods

To experience OOPS, it is best to use a dedicated analogue vocal eliminator circuit, or in the digital domain, an audio editor such as CoolEdit, as already described.

However, if you do not have access to these, then these three methods will allow you to experience the OOPS effect. You don't get the level of control, but it is still usable.


Method 1:

By reversing one of your speaker or headphone leads, you can put the sound "out of phase". The effect is strongest when you listen from a point exactly between the speakers, but this is not a very reliable method. I've seen this documented elsewhere, but I'd skip over it. I mention it only for completeness.

Diagram

Some people inadvertently wire their speakers like this when setting up their stereos. Of those people, some don't notice that it's happened. That shows how weak the OOPS effect is. Of course, it could show how little attention they are paying to their music.


Method 2:

You can also do something similar by disconnecting the ground (-) wire to the speakers, and then connecting the ground wires of the speakers together. It works much more effectively than the above method, in that the position of the listener is not critical.

Diagram

Method 3:

An alternative method to the above is this. It works just as well, although sound only comes from one speaker.

Diagram

On this subject, Tim Scott offers the following advice

"Basically if your CD player (or whatever you will be using) is wired in stereo to your amp, by bridging your speaker the left channel will push the drivers while the right channel will pull them. Thus any signal which is equal to both will be cancelled out, thus giving OOPS. Connecting both of the negative outputs doesn't seem to make any difference to the signal but this is how I have been taught to bridge professional power amps so there must be a reason. Maybe it's something to do with professional amps having circuit logic checks in them which need the two negative poles shorting otherwise it thinks there is a fault [ ... ] If however you had a mixing desk and power amp (instead of a hi-fi system) and had your CD feeding through two mono channels you could then use the pan control to affect how much of the centre channel is cancelled."


Other methods of inducing an OOPS effect into your audio have been discovered :-

Method 4:

Shane Guy adds this useful advice

The stereo that I bought has Dolby 'Pro Logic' circuitry. This involves using five speakers - two front stereo speakers, a centre channel speaker, and two rear speakers. You know what I found out when I disconnected the front three speakers? That the rear channel is, in fact, OOPSed. So when I played "Birthday" through the rears, I got only the vocals. I also played "Helter Skelter" the same way and got to hear the 'hidden' vocals toward the end you mentioned. If you could pass this on, then it would be put to good use with people who have Dolby 'Pro Logic' circuitry in their stereo systems.

Not to be confused with a proper "5.1 surround" system, this is an earlier attempt at surround sound, which works by injecting a rear ambience channel into the music. This unconventional method of generating OOPS works as a side effect of the Dolby Pro Logic decoding process.

With Dolby Pro Logic, your ordinary left and right stereo signals are passed to the front left and right speakers as normal. The extra rear signal, which is in mono, is added partly to the left channel and partly (but polarity inverted) to the right channel. Doing this means that the rear signal mostly cancels out when played on a normal stereo system, and will entirely cancel out listened to in mono. So compatibility with older systems is preserved.

The Pro Logic decoder takes the left channel, and subtracts it from the right channel to recover this rear signal, and plays it through the rear pair. The rear pair get very little of the main signal - which is mostly cancelled out, but lots of the rear signal.

So how does this relate to OOPS? Well, your Beatles recording has no "rear" encoding in it, but the decoder will still attempt to subtract left from right. That's the definition of an OOPS effect!

Just disconnect, or fade down, your front speakers and turn on the "Surround" option, and listen away.


Method 5:

Fiddling while a ROM Burned, Ugo Coppola writes

While fiddling with the filters on the Nero Burning Rom progam, I found another way to obtain the OOPS effect. Nero's Karaoke filter works, according to the program's "Filters" dialogue window, "by removing those parts of the song that are equal on both stereo channels"... exactly what OOPS is supposed to be

The version of Nero I have, 5.5.1.1 (which I think is one of the latest) also features a "Vocal Balance" slide, which should tell the program where the vocals are. Well... I tested it with the well-known OOPS effect on Birthday... and, lo and behold, it worked perfectly

1. Rip the song(s) you want to apply OOPS on into (an) mp3 file(s). (Lots of commonly available encoders can do this.)

2. Open Nero, select 'Compile a new CD', then "Audio CD", click OK.

3. Locate the file(s) you made in 1., drag and drop it/them into the Audio1 window on the left.

4. Right-click on (each of) the file(s), select Properties, then the Filters tab.

5. Tick the Karaoke box.

6. On the dialog window, bring the upper slide to 'Maximum effect'.

7. Click 'Test selected filters'. This will play the song.

8. (For those with Nero Version 5.5.1.1 or later) Move the 'Vocal Balance' slide till you are satisfied with the effect. [Note: On 'Birthday' it works best, IMHO, at its extreme left. But on most Beatles songs it works very well in the middle. It depends on the song.]

9. There you go

I would only add the comment that ripping to a lossless format, such as uncompressed WAV, would be preferable to ripping to mp3, if Nero will let you do that.


Method 6:

Ugo also suggests

Another program I have, called Samplitude, works on .wav files and applies the same kind of filter as Nero does (called Stereo Cancel on Samplitude) to specific sections of a song. You don't have to sample it all, you just choose the bit that you're concerned in and there you go.


Whichever method you choose, if you want to be sure you've wired it correctly, find a CD that is in mono, for example the early Beatles (Please Please Me, With The Beatles). Listen to it through this wiring - you should hear very little sound, as all of it is "in the middle" and cancels out. Try it on a track with vocals in the centre, and check that the vocals vanish. Try it on "Birthday" and see how the music cancels out, but the vocals remain ...

Claude Flowers, among others, poses the question

Secondly, is there any way to "reverse" the out of phase process? i.e. instead of removing shared signals, I'd like to KEEP the shared signals and remove any signals unique to the left and right speaker. Can this be done? Any suggestions?

The short answer is no. Not by the simple subtraction and addition presented so far. Too much information is destroyed in the OOPS process, so any ideas about taking the OOPS'd signal and subtracting that off the original stereo are no good, the maths of it falls apart, and practical tests of this in CoolEdit show that it won't work.

I know it's appealing to think that if you've got a signal with, for example, left: a guitar, right: drums, centre: vocals, that vocal elimination leaves you with "guitar and drums". Therefore, subtract that off "guitar and vocal and drums" and bingo! This is great until you actually work out the polarity of each signal. It stops working at that point.

There are many people that claim it can be done this way, but they can never substantiate this. It's always "a friend showed me" or "I heard that someone ...", or "I deleted it once I'd done it".

This is a shame, because that would be very cool. It was a persistent question on the CoolEdit forum. It was one of the threads not to start.

The only way I've ever achieved something similar is to take the original track (stereo), vocal eliminate it (to mono), and then use Cooledit's Noise Reduction to try and "Spectrally fingerprint" the music content. Then, treat this music as a noise that needs to be statistically removed from the overall stereo signal. This has to be done in tiny little segments for it to work. Each time the background sound changes to any significant degree, a new fingerprint is needed, and that fingerprint can only de-noise the exact part of the song it came from. It's painfully slow, the results are not wonderful, and very very dependent on the source material being just right. It would be hard to automate this process so it "just works", although some newer software such as Elevayta's VST tools comes close to this kind of method.



What Goes On V3.09 © 2010 Mike Brown. Visit the website at http://wgo.signal11.org.uk for the latest version and contact details.
noclick