Skip to main content

I meant to mention this a long time ago and I just remembered it again. When I was trying to mix and master for the first time on my band's album (and I am no mastering engineer for sure) I was using Izotope's "Ozone" for master bus compression and eq. Well one thing Ozone can do is analyze the overall frequency of other tracks and kind of give you a goal to shoot for. Well at the time I used Metallica's "Black" album for the reference. I dont remember which track in particular. Anyway, when I referenced the track, it showed like a straight verticle rolloff at about 18k straight off the chart. How the heck did they do that and is that a standard practice?

Comments

anonymous Wed, 03/11/2009 - 16:33

You could re-sample your audio at 36kHz. That would do it. You could use [[url=http://[/URL]="link removed[/]="link removed[/] and "crush" it 20%. Although I have a hard time believing that album really has that steep of a roll off. Maybe the settings on Ozone are for the wrong sample rate? Or it could be the source file. Did you rip it as a WAV and not am MP3 from the CD or is a different source?

Boswell Thu, 03/12/2009 - 03:59

It's not difficult to apply very steep filters at 18KHz to audio tracks. It's done, for example, where frequency space needs to be cleared for a 19KHz sub-carrier for 1KHz bandwidth AM control signals. Using a compatible playback system, these control tracks can be demodulated and used to adjust other variables in the setup, start the video projectors, control the room heating, order a pizza and otherwise play with your mind.

soapfloats Thu, 03/12/2009 - 09:35

I was just at another engineer's space yesterday looking at spectral analysis of Dire Straits' On Any Street (he uses it as a reference CD).
One track (forget which), I noticed a very similar thing. At about 18k, there was a "wall" beyond which nothing showed up - until later in the track, there were small flashes (~25% of other frequencies) in the 18-20k range.

I saw this thread just after heading over there, and thought it very interesting. The album was early 80s, I believe.

Boswell Thu, 03/12/2009 - 10:58

soapfloats wrote: I was just at another engineer's space yesterday looking at spectral analysis of Dire Straits' On Any Street (he uses it as a reference CD).
One track (forget which), I noticed a very similar thing. At about 18k, there was a "wall" beyond which nothing showed up - until later in the track, there were small flashes (~25% of other frequencies) in the 18-20k range.

I saw this thread just after heading over there, and thought it very interesting. The album was early 80s, I believe.

Was it something like this? http://eetimes.eu/showArticle.jhtml?articleID=204701099&queryText=spectrums
This is not quite what I was referring to as these examples still have the underlying audio extending beyond 18KHz.

soapfloats Thu, 03/12/2009 - 11:08

No, not like that.
Imagine that at the furthest right vertical line (can't read the Hz), there is nothing showing up beyond it. And it's not a gradual roll-off either.
It's like someone took an eraser and cut everything beyond that mark.

Eventually, a little bit did show up in that range.

I found it particularly interesting that this album was pre-mp3 compression (I know they sometimes cut extreme high and low freqs).
At first I attributed it to a lack of content in that range (either the source, or mics w/ a frequency response topping out @ 18k). Until that little bit showed up at the end.
Just curious, as I noticed the same thing the OP did, on a much older album.

RemyRAD Thu, 03/12/2009 - 20:10

I'm sorry, I have to make some folks aware of some things.

Codemonkey indicated that anything sampled at 44.1kHz will have response from zero out to 22kHz. That's incorrect. It has a steep cut off at 20kHz. 48kHz sampling will get you out to 22 kHz.

Any studios that are near any broadcast transmitters, AM, FM, TV, can and do cause all sorts of frequency dependent, harmonically related RF interference that can easily be picked up by many/most condenser microphones. These can be of lower harmonics than the actual assigned broadcast frequencies.

Then there are the Operational Amplifiers, particularly IC chip types that can become unstable & oscillate at a myriad of high frequencies, not necessarily heard. This then has to be dealt with some extreme filtering. The 15kHz horizontal sync signal produced by the CRT fly back transformer, is a HUGE problem at many studios. Some of the extraneous frequencies of these spikes and other than 15kHz can be just as easily produced by computer monitors running at something other than 60 hertz refresh rates and/or 50 hertz refresh rates. This will cause a variation in the horizontal synchronization frequency.

What I was transferred from NBC radio to NBC-TV, I thought I was going to lose my mind. TV monitors, without any input source in the control rooms, turned into freewheeling high frequency oscillators, even more intense & at other high frequencies other than 15kHz. Thankfully, my brain finely configured a built-in notch filter that kept me from resigning!

The worst example I every experienced was in 1976. That was the year that the Rolling Stones came out with their album, "Black & Blue". I was working at the #1 rated Progressive Rock & roll station in Baltimore as the overnight disc Jockey. I cued this vinyl record up and when I hit it, I couldn't believe the 15kHz oscillation on this piece of vinyl! I surmised that Mick Jagger must have been watching a soccer game, while cutting his vocal track, with the sound down on the TV. That didn't make any difference since, these guys must all have been already deaf since it was so intense. I mean, not even the mastering engineer caught this. Not even for vinyl. It was really the only time in my life I ever questioned the competency of the British blokes involved with this production. It was horrific, truly. How that made it to the final vinyl still stays with me as deep confusion & curiosity. But then, I was only 19 & my hearing was quite acute extending out to 19kHz.

I'll frequently use FFT filtering to create extreme notches to prevent spikes such as the ones we're talking about. Sometimes, I'll just cut off higher frequencies that are not necessarily integral to the musical content that most "average" people cannot perceive. This can also be necessary so as to not "confuse" any kind of dynamics processing from operating/triggering upon spurious frequencies. This was also a little trick I utilized at NBC-TV, Washington, DC, to improve the operational characteristics of the onboard dynamics processing of our digital SSL console. That 15kHz flyback transformer TV blah blah, is flying around all over the studio & affects the operation of the compressors/limiters. This provided NBC 4, with an audible edge over the competing stations. So it's good to understand how and why you should filter certain things out. In that particular instance, it wasn't about audio quality or Fidelity. It was all about intelligibility. One is not necessarily related to nor important to the other. Talking heads don't necessarily need the same kind of broad response that music requires.

Now while my information is not necessarily firmly based upon theoretical electronics academia & research, it is from my real world observations & experiences. So take this with a grain of salt or Scotch. Maybe tequila instead of Scotch?

What's that sound I'm hearing? Salt shaking?
Ms. Remy Ann David

soapfloats Thu, 03/12/2009 - 21:41

hueseph: And that is why many of us should feel humbled and privileged to be able to soak some of these people for experienced information

Just thought that needed repeating.

Sometimes Remy drops an entire course worth of education into a single post!

That covers the presence of weird spikes, but...
The question of the massive roll-off / wall still hasn't been answered IMO, though.
Is it an intentional extreme LPF, or a matter of lack of frequency response?

Not that it's really important, but I believe that's what the OP was asking about, and having seen something similar, I'm curious as well.

x

User login