I'm very unhappy with both harman target especially with the in ear target AND GRAS's new hires head and variants.
There are two main characters of over ear/2013 original harman target. First, bass increase strictly below 200hz. Second rolled off highs after 10khz. Both are artificial or incorrect. 200hz is manually selected as bass frequency point, which doesn't related to loudness contour or speakers in room response. It's completely arbitrary and unnecessary. Second the high frequency roll off. Neither of tyll's attempt and early in room response(eg etymotic and other small room responses) or diffuse filed or free field has rolled off highs after 10k. (at least not that fast). Both are the main characters different from early targets and both are wrong essentially.
Good thing is the first one isn't that bad. It sounded 3 dimensional and somewhat enjoyable.
Then there is the in ear response. It's very bad and sounds like sh* basically. It has no high extension and huge amount of emphasis around 7khz. The reason was the fact that most in ears do measure like that and for shallow insertion it's not easy to change the peak around 7k. But using the averaged response of those slightly tuned in ears is very bad idea. This is the main obstacle to propagate harman target. And it's really confusing. Etymotics target is far more neutral and is done around 30 years ago. (you may think their old heads aren't accurate or advanced as now) Yes they are. They used the kemar from Knowles (later on sold to GRAS as we know now) And used swizlocki coupler instead of the iec coupler we now know. Both couplers are extremely close that most swizlochi are actually tunable(I even have one of those bought last year as a nerd). And GRAS used the same coupler response (basically ). For short, it's still relevant and accurate. Won't verbose. So basically etymotic target good yet not perfect but much better than harman target.
Now the GRAS hires head. They changed the input impedance of the coupler to be "suitable" to measure HIRES content. This makes the coupler acoustic impedance different from human ears. Because all transducers are not perfect voltage source the input impedance of the next system is very important. In the case of human ears is reactive. There is article from B&K showing the early stage of design of couplers with real measurements of human ears. The issue is that new head is still complied to the iec standard. How could that be? The standard(iec 711 or iec 60318-4) only specifys the peak frequency of the impedance not the amplitude, which is at 13.5khz +-1.5k iirc. The new coupler still has the same peak frequency just reduced the amplitude. Is that bad? Refer to the article from bk, yes it no longer resemble the impedance of human ears. Then people will then ask ' it's only over 10khz which wasn't accurate anyway. We are only using under 10khz'. Yeah two problems, first of all if only 10khz and below is used why bother with hires head to begin with? Secondly, which is more importantly, the impedance is measured at reference plane which translate to the ear drum. Only etymotic er4 and some custom iems are inserted at that depth. What's for the other insertion like shallow insert inear and over ear headphones? 0db.kr has the answer. Generally minor different at around 3k 5k and huge difference around 10khz. That's also the reason why jude from headfi always has less 10khz in his measurements and hence the infamous z1r 10k peak confusion. Almost no one knew why at least in the public. It's basically unusable. And some people even think the difference is better or more accurate. No! If you try 10s of pairs of headphones earphones and use sine wave /tone generator to hear whether there is peak of the earphone or the high frequency extension, you will realize the response is completely incomprehensible at that range. Here is another thing, coincidentally the hires head and the harman target both have high frequency roll off. Is it gospel? No, some earphones have peaks and some don't and non are present in hires head. Do you think that's true representation? So in conclusion the new hires head from gras adds more confusion to the field which was already chaotic(iec 318 type, diy mic + tube type, 2cc type, cheap iec711 type) and provides no actual benefits to the user. One side note, someone posted the oratory1990. He is one affected one that's more embarrassing. Which is sad to him because he paid really a lot for his setup. He used old head and new pinna. Think about it. It doesn't work together. Resulting 5-7khz peak for many over ear headphones that's not present in measurements with older pinna. The head is more expensive than apx555 I tell you that much. and pinnaa along are more than 6000 dollars. ( i don't know exactly how much but only more than 6k not less). Tyll is gone, so is inner fidelity. I miss him a lot. We had speakerphone clarityfidelity, he doesn't update anymore. We had 0db.kr, they upgraded to new head. Naver ear fi is still good but they rarely measure over ear(still do just rare) and is one of the last hope. Oratory1990 was probably the new hope but ruined by both harman and GRAS. I also miss m.r.o which last update was a few years ago. Also goldenears(both global and kr site) don't update anymore. They did even in 2016. Seeko.kr doesn't update anymore last was 2017 I believe. It's so sad we don't have many websites to read. I personally still have friend who has b&k 4128c(equivalent to old gras) I can see some measurements. It's sad and worrying and genuinely powerless facing the measuring field.
