SurroundByUs.com

I was recently told that a conversion I did was out of phase. I respect the opinion of the person telling me this (since I really dont know exactly what im doing yet) so i have to ask..... what does that mean exactly. I think i have an idea but would really like a better explanation of this term. Another thing I was told by a different person was that the voice was coming from one of the rear channels. I understand that part. My question is this...Should the vocals come from the center channel? I personally thought they should but would like the advice of people that know more than I do. With that in mind..using spec..how does one get the vocals to change where they come from? Again I think i might know the answer to this but would rather ask than show the depth of my ignorance lol.

With the "out of phase" question, I would simply make sure that I was running the latency test at the beginning of the SPEC guide to make sure my FFT size in SPEC matched what the Plogue settings are. In a future build of SPEC, we're going to eliminate that step entirely but, for now, it's needed.

As for vocals in the rears, it's going to be both source-dependent and how meticulous you really want to be while live monitoring, if you're live monitoring at all. We aren't a bunch of studio mixers sitting at a mixing board with multitracks here. We start off with limitations right off the bat. It doesn't mean we give up and say it's not worth doing, although there certainly is a part of surround community which feels that way. The fact that we can get to where we can get to is amazing, with credit to be given to both Zeerround and those who paved the way with earlier stereo-to-surround methods. If the vocal is perfectly centered in the stereo, you stand a much better chance of centering it in the conversion. If the vocal tends to expand to the entire soundfield, you have a bigger chance of hearing it in the rears. You can be aggressive with your separation in the rears with all the SPEC variants, but also need to remember that, as you remove vocal, you're also removing other parts of the mix. You may be left with a lot of partial instrumentation and vocal in your rears which may result in some audible distorted sounds in your final mix if you get too aggressive. You also might leave just enough in there that it's masked by the rest of the mix and you can really only hear it if you put one ear extremely close to the speaker. The balance between separation and soundfield is one best left to the person doing the conversion. ArcTan, without the Slice rear blend, may offer you a good balance. There are other methods out there on the net which are completely soundfield-dependent, but aren't going to work well with denser mixes. Lastly, we all have drastically different speaker and listening setups. What you hear on your speakers may be different than what I hear on mine. That fact alone explains a good 50% of the differences of opinion that go on on every surround-oriented group out there.

I had the good fortune of having you share some of the work you're referring to with me. It was music that is now 40 years old. There was vocal in one of the rear channels but, when sitting in the sweet spot, the vocal appeared entirely centered. That may be the best you can do with that specific album. Someone doesn't like it? They can move on, I say.

I'd also say that there's plenty of retail mixes, quad transfers, etc., out there guilty of rear artifacts, partial vocal, etc., and we love them.

In the end, do you like what you hear? Would you feel proud sharing it with a friend? That's what matters.

More explanation on "out of phase":

If you send the same signal to all 5.1 speakers the sound from each speaker should arrive at your ear at the same time. That would be "in phase". For a surround system it depends on the time delay being set correctly so that if the speakers aren't in a perfect circle the closer ones are sent time delayed signals so that the sound from all speakers arrives at the listening position "in phase". It also depends on the speakers being wired correctly, all the "plus" outputs of the amp connected to the "plus" input of the speaker. If one or more of those is "flipped" those channels will be 180 degrees out of phase with the ones that are wired correctly.

So "phase" is just a really short time delay. The above talks about phase errors in your system, but you can also introduce phase or time delay errors in your conversions. More on that in a minute.

With practice you can detect (larger) phase errors by ear quickly. Here are a couple of methods:

1) This is really more applicable to stereo but if you move two speakers close together, and the phase is off, the bass will reduce in volume.

2) If you move around the room, when there is a phase error, you can detect a "swishing" sound that is not there when phase is correct. This is easiest to hear on noisy signals like applause. Many HT setup DVDs have phase tests that work like this.

So, coming back to phase errors in conversions, you need to ensure that the path from the audio player to the recorder has the same amount of delay, or latency, for each channel, so they are all "in phase" at the recorder.

In SPEC, the process of moving sound into, and then out of the frequency domain for conversion has latency. If all 5.1 channels go through the same FFT-->iFFT processing then you wouldn't care (for phase reasons, but you still need to care for perfect overall time alignment of your recording, but that's another topic) but in the different SPEC conversion methods different channels do or don't go through the FFT-->iFFT processing.

In LCR, only the rear channels go through FFT-iFFT
IN CC, only the center and rears
In SLICE and ArcTan all five channels
and, if you are using LFE, it doesn't go through FFT-iFFT in ANY method.

If you open up SPEC you will see that there are delay bidules to keep all these channels "in phase" but it depends on you correctly measuring the FFT latency as described in the instructions (and this will be automatic in the next version of SPEC).

So, if you follow the instructions and use the supplied layouts your conversions will always be in phase.

However, if you are adding additional VSTs for pre/post processing, you need to keep any VST latency in mind and again compensate with delay bidules for any signals that don't go through VSTs with latency.

Most VSTs that have latency will have an "extra" pin on the output that is the amount of latency in samples (just like SPEC). Wiring that pin the the delay input of a delay bidule (for channels that DIDN'T go through the VST) is how you would put things back in phase.

If you have more specific info about what channel(s) in your conversion were deemed to be out of phase and how you processed them I'd be happy to provide more specific guidance.

Cheers,
Z

Thanks for the input guys. I will continue to experiment with these various settings in SPEC. I am finding out that there are some albums that just dont convert well. As for the one that River listened to I personally enjoy listening to it but I am always willing to try to make something better if possible. Pretty steep learning curve on some these things but man is it fun.