It doesn't look like DC offset.
It looks like phase shifting which is what high pass filters do.

BTW if it's the same audio clip on both images you sure got a problem :lol:

hi,

Thanks for the reply...not the same clip.... 2 takes. However, this is a typical example of how different the wforms look with and without the high pass.

Thanks.

-paul.

Not exactly sure what i'm looking at but if the red line is the zero mark, you do have DC offset to begin with.

The filter is causing the signal to be a-symetrical, it's not uncommon for this to happen on vocals, horns, etc.. Radio stations use phase rotators to correct for this and make on air voice more symetrical. You can try using a less phasey eq if you want your waveform to look more symetrical. otherwise, I wouldn't worry about it.

Mike,

Thanks...can you explain this "Radio stations use phase rotators" and give me an example? [component or software]

Thanks.

-paul.

Basically it just shifts the phase of the lower freq's to come back inline with the higher to make it more symetrical which means a lightly higher output which means a further signal. I don't know of any software solutions only very expensive hardware ones like the optimod or omnia. I wouldn't stress over it, it's going to happen on anything you filter or eq if the eq has any phase shift which most of them do, even if you can't see it on the waveform. You can try using a linear phase eq and you will see an improvement but part of the sound of a great analog eq is the phase shift that it creates. If it sounds good then who cares what it looks like. If you were using 2" tape, this wouldn't even be a subject because you would never have seen it.

Michael Fossenkemper wrote: Not exactly sure what i'm looking at but if the red line is the zero mark, you do have DC offset to begin with.

That was what I thought first, but then I saw the null in the middle.

Mike,

Thanks again for being so helpful.

-paul

nyc ny

As Michael says you have to care about how it sounds and not how it looks. Pulling the signal through a device that can correct this will always ad a color of its own no matter what you use and how much you have paid for it.

Sometimes i have a hard time figuring out a positive side of using a computer with audio? It seems like it only brings on doubt, false aqusations and false conclutions.

Best Regards

Henrik,

Thanks for your help as well. This place is just a fantastic resource.

-paul.

nyc
ny

Michael Fossenkemper wrote: I don't know of any software solutions only very expensive hardware ones like the optimod or omnia.

I was just changing presets on my Optimod 8200 at the station today. Thumbing through the manual I came across a schematic of one of these infamous 'phase rotators': a $5 circuit with two 2nd-order all-pass filters. $20 if you use audiophile-grade parts.

An single-section all-pass filter does not impact the amplitude of the signal, only its phase - and only over a limited bandwidth. Call it what they will, but it's a time-distortion generator since it changes the phase content of the signal spectrum. To cover more bandwidth, multiple stages are cascaded. (I had to design one of these things to cover a 10 Hz - 2500 Hz bandwidth early in my electronics career. Tuning that thing to get the phase response I needed was a nightmare!)

If it makes the signal more symetrical then it does impact the amplitude. Being that it is usually signal with lower peak to average ratios this really doesn't become an issue as other elements in the mix will peak before these do. in the optimod or the omnia, you have to be carefull when using this as it will make the voice more symetrical but can damage the music. If i'm not mistaken, this phase shift is in the 200hz or so range. I've never tried to deal with this or tried to fix it so i can't tell you the best way or if there is a good way. maybe dpd can give us a run down of the circuit. I'm wondering now if something like the bbe sonic maximizer does something like this phase shifting the lower freq's in reference to the higher ones. I have an older one laying around somewhere, if I have time I'll run some noise through it and see what kind of phase shifting it does.

^^^ if you sweep the phase rotator / all-pass filter with a sinusoid it will measure with an absolutely flat amplitude response and the phase response of a LP filter . I have some design notes in my files at work - I'll try to dig them out next week and see if I have any response curves.

Simple, 2nd-order analog all-pass filters generate 90 degrees of phase shift at the filter's center frequency, zero shift well below that frequency, and 180 degrees (e.g. polarity inversion) well above it. Cascading filters can produce much greater phase shift over a small frequency range or can adjust the phase by a smaller amount over a greater frequency range.

When complex signals get passed through these filters the phase of the different frequency components add/subtract differently, thereby affecting the overall amplitude of the waveform via symmetry.

The FM processors use this to reduce the dynamics of the voice region (particularly) so that it can be further compressed. Thus, the average modulation can be increased - making the station 'louder' over the air.

Part time, I'm a broadcast engineer for classical and jazz music stations. I'm staying out of the loudness wars. And, I'm realizing that if I don't preserve dynamics and signal fidelity in my recordings, they don't stand of chance of ending up that way over the air.

dpd,

If you have some time I would love to talk you you about a few aspects of broadcast production. Email me if you can

-paul
ptfiggMac.com