[Thrifty Spender]

By the way, guys, does anybody understand how this Adaptive ANC in H95 works? I can't feel it and it seems it doesn't change any levels that I am able to see in application.

Make sure that the ear cup seals well around your ear. It only works if it covers your ears. To be honest, the levels difference is quite subtle.

 

[jagwap]

When you play a file through bluetooth it needs to be encoded to the bluetooth CODEC in use, and back again. All fine, and understood. Once you play a lossy file, it needs to be decoded and re-encoded to the bluetooth CODEC. So the music goes through two sets of lossy encoding and decoding. Transcoding. Each CODEC is going to have a different data rate, so they cannot be sent directly. This transcoding process is not ideal. Each system throws away some different data. The only way to reduce it to one is to use FLAC or another lossless format on the source.

 

[MayaTlab]

When you play a file through bluetooth it needs to be encoded to the bluetooth CODEC in use, and back again. All fine, and understood. Once you play a lossy file, it needs to be decoded and re-encoded to the bluetooth CODