The fizzy frequency

[Ralf_T]

I was going to add this to the "High pass and low pass filter for distorted guitars?" thread, but than I thought it deserved its own post for discussing.

Like most of you, I have also read the article of Craig Anderton dealing with the "fizzy frequency" that is said to be inherent in modeling. And like some of you, after reading this, I was also p!$$ed knowing the only solution to this issue is sacrifying one effect slot for the PEQ ...

But I also was shocked to read that Craig found the annoying whistle on the 11R centered around 7.9kHz. And the scientist in me was asking for the evidence, meaning a comparison with some recorded real amps that serve as a control group ... . So I took some audio samples of recorded amps and ran the same kind of analysis, and - you guessed it - I also found fizzy frequencies. I have uploaded three sound samples that demo this.

The first sound examples contains of sound clips from a microphone comparison featured in last month's issue of "Gitarre&Bass", a German guitar player magazine. Just in case your interested in the files, you can download the files in AIFF format using the link below:

http://www1.gitarrebass.de/magazine/...ff-dateien.zip

And here is the link for accessing the sound samples:

http://soundcloud.com/ralf_t/sets/fizzy_frequency

Fizz mic samples

I have selected five popular mics for guitar recording. C414, MD421, R121, SM57, and an e906. You will first hear the initial sound file for a mic followed by the according fizzy version. For all mics the "annoying" frequency could be found around 4.6-4.7kHz.

Fizz on albums

Of course I was curious whether you can hear the fizzy frequency on professionally produced album. The answers is: Yes, you can. This sample contains three short clips

(CAUTION: Make sure you start listening to this a low volume, especially when listening over headphones!)

The annoying frequency for those clips was centered around 4kHz.

Fizz ubossert

For the last test I used a DI track from the CD accompanying Christoph Reiss' book "Guitar Recording". The track was played and recorded by Uwe Bossert, guitarist with German band Reamonn (their singer is Irish, just in case you wonder about the name ...), in his studio. The book's accompanying CD also includes the wet tracks of those recordings. I loaded the tracks to Pro Tools, created a preset with the same amp and mic models that have been used. Unfortunately there was no hint on the amp's settings, so this comparison is a bit skewed, but anyway ...

At first you will hear the original take; a Marshall JTM-45 through an Orange Box loaded with Celestion Vintage 30s miced with an SM57 and an MD421. After that you will hear the same track with the fizzy frequency. Then you will hear the reamped guitar track with the preset I made, no fizz here. The fizz appears in the last 4 bars.

You can hear that the fizz on the reamped track is louder. But interestingly the annoying whistle is centered around the same frequency, namely 4.29kHZ Now, that is what I call "accurate modeling"
I did a lot more of those analysis for tracks I have recorded with the Eleven Rack, and for most of them the fizzy frequency was centered around 4kHz. No obvous fizz in the 7kHz area.

That being said, my conclusion is that the fizzy frequency is an essential part of the sonic characteristcs of the real amp/cab/mic combination. As far as my observations go, it is obviously more present in modeling then in real recorded amps, but carefully dialing out a tiny bit will help. Especially with clean sounds on a Bassman or Tweed model I found cutting that frequency to -18 dB will take a lot of their character from them instead of adding to the potential richness of the amp sound.

Be careful with the gain control when cutting the fizzy frequency

[Muzza_Boogie]

awesome work, cheers for the sharing of knowledge.

[singtall]

i find that some fizz is needed for a recording to sound right. here is an example of two presets i made:

without fizz:

http://soundcloud.com/singtall/metallica-black


with fizz and added high mids (superfizz):

http://soundcloud.com/singtall/metallica-sbt

[aaron.ferrer]

I don't get what is happening in the Mic Samples file. The first clip is the direct recording from the mic? What is the "fizzy version"? Added gain in the 4.6-7khz?

[Chris Townsend]

Very interesting post. I think there is an important distinction between "analog fizz" from a tube amp (or accurate amp model) and "digital fizz" you can get from some poor quality amp modelers. And there are various flavors of analog fizz, such as what can occur from crossover distortion or speaker breakup. Crossover distortion is kind of a "phasey fuzzy" sound. Speaker breakup is smoother and sort of "noisy" or "hissy". What I would call "digital fizz" sounds harsher and more separated from the guitar track.

