Is there a reason you used the 3PPO 5128 DF HRTF as measured by B&K instead of the Diffuse Field for the same head calculated by Oratory1990 from the relatively much more precise 12PPO Free Field data supplied by B&K?

Sean Olive from Harman, Linus Media Group, and Headphones.com are all using the latter for the baseline DF response, so I’m just curious if there’s a particular reason you chose the former. Thanks!

Thanks for the question listener, I am aware of this calculated diffuse field curve. In our view the variations that this higher resolution brings, mainly in the 8-10k region are quite minimal. The perspective that compensating a headphone response against these small wiggles could tell us something on how the said headphone could be globally preferred would in our view be far fetched and it was clear after our listening tests that it was not the direction we would be taking. That said we certainly value the work done and we want to get this graphing tool up and running sooner than later so we can showcase what these different approaches bring to the table.

Bravo Pierre !

Have RTINGs considered conducting listening tests with IEM and earbud style products too, due to the lack of a pinna interaction likely resulting in a different preferred FR beginning from ~4kHz onward.

Sep 13, 2024 04:39 AM By sylvieqq

3

Have RTINGs considered conducting listening tests with IEM and earbud style products too, due to the lack of a pinna interaction likely resulting in a different preferred FR beginning from ~4kHz onward.

Hi Sylvie, this is a great suggestion. Indeed there is some basis to believe that the lack of Pinnae interaction could potentially influence what is a preferred response in the High frequencies. That said tough, this would be an experiment with considerably more variable than doing this with an open back over-ear set of headphones. The pinnae does not interact as much as you mention, but the ear canal resonance would influence order of magnitudes more since it become a sealed cavity. Insertion depth would influence along with tip size, ear canal shape and size would influence order of magnitude more. Think of evaluating a pair of speakers where every subject would do their listening in a different room. They don’t hear the same thing. Also with headphones to a degree, but more so with canal blocking IEMs. We feel that we would need many more subjects and a much more complicated process to arrive at something of significance

On that I admit that we are somehow standing by. I know the Harman are hard at work with IEMs preference curve and are also doing evaluations on the 5128 platform. We are not going to pretend they have more resources and are surely more advanced that we can be on these considerations. Something tells me that we may have more Insight on this after the AES convention, maybe.

Does this update mean that your testers will be switching to AP across the board?Including the group delay and phase

No shade guys but it seems like you

• made a study that didn’t produce any statistically significant results
• drew the wrong conclusion (it may be correct but it’s completely unrelated to your results)
• just winged it, 1 month before the 5128 Harman target releases

Sep 16, 2024 10:40 AM By andibrema

120

No shade guys but it seems like you

• made a study that didn’t produce any statistically significant results
• drew the wrong conclusion (it may be correct but it’s completely unrelated to your results)
• just winged it, 1 month before the 5128 Harman target releases

Hi Andibrema, we appreciate all comments and concerns, including criticism, that’s how we can improve. I would need a bit more info on why you think that tough but I can quickly try to address your points.

• The experiment (we don’t refer to it as a study) show no statistically significant preference between 4 curves. This in itself is significant enough to disclose.

• Sorry on the second we would need a bit more understanding of what you mean. The conclusion is that in our group of participant, any of those four curve can be preferred, I don’t think that we conclude anything else than that. That’s the results. Did you read something else in what we conclude?

• On this that’s great news! We did not wing it, we’ve been at this a few months but as mentioned in the article, if there are curves that show to be valid we will include them in our next test bench.

Sep 14, 2024 02:28 AM By Tim1992

15

Does this update mean that your testers will be switching to AP across the board?Including the group delay and phase

Hi Tim, thanks for the question. Yes we are, as we progress into measuring headphones frequency response for our Test Bench 1.8, we are also gathering results for group delay, phase, and more. In our next test bench they will be published since our process involve more than just measuring as you may see in our reviews. Stay tune they are coming.

One comment I have with your test bench is that the voice isolation/noise performance results aren’t valid anymore because many truly wireless buds have a voice pickup unit that uses bone conduction to help drive a AI algorithm to filter background noise. For example if you look on YouTube for the spectacular performance of the Air Pods 2 Pro with voice isolation or the Galaxy Buds 3 in very loud environments they are completely silent with only your voice being heard. But on rtings tests that isn’t reflected because you’re only getting acoustic conduction via your speaker to the earphones mic.

It needs testing on a real person who’s mouth moves. Otherwise the test is invalid. Off the top of my head the ones that do this are: Air Pods Pro 4, Air Pods Pro 2 these two only with iOS. Galaxy Buds 3 and Sony LinkBuds Donughts

Oct 07, 2024 08:53 AM By Dush

100

One comment I have with your test bench is that the voice isolation/noise performance results aren’t valid anymore because many truly wireless buds have a voice pickup unit that uses bone conduction to help drive a AI algorithm to filter background noise. For example if you look on YouTube for the spectacular performance of the Air Pods 2 Pro with voice isolation or the Galaxy Buds 3 in very loud environments they are completely silent with only your voice being heard. But on rtings tests that isn’t reflected because you’re only getting acoustic conduction via your speaker to the earphones mic. It needs testing on a real person who’s mouth moves. Otherwise the test is invalid. Off the top of my head the ones that do this are: Air Pods Pro 4, Air Pods Pro 2 these two only with iOS. Galaxy Buds 3 and Sony LinkBuds Donughts

Thanks for the comment. You bring valid points. We do measurements on humans for On-ear/Over-ear but in terms of IEMs it would be impossible to do so, a measurement microphone would need to be in the ear canal, between the eardrum and the IEM tip. This would invariably break the seal. Some past research have been able to do such measurements but it involve modifying the earbud considerably, it is used in development but it’s not the same earbud anymore. We will see in the future if there are ways to simulate this with a mannequin head but at the moment it’s a limitation of the test. The amount of Noise reduction is still valid, we just can’t test the voice isolation.

Great article, and prefer the new rtings curve than harman Do you plan to add an “auto EQ” system in the futur graph tool? like we can find on hangout audio. Cos you get very good measurement technic and explain, and can be cool to get EQ for correct headphone. Thank

Nov 02, 2024 11:40 AM By masteringheadfone

1

Great article, and prefer the new rtings curve than harman Do you plan to add an “auto EQ” system in the futur graph tool? like we can find on hangout audio. Cos you get very good measurement technic and explain, and can be cool to get EQ for correct headphone. Thank

Thank you for the kind words. We really like the new Hangout Audio Squiglink Interface too. We may consider adding EQ parameters down the line as we plan to keep improving our Graph Tool. We’ll look at what is realistically feasible as we progress.