Headphone EQ Artifacts: Phase Shifts, Pre-Ringing, and Their Impact on Headphone Reviews?

[BZ22]

Since we discussed this topic a bit in another Thread, I wanted to start a separate thread for it.

Considering which EQ one uses, an EQ will either introduce phase shifts or pre-ringing artifacts, or even both.
I can clearly hear the difference and am not deceived by the EQ change.

Many reviews focus primarily on gaming and soundstage, but unfortunately, the properties relevant for mixing are rarely mentioned, and when they are, it's usually only in general terms.
I find it critical to assess transient reproduction when an EQ is used and there is phase shift or pre-ringing present.
I have seen headphone reviews where it was immediately clear to me that the person had latency issues and was blaming the pre-ringing problem on the headphones...
the "suckback echo"
- I also want to make it clear that I am not making fun of headphone reviewers, nor am I an expert myself.
I greatly appreciate their work and support it, but I would wish that they also open a DAW sometimes.

My HD 600 with EQ does not appeal to me at all, whereas I find that the Edition XS handles a 3-4 dB low-end boost very well.
I do not use EQ for my headphones or studio monitors while producing or mixing because I know it affects the transients and the phase.

 

[Doodski]

This is a good thread for @solderdude. ShapoW!... Sooper Solderdude appears.... LoL,.

 

[ZolaIII]

Well no one is stopping you to bake and use minimum phase FIR EQ with Auto EQ and such. Of course you won't push what certain drivers/design's can't do. Just remember Harman bass is nothing else then psy equal loudness compensation on medium SPL.

 

[bmc0]

an EQ will either introduce phase shifts or pre-ringing artifacts, or even both.

Headphones are substantially minimum phase devices (save for multi-driver types with crossovers), so accurately correcting the magnitude response with a minimum phase EQ will improve the phase response, not degrade it. Minimum phase EQ does not cause pre-ringing.

I can clearly hear the difference and am not deceived by the EQ change.

Proof? I seriously doubt it. Unless you're doing something incredibly bizarre, the change in magnitude response will be far more audible than any change in phase response.

I have seen headphone reviews where it was immediately clear to me that the person had latency issues and was blaming the pre-ringing problem on the headphones...

Please define "latency issues" in this context.

 

[BZ22]

Headphones are substantially minimum phase devices (save for multi-driver types with crossovers), so accurately correcting the magnitude response with a minimum phase EQ will improve the phase response, not degrade it. Minimum phase EQ does not cause pre-ringing.

I don't quite understand your statement that it actually 'improves' the phase response

Proof? I seriously doubt it. Unless you're doing something incredibly bizarre, the change in magnitude response will be far more audible than any change in phase response.

It sounds like a slight chorus effect on the recording, you can particularly hear it in the higher frequencies.
I could just as well ask you for proof, as well as precise measurements and analyses of the minimal phase behavior of various headphone models and the effects of using a minimum phase EQ on them to ensure neutrality for mixing purposes

Please define "latency issues" in this context.

I meant pre-ringing, which I also mentioned

 

[Keith_W]

I think your options are a bit confused. First, are you EQ'ing with minimum phase filters, or linear phase filters?

If you are doing EQ in linear phase, then there may be pre-ringing, but there won't be any phase shifts. At least, there won't be if you have done it properly.

However, most (nearly all) EQ is done in minimum phase. These can introduce phase shifts, and it may be audible if there is enough shift in the same ERB (@j_j says 15 degrees per ERB). There is no pre-ringing with minimum phase filters. Only post-ringing.

Phase shift is much more audible with headphones than with speakers. Most of the studies on group delay I am aware of are done with headphones.

 

[BZ22]

But what I wonder is that most people don’t use EQs for minor adjustments but rather for significant changes to the entire frequency response.
I really can’t imagine that this would be beneficial for mixing purposes.

 

[solderdude]

A: do not use high Q filters for corrections of headphones.
B: I have measured improvements (using passive filters) in both amplitude and phase when applied correctly.
C: my experiences do not seem to mirror that of the OP.

 

[markstein]

Seems that linear phase FIR filters and minimum phase IIR filters are mixed up by the topic starter.

IIR filters do no cause pre-ringing and should not have latency issues.
Edit: if realized in hardware DSP. As Keith_W points out, there might be considerable latency by the audio driver and buffer framework when using PC based DSP.

For frequency response correction you should use IIR filters.

 

[bmc0]

I don't quite understand your statement that it actually 'improves' the phase response

If a system is minimum phase, its phase response is inextricably linked to its magnitude response. In other words, given either the magnitude or phase response, you can compute the other. Further, it follows that a minimum phase filter which exactly corrects the magnitude response of a minimum phase system will also exactly correct its phase response. No ringing, no phase distortion. Mathematically, this must be true.