There are two main characters of over ear/2013 original harman target. First, bass increase strictly below 200hz. Second rolled off highs after 10khz. Both are artificial or incorrect. 200hz is manually selected as bass frequency point, which doesn't related to loudness contour or speakers in room response. It's completely arbitrary and unnecessary. Second the high frequency roll off. Neither of tyll's attempt and early in room response(eg etymotic and other small room responses) or diffuse filed or free field has rolled off highs after 10k. (at least not that fast). Both are the main characters different from early targets and both are wrong essentially.
Good thing is the first one isn't that bad. It sounded 3 dimensional and somewhat enjoyable.
Then there is the in ear response. It's very bad and sounds like sh* basically. It has no high extension and huge amount of emphasis around 7khz. The reason was the fact that most in ears do measure like that and for shallow insertion it's not easy to change the peak around 7k. But using the averaged response of those slightly tuned in ears is very bad idea. This is the main obstacle to propagate harman target. And it's really confusing. Etymotics target is far more neutral and is done around 30 years ago. (you may think their old heads aren't accurate or advanced as now) Yes they are. They used the kemar from Knowles (later on sold to GRAS as we know now) And used swizlocki coupler instead of the iec coupler we now know. Both couplers are extremely close that most swizlochi are actually tunable(I even have one of those bought last year as a nerd). And GRAS used the same coupler response (basically ). For short, it's still relevant and accurate. Won't verbose. So basically etymotic target good yet not perfect but much better than harman target.
Now the GRAS hires head. They changed the input impedance of the coupler to be "suitable" to measure HIRES content. This makes the coupler acoustic impedance different from human ears. Because all transducers are not perfect voltage source the input impedance of the next system is very important. In the case of human ears is reactive. There is article from B&K showing the early stage of design of couplers with real measurements of human ears. The issue is that new head is still complied to the iec standard. How could that be? The standard(iec 711 or iec 60318-4) only specifys the peak frequency of the impedance not the amplitude, which is at 13.5khz +-1.5k iirc. The new coupler still has the same peak frequency just reduced the amplitude. Is that bad? Refer to the article from bk, yes it no longer resemble the impedance of human ears. Then people will then ask ' it's only over 10khz which wasn't accurate anyway. We are only using under 10khz'. Yeah two problems, first of all if only 10khz and below is used why bother with hires head to begin with? Secondly, which is more importantly, the impedance is measured at reference plane which translate to the ear drum. Only etymotic er4 and some custom iems are inserted at that depth. What's for the other insertion like shallow insert inear and over ear headphones? 0db.kr has the answer. Generally minor different at around 3k 5k and huge difference around 10khz. That's also the reason why jude from headfi always has less 10khz in his measurements and hence the infamous z1r 10k peak confusion. Almost no one knew why at least in the public. It's basically unusable. And some people even think the difference is better or more accurate. No! If you try 10s of pairs of headphones earphones and use sine wave /tone generator to hear whether there is peak of the earphone or the high frequency extension, you will realize the response is completely incomprehensible at that range. Here is another thing, coincidentally the hires head and the harman target both have high frequency roll off. Is it gospel? No, some earphones have peaks and some don't and non are present in hires head. Do you think that's true representation? So in conclusion the new hires head from gras adds more confusion to the field which was already chaotic(iec 318 type, diy mic + tube type, 2cc type, cheap iec711 type) and provides no actual benefits to the user. One side note, someone posted the oratory1990. He is one affected one that's more embarrassing. Which is sad to him because he paid really a lot for his setup. He used old head and new pinna. Think about it. It doesn't work together. Resulting 5-7khz peak for many over ear headphones that's not present in measurements with older pinna. The head is more expensive than apx555 I tell you that much. and pinnaa along are more than 6000 dollars. ( i don't know exactly how much but only more than 6k not less). Tyll is gone, so is inner fidelity. I miss him a lot. We had speakerphone clarityfidelity, he doesn't update anymore. We had 0db.kr, they upgraded to new head. Naver ear fi is still good but they rarely measure over ear(still do just rare) and is one of the last hope. Oratory1990 was probably the new hope but ruined by both harman and GRAS. I also miss m.r.o which last update was a few years ago. Also goldenears(both global and kr site) don't update anymore. They did even in 2016. Seeko.kr doesn't update anymore last was 2017 I believe. It's so sad we don't have many websites to read. I personally still have friend who has b&k 4128c(equivalent to old gras) I can see some measurements. It's sad and worrying and genuinely powerless facing the measuring field.