What is PRTF/HRTF Accuracy?

[HeadphoneFan]

https://www.rtings.com/headphones/tests/sound-quality/soundstage

Hello everyone! I was wondering if some of you could help me with this.

I’ve read rtings.com’s little blurb on their PRTF accuracy measurement numerous times and understand what they do to measure this, but I’m still confused as to the relevance of the stat to my listening experience. Isn’t imaging primarily responsible for giving accurate positional information to the Pinna? How would PRTF accuracy differ from that? What does a natural soundstage sound like? I have an Audio Technica ATH M40x (does very well on this test) and a Philips SHP 9500 (does poorly on the test) but I’m not convinced the ATH M40x sounds more natural...

Any help would be greatly appreciated

 

[solderdude]

I have talked to Sam about it (at Rtings).
It's great that he measures it, really it is as I use that info, but don't agree with the conclusions.
His answer is more or less that he too is still strugling with it but as his website is all about objective measurements and rating them, and not so much the listening side of it, he has to put some numbers on it that weigh along with the total score.
The basis for his conclusions is how the best sounding headphones he measured (HD800S and Ed X) react to this.
As these have fairly similar responses and he likes the stereo imaging it is easy to conclude that it is related.
Personally I think driver-ear distance, angle of the driver, driver size, position opposite the ear and impulse response/phase response as well as FR have a lot to do with how well/easy the brain can 'reconstruct' a sense of depth/realism/instrument pinpointing/separation from the 2CH info the ears receive.

Still the measurements are valuable to me as I measure without Pinna and can see the influence of it (relatively) to my measurements when we both have measured the same headphone.

I ignore the rating myself and just look at the effect between measurements with and without Pinna to see what part of the FR is affected and how much.

 

[HeadphoneFan]

I have talked to Sam about it (at Rtings).
It's great that he measures it, really it is as I use that info, but don't agree with the conclusions.
His answer is more or less that he too is still strugling with it but as his website is all about objective measurements and rating them, and not so much the listening side of it, he has to put some numbers on it that weigh along with the total score.
The basis for his conclusions is how the best sounding headphones he measured (HD800S and Ed X) react to this.
As these have fairly similar responses and he likes the stereo imaging it is easy to conclude that it is related.
Personally I think driver-ear distance, angle of the driver, driver size, position opposite the ear and impulse response/phase response as well as FR have a lot to do with how well/easy the brain can 'reconstruct' a sense of depth/realism/instrument pinpointing/separation from the 2CH info the ears receive.

Still the measurements are valuable to me as I measure without Pinna and can see the influence of it (relatively) to my measurements when we both have measured the same headphone.

I ignore the rating myself and just look at the effect between measurements with and without Pinna to see what part of the FR is affected and how much.

Thanks for the reply! Don't agree with what sorts of conclusions? You mean the weight that Sam assigns to PRTF accuracy? I certainly don't pay much attention to the weightings/overall scores either, I just look at the data in terms of what matters to me and come to my own conclusions. The fact that headphones which score well in PRTF accuracy (i.e akg k702) are consistently identified in subjective listening tests as having natural-sounding stages is perhaps a point in Sam's favour. I also find it interesting that most gaming headsets (which presumably are aiming for accurate sound placement/separation) score well in both imaging and PRTF accuracy. I think it is largely what you say: that PRTF accuracy is one of many factors that contribute to a natural-sounding headphone.

How does the PRTF accuracy affect FR? I don't fully understand this. It's fun to have discussions like this about audio on a forum where people actually care about audio.

While we're on the topic of Sam's soundstage measurements, what do you think of the other PRTF measurements? Size seems fairly easy to understand but distance is a little confusing. How does this "10khz notch" help create the perception of audio being outside the head?

 

[solderdude]

I think soundstage is not to be judged on Pinna effects only which is how he rates them.
I don't think he is totally wrong and has some correlation.

How does the PRTF accuracy affect FR? I don't fully understand this. It's fun to have discussions like this about audio on a forum where people actually care about audio.

The soundstage plot shows the difference in Frequency Response between a measurement with and without Pinna where the 'without' Pinna becomes the reference (i.e. a flat line) and the plot is the change in FR due to the presence of the Pinna.
I measure without a Pinna and use his plot to estimate EQ needs due to my lack of Pinna during measurements.
There is something to be said for measurements with and without Pinna.
So far I have only seen Sam do these measurements which I think give lots of info.

While we're on the topic of Sam's soundstage measurements, what do you think of the other PRTF measurements? Size seems fairly easy to understand but distance is a little confusing. How does this "10khz notch" help create the perception of audio being outside the head?

The 10kHz notch is his conclusion and likely only relevant for his measurement setup.
He is in the best position to evaluate his measurements having the headphones he measured around.
Like me, he does a lot more measurements than the ones published and has learned a lot about it.
Any other Pinna and or fake ear canal may have a dip at another frequency.

 

[HeadphoneFan]

I think soundstage is not to be judged on Pinna effects only which is how he rates them.
I don't think he is totally wrong and has some correlation.



The soundstage plot shows the difference in Frequency Response between a measurement with and without Pinna where the 'without' Pinna becomes the reference (i.e. a flat line) and the plot is the change in FR due to the presence of the Pinna.
I measure without a Pinna and use his plot to estimate EQ needs due to my lack of Pinna during measurements.
There is something to be said for measurements with and without Pinna.
So far I have only seen Sam do these measurements which I think give lots of info.