I could just as well ask you for proof, as well as precise measurements and analyses of the minimal phase behavior of various headphone models and the effects of using a minimum phase EQ on them

There are some papers on the subject; mostly from Harman (Sean Olive, Todd Welti, and others). See "A Virtual Headphone Listening Test Methodology" and "Validation of a Virtual In-Ear Headphone Listening Test Method". It seems well-established in the literature that most headphones are minimum phase over most of the audible spectrum. There is often some non-minimum phase behavior at high treble frequencies, however.

I meant pre-ringing, which I also mentioned

To me, latency typically refers to input-output delay which is unrelated to pre-ringing.

 

[solderdude]

To me, latency typically refers to input-output delay which is unrelated to pre-ringing.

Filters with pre-ringing need a lot of clockpulses before the actual sample is being converted which is latency, with non pre-ringing filters there are no preceding clock pulses needed and the sample is clocked out almost directly.
I assume this is what OP means.

For music reproduction this is not important. Maybe for video and gaming...

 

[Keith_W]

To me, latency typically refers to input-output delay which is unrelated to pre-ringing.

Latency is related to constant delay filters, AKA linear phase filters, and linear phase filters are related to pre-ringing. The impulse is at the centre of the filter length. The latency can be calculated with (N-1)/2Fs where N is the number of taps, and Fs is the sampling rate.

As has been mentioned, there is less latency with variable delay filters, i.e. minimum phase filters. The impulse is at the beginning of the filter length. I would prefer if they were called "minimum delay filters" because to me that terminology seems more accurate.

There are other causes of latency as well, e.g. ASIO itself has a latency of about 5ms, and older Windows driver modes (especially WASAPI shared, DirectSound, MME, etc. which use the Windows Mixer) may have a latency of up to 30-40ms.

 

[bmc0]

Latency is related to constant delay filters, AKA linear phase filters, and linear phase filters are related to pre-ringing. The impulse is at the centre of the filter length. The latency can be calculated with (N-1)/2Fs where N is the number of taps, and Fs is the sampling rate.

I'm well aware of the basic characteristics of linear phase filters. The point is that if someone is using a filter with objectionable pre-ringing, they don't have "latency issues". They have filter design issues. Latency doesn't cause pre-ringing—the filter design does. Perhaps "unrelated" was the wrong choice of word, but the two terms are not interchangeable.

 

[BZ22]

Of course you won't push what certain drivers/design's can't do. Just remember Harman bass is nothing else then psy equal loudness compensation on medium SPL.

I completely agree with this statement, it is logical and at the same time contradicts the headphone reviewers who recommend that everyone adjust their headphones to a specific EQ target curve.

Blind trust plays a very big role here, and I always associate it with the philosophy of mixing and sound engineers who have been dealing with this every day for decades.

For example, many producers and sound engineers have Sonarworks SoundID on their systems but never use it because it didn’t work for them, and they can't work with the EQ correction that SoundID claims to be FLAT.
Their mixes tend to sound too bright.

Yes, I have a problem with so-called experts and headphone reviewers recommending EQ to everyone.
I also have a problem with superficial reviewers who don't even begin to understand the processing methods used by sound engineers.
In the same breath, it is claimed that by using their EQ settings, we listen to the music as it was intended by the artist and sound engineers.

And it wouldn't be that hard if they simply opened a DAW and learned the basics of sound engineering.
This would help them immediately recognize certain characteristics, enabling them to make fair and realistic assessments.

There are different EQs in sound engineering, whether digital or analog, that have been specifically made for the master channel, individual tracks, or group buses for decades, and not without reason.
And even then, only minor adjustments are recommended.

 

[usern]

Before you call users in this forum "a problem", please answer the question if you were using minimum phase or linear phase filters to EQ headphones when you heard the EQ artifacts?

And it wouldn't be that hard if they simply opened a DAW and learned the basics of sound engineering.
This would help them immediately recognize certain characteristics, enabling them to make fair and realistic assessments.

I have FL Studio, teach me.

 

[KSTR]

@BZ22 , could you kindly provide details of what exactly you did, EQ type/settings etc?

And my 2ct:
Corrective Headphone EQ to a given target is no different than "Room EQ" for speakers, if you overdo it or use wrong general settings it can easily fail and give worse sound than before.

Personally, I use IIR (min-phase) EQ in form of Lo- and Hi-shelves and some low-to-moderate Q parametric bands to "bend" the overall response closer to what I feel is required for a balanced sound, spectrally. On top of that I have some "taste EQ" to tailor the spectrum for recordings that are skewed.

Overall, a rather gentle correction, leaving all the small warts and wiggles in place (when looking at any unsmoothed measured response).

To me this is a completely transparent solution, that is, no artifacts of any sort (beware of headroom issues, though), and the perceived results are better, more balanced than without EQ. The "transient character" as well as some hints of standing waves/resonances is still similar to non-EQ'd, perhaps a bit better.

Important thing to note here, the corrective EQ must be minimum phase, linear phase will sound weird with percussion (pre-ringing).
And the corrective EQ must not try to brutally invert the response in the tiniest details, rather it should only adjust the overall curvature of the frequency response.

Minimum phase problems call for minimum phase corrections to keep the response minimum phase.

