Login / Register 0 items | $0.00 New @ KVR
Mr Entertainment
12016 posts since 29 Apr, 2002, from i might peeramid

Postby xoxos; Mon Apr 07, 2014 4:22 pm Re: Free EQ Tutorial about minimum the distortion and negative effects

bbaggins wrote:The article that is the subject of this thread suggests that long group delays are both commonplace and a practical problem in real life. I'm trying to establish whether or not that's the case.

how big are your speakers? what's your musical application? ;)
you come and go, you come and go. amitabha xoxos.net free vst. neither a follower nor a leader be
tagore "where roads are made i lose my way"
where there is certainty, consideration is absent.
729 posts since 5 Feb, 2005, from Canada

Postby shane1980; Mon Apr 07, 2014 11:42 pm Re: Free EQ Tutorial about minimum the distortion and negative effects

kmonkey wrote:
shane1980 wrote:
bbaggins wrote:Is 80ms realistic in the real world? Or 10ms, for that matter. Can you describe a test scenario that would demonstrate group delays of this magnitude?

I also think about a test scenario for that, but after I found it is too difficult to get done and accepted by most common audiences. In the original tutorial. I use a very bad Reverb plugin to show the effect. But soon I found that it also requires trained ears to hear the metallic effect on the highs.
Group Delay is one of the most hard-to-understand concept in DSP. And even harder to let people know its exist and how it sounds like. The real problem here is how to let people know the effects of group delay, not what exactly it is.

dude what are you talking about? Really...Reverb effect is literally created with series of delays(literally it's working like that). This is not same as EQ. You can't talk about something of being a problem in Reverb then translate this to EQ. FYI bad reverbs do sound metallic.

And even my friend little sister can notice 80ms of delay. Are you sure you are on to something real here.

I don't understand. What practical gain one have if your problem in reality is as you say hard to understand and how it sounds alike if it can't be heard. We are not talking about black holes here.. Come on it's 2014 here. Post some examples, your testing method and we'll see..

Most reverb is cascade filters in structure, so group delay is more likely being accumalted and easier to be heard. That's why you can hear metallic sound in most bad reverbs. But EQ is a bit difficult to hear that. Reverb just lower the request for the trained ears on this point.

BTW, I mean algorithm reverbs and you can easier hear similar distortion on reverbs.
And for EQ,when you boost 18dB with high Q, it needs special signal to hear the distortions
729 posts since 5 Feb, 2005, from Canada

Postby shane1980; Tue Apr 08, 2014 12:04 am Re: Free EQ Tutorial about minimum the distortion and negative effects

quayquay17 wrote:So if when taken to extremes (and isolated) only trained ears can hear the difference, is it really making a difference on the clarity of entire mixes?

In most cases, untrained ears can only feel that it is not sounds good or right.
999 posts since 12 Jul, 2009, from Brighton

Postby DaveGamble; Thu Apr 10, 2014 8:28 am Re: Free EQ Tutorial about minimum the distortion and negative effects

Please excuse me for crashing in uninvited with a bucket full of science.

1. Group delay is the same thing as phase response.
See: http://en.wikipedia.org/wiki/Group_dela ... hase_delay
group delay is the normalised derivative of the phase response.

2. Describing group delay as "distortion" is profoundly and deeply misleading.
I figured we had roughly all agreed that "distortion" (without a more specific qualifier) refers to "Harmonic Distortion"; the generation of harmonic content from the signal. Group delay generates nothing that wasn't already in the signal; it just offsets phases of different frequencies, in PRECISELY the same way that the EQ curve modifies the relative levels of different frequencies.
In any case, phase response (which is what we're discussing) is an inherent property of any linear system. A linear system (by definition) is one that does not generate distortion.
As a rule, usage of the phrase "phase distortion" is overcomplicating something very straightforward. "Phase response" describes the same thing without being misleading. Sorry for getting into semantics.

3. EQ=Reverb=Delay.
Unless we're talking about modulated reverbs, these are ALL examples of linear systems - they can't generate harmonics. They don't distort. EQs tend to have extremely short/simple impulse responses, Delays tend to have extremely sparse impulse responses, Reverbs tend to have longer and more complicated impulse responses. The name we give to an effect correlates precisely with what we hear it doing to the signal. Hence the grey area between "delay" and "reverb" that one might call "echo". Again, semantics.

4. "Fact 1: When you boost or cut frequency bands, it will boost or cut something unwanted or unexpected."
The truth of this "fact" depends on your experience, your signal, and your expectation.
When I boost or cut a frequency band, I have a pretty good idea what it will do.
I tend to use cuts for something unwanted and boosts for things I like. See how I twisted that? Again, semantics.

5. "Fact 2: EQ can indeed bring distortion, due to an effect called "Group Delay"."
As above. Not distortion in the sense that you would usually understand it. Nothing is being added to the signal; it's just being phase-realigned. This is a natural and expected consequence of EQ in both analogue and digital domains.

6. Prescriptive strategies for EQ are unhelpful in general.
When cooking a meal, how much salt should you add?
Hopefully, you'll agree that it depends on what you're cooking, and whether or not you find that salt improves the flavour. EQ is the same story.

7. Group delay can be implemented analogue, using all-pass filters.
It's a bit wobbly, but works fine! My favourite example was the old Focusrite Blue mastering limiter - completely analogue, but with a 10ms lookahead delay implemented with a network of allpass filters. Ace.

8. Delays and reverbs USE group delay in order to provide.... delay.
Hopefully you can see a link here. The clue is in the name. If, every time you read "group delay", you skip "group" and just read "delay", you're close to a decent working understanding of the phenomenon.
I've seen 2-4 samples (at 44100 = ~50microseconds) of delay from analogue EQs. That's about what you expect from a 5band set up to do stuff. 50us isn't much though. Reverbs use filters to achieve MUCH more delay. Typically at the cost of being able to EQ too much. Hence using all-pass - they're flat in magnitude response but give you the phase response (group delay) you want.

Enough rambles. Hope this helps.


Edit: Dammit. Just realised that article I was responding to was an advert. Sorry. I keep doing this. Anyway, hope the above is helpful. I'll be where I usually am if anyone needs me.
[ DMGAudio ] | [ DMGAudio Blog ] | dave AT dmgaudio DOT com
1399 posts since 8 Feb, 2012

Postby antithesist; Thu Apr 10, 2014 8:16 pm Re: Free EQ Tutorial about minimum the distortion and negative effects

I <3 Dave
WEASEL: World Electro-Acoustic Sound Excitation Laboratories
3817 posts since 17 Aug, 2004

Postby kmonkey; Fri Apr 11, 2014 10:22 pm Re: Free EQ Tutorial about minimum the distortion and negative effects

antithesist wrote:I <3 Dave

Yes me too. I waited someone like him to jump in.. :love:

Moderator: Moderators (Main)

Return to Effects