The 10kHz notch is his conclusion and likely only relevant for his measurement setup.
He is in the best position to evaluate his measurements having the headphones he measured around.
Like me, he does a lot more measurements than the ones published and has learned a lot about it.
Any other Pinna and or fake ear canal may have a dip at another frequency.

What can soundstage be judged by other than imaging, openness and Pinna effects? I don’t know much about this sort of thing. I guess you mean FR should play a part as well? I do think THD could affect the perceived size/accuracy of soundstage (cleaner audio = more spacious-sounding/accurate). Strange he puts so little weight on THD, the combination of THD and Treble overemphasis makes some headphones unlistenable. Nonetheless, I greatly appreciate his work, as do you it would seem.

Interesting, I can see how measuring the effects of Pinna on FR could be useful.

And the PRTF distance calculation is likely only relevant for Sam, ok will keep that in mind. He puts a fair bit of weight on it too...

 

[solderdude]

THD in lower frequencies is not very problematic as our ears are not very sensitive to lows but is important in the mid range to treble range.

Sam's weighting and numbers are still a work in progress at least that's what he told me a while ago.

The openness has little to do with soundstage. Have heard closed ones with a good 'soundstage' and open ones which are narrow soundstage.
I often see people stating these are related... not to me. I totally not buy that reflections on walls from open headphones are audible these will be attenuated way too much certainly compared to the SPL coming from the driver at a few mm from the ears.
A bogus argument just like the crosstalk coming from around the head. I suspect bones in the head give much more crosstalk than sounds bending around the head reaching the other ear.

Pinna effects do have an influence of stereo imaging but it depends on the angle as well.
Only the K1000 and a stax do have sound really coming from in front of you.
The 'angle' some headphones make are alike speakers positioned from the side of ones head but being positioned slightly more forward.
It does have some influence on Concha activation though.
The effect of the slight angle is overrated IMO.
Phase will likely have to do with imaging as differences between L an R at certain frequencies are likely to f things up.
A gentle phase shift over the FR range probably has little effect.

Besides, even with trickery the sound is still IN my head.
For me 'soundstage' is how easy it is for me to separate instruments and pinpoint them in L and R postioning.
Some claim to hear the sound coming in front of them.
I can't (or my brain can't) but possibly other peoples brains can (re)create a 3D image using normal headphones. Let's not talk about some specific binaural recordings here.

Yes, Sams work is much appreciated !
I like the guy too.

 

[HeadphoneFan]

Interesting stuff about THD. I had heard it was less noticeable in low frequencies due to masking.

I find that sound placement is easier with my open SHP 9500s than my closed ATH M40x but I don’t put a great deal of confidence in my personal impressions. I know the SHP 9500s has better imaging if Sam’s measurements are accurate. AFAIK still no scientific evidence or ABX test suggesting openness improves soundstage, to your point. Even if I were to assess the affect of openness on soundstage purely according to my own impressions, I would say its effects are vastly overstated. Yeah being able to place instruments seems far more important than the perceived size of a sound field.

 

[solderdude]

Have a look at the Phon curves and you will see that the actual SPL and perceived SPL (Phon) don't match up for lower frequencies.
This means harmonics in low frequencies (say at 60Hz and 70Phon as 1st harm. of 30Hz) that measures say -33dB (2%) are perceived as -46dB and thus are less audible as we would perceive that as 0.5% instead of 2%.

 

[HeadphoneFan]

Phon curves? Where is this sorry? Rtings again? This is interesting. So THD is less audible in lower frequencies simply because lower frequencies are in general less audible than higher ones, or because lower frequencies are more similar to the harmonic distortion projected than most of the treble or mid range would be?

 

[solderdude]

Just google 'Phon curves' or Fletcher Munson

We measure SPL and convert (or show it) in dB relative to 20 micro Pascal which = 0dB.
We hear in Phon. These are only the same around 1kHz but differ at other frequencies.

 

[HeadphoneFan]

Ok I will look that up. That's good to know

 

[pwjazz]

This is a quality thread, much appreciated!

 

[pwjazz]

I know a bit about the pinna's effect on frequency response, but does anyone know if it also introduces any audibly significant frequency dependant phase shift?

 

[solderdude]

What I understood is that people are not as sensitive to (small) phase shifts as most folks think/suspect/assume it is.

Have passed on the question to Sam (Rtings) as he makes measurements with and without Pinna as the question got me curious.
Biggest measured phase shifts may be caused by the ear canal itself.

 

[solderdude]

Got an answer from Sam (Rtings) and he sent me some raw files to analyse with and without Pinna.

Without posting plots it is easy to see the Pinna does change phase of the headphones.
Unfortunately the phase change does not have any relation between 5kHz and 20khz but do change phase similarish between 2 headphones between 400Hz and 4kHz.
Phase changes are about max 100 degrees at certain frequencies and at some frequencies almost 0 yet at other frequencies -60 degrees.
There is no 'target phase' to be found between 2 different headphones. Of course one would need to analyse much more headphones to be able to really say something meaningful that would hold up in court.

Given the video above I would think the phase shifts in this part of the frequencies are inconsequential to the sound. The amplitude differences, however, are quite audible and also differ per headphone so for that too there is no 'target' here as well.

Interesting excercise none-the-less.