 

[Chagall]

I completely agree with this statement, it is logical and at the same time contradicts the headphone reviewers who recommend that everyone adjust their headphones to a specific EQ target curve.

Blind trust plays a very big role here, and I always associate it with the philosophy of mixing and sound engineers who have been dealing with this every day for decades.

For example, many producers and sound engineers have Sonarworks SoundID on their systems but never use it because it didn’t work for them, and they can't work with the EQ correction that SoundID claims to be FLAT.
Their mixes tend to sound too bright.

Yes, I have a problem with so-called experts and headphone reviewers recommending EQ to everyone.
I also have a problem with superficial reviewers who don't even begin to understand the processing methods used by sound engineers.
In the same breath, it is claimed that by using their EQ settings, we listen to the music as it was intended by the artist and sound engineers.

And it wouldn't be that hard if they simply opened a DAW and learned the basics of sound engineering.
This would help them immediately recognize certain characteristics, enabling them to make fair and realistic assessments.

There are different EQs in sound engineering, whether digital or analog, that have been specifically made for the master channel, individual tracks, or group buses for decades, and not without reason.
And even then, only minor adjustments are recommended.

I have yet too hear a headphone that wasn't improved by Oratory's EQ. Sure, I didn't like the Harman amount of bass on HD600 and HD800s, but that was always up to the user's preference.

Regarding "recommending EQ to everyone" comment. What is wrong with that? Research found that there is a headphone frequency response most people will find preferable. Users are free to try it and fine-tune it as they see fit.

Regarding sound engineers and the way it's been done for decades. Did they have access to headphone measurement rigs? Did they take into account Harman research? What is the most popular mixing headphone in the last 30 years and how does it measure?

 

[DVDdoug]

But what I wonder is that most people don’t use EQs for minor adjustments but rather for significant changes to the entire frequency response.

I guess it depends on what you like or what you're trying to accomplish.

I really can’t imagine that this would be beneficial for mixing purposes.

Mixing is a different story. Pros don't generally mix on headphones. If they do, their headphones should sound like good speakers/monitors in a good room (the Harman curve or similar) and/or they need to learn what a good mix sounds like on their headphones, and keep a known-good reference recording handy.

Here are some excerpts I've posted a few times:

This is from Recording Magazine "Readers Submissions" where readers send-in their recordings for evaluation:

As those of you who have followed this column for any length of time can attest, headphone mixing is one of the big no-no's around these parts. In our humble opinion, headphone mixes do not translate well in the real world, period, end of story. Other than checking for balance issues and the occasional hunting down of little details, they are tools best left for the tracking process.

And this is from a mixing engineer, also Recording Magazine:

Can I mix on headphones?

No. But in all seriousness, headphones can be a secret weapon and it really doesn’t matter what they sound like…

Over time, after constantly listening back to my work from different studios on those headphones I really started to learn them. They became sort of a compass. Wherever I went… It became a pattern for me to reference these headphones to see if what I was hearing was “right”…

I learned them, I knew them, I trusted them. It didn’t matter whether or not I loved them…

So, can you mix on headphones? Probably. I just think you really need to put some time into learning them first…

This is from Floyd Tool's book, Sound Reproduction

Headphones entertain masses of people. Professionals occasionally mix on them when conditions demand it. Both rely on some connection to sound reproduction, that is, loudspeakers in rooms, because that's how stereo is intended to be heard. Stereo recordings are mixed on loudspeakers.

 

[BZ22]

@BZ22 , could you kindly provide details of what exactly you did, EQ type/settings etc?

With SoundID in Zero Latency Mode using my HD 600
Do you use reverb first and then EQ and compression, or the other way around?
Or do you use it as a send effect?


I think that room EQs on the master bus or in the system do not provide an optimal solution for neutrality in sound engineering.

I also have the impression that minimum-phase EQs alter the phase of the signal in such a way that, for example, impulse responses from reverb units, which are very sensitive to phase shifts, become less precise or clear.
These are characteristics that I believe are often overlooked.

 

[ZolaIII]

@BZ22 huh uhh... there is physiology (outer ear, ear channel) difference impacting how we will objectively perceive highs on earphones/headphones. Then there are hearing health state (range influenced by age and illness's) and lastly there's very little headphones which will do highs rather good. Some things they can do better as separation and lower THD but I won't use them for most mastering work especially not highs. To check transistance sure, after all most listening would be done on such. Meaning also a long set of rules when you mix and master on speakers like no willed separation placing only on one chenel pretty much anything, no willed panning and other funny stuff which makes it practically broken for consumption on such. It's not basics it's a process and either you have talent and practice a lot so things come up by them self or it's better that you do something else in your life instead. So much ruined material's out there that it's hart braking. You won't do critical mastering on average mid levels (about 76 dB programme). You will crunch it up to a point when you don't need any equal loudness compensation (86~89 dB programme SPL and +20 dB or more [surround big screen movie mixes usually up to 24 dB] true peeks maximum or close to it as you need to hear every detail and that's why more flat bass headphones are preferred for such things. That's only in short without going into any details.