I thought Joe Gore's recent blog posting Can You Tell Amps From Models has some good illustrations of tube fizz. In particular, track 6, which is an actual 50 Watt Marshall Super Lead, has tons of that phasey fuzzy sound that these amps are known for. Track 5 is the model used as a comparison and there is very little of that effect happening (although I think it does sound good in it's own right). It's definitely worth checking out.

http://www.seymourduncan.com/tonefie...om-amp-models/

Also, I've downloaded tons of raw guitar tracks that were ripped from guitar hero, and it's amazing how grungy (and fantastic) most of these tracks sound in isolation. These youtube tracks have quite a bit of data compression artifacts, but I think you can get the idea. You can find higher quality versions on the web as well.

EVH - Running with the Devil
http://www.youtube.com/watch?v=DWUsyZVr4yM&hd=1

EVH - I'm the One
http://www.youtube.com/watch?v=3K0LAE

Sweet Child
http://www.youtube.com/watch?v=WDUiesKES90

Welcome to the Jungle
http://www.youtube.com/watch?v=xr-0CreXEL8

Rush 2112 Overture Isolated Guitar Track
http://www.youtube.com/watch?v=3rRG0GouNZY

Red Hot Chilli Peppers - Under The Bridge
http://www.youtube.com/watch?v=8xqXrAn3S5E&hd=1

Beatles - While My Guitar Gently Weeps
http://www.youtube.com/watch?v=t2enJ7YJyAI&hd=1

The Beatles - I've Got A Feeling
http://www.youtube.com/watch?v=e_0P6Tcm2ko&hd=1

Another good one is Jimmy Page's guitar intro on the Lemon Song, but there are tons of others. Anyhow, it really is true that tube amps have "fizz", and I love every bit of it.

Also,I think there are some pretty serious audio artifacts on the soundcloud clips you posted due to data compression or poor sample rate conversion. I hear this on youtube all the time. If you upload a video to youtube at 48k it will auto-magically convert to 44.1kHz before compressing, but the sample rate conversion is dreadful and introduces aliasing artifacts. The soundcloud clips sound like that to me.

[bredo]

There's no such thing as a "constant" fizzy frequency (but some are more annoying to the human ear).

With the same amp, same setting (or on a modeler), this will change with what type of guitar, pickups, strings, picks, picking technique, where you hit the strings, cables and so on.
All these parameters will affect which overtones and harmonics are produced. Some will spike like crazy, some will phase out, some will....... you get it.

In Craigs article, the most important is to get familiar with the procedure to find "your" trouble frequency, and how to tame it.

[Vercingetorix]

But what about people who have a "fizzy" frequencies present in their ears at all times?

Can I use a PEQ to dial out the tinnitus? .

Anyway, thanks for posting that article Ralf T, it made for an interesting read and thanks also for chiming in there Chris T., it was good/surprising to hear those guitar parts in isolation. Even accounting for YouTube compression there is a lot more fizz on some of those tracks than you can hear in the final song!

[dmock66]

Excellent article - thank you!

I agree - there is some fizz that adds the good nasty to your tone - but also it's helpful to remove all the extra noise/fizz that causes the problems.

Just to assure I don't assume incorrectly - the cuts at 4.6/4.7 and 7.9 are at a very narrow Q (like 10.0 in the Para EQ), right? Also, can you accomplish that using the end two blocks of the Para EQ? If so, what parameter setting do you use (Peak, Shelf, LP/HP6/12/24 etc) to accomplish that?

Thanks in advance!
David

[dmock66]

... another question...

If I'm already using a LP filter at say 6.1kHz do I need to concern myself with a 7.9kHz cut? I plan to test this on my own when I get a chance - but in theory shouldn't that frequency already be taken care of by the LP filter?

[madgt621]

To your first: "Peak"

To your second: depends on the roll off of the lp filter. you can go above and cut it IF it's still there and you want to keep whats left of the higher frequencies intact. Technically. Whether or not you should do it is up to your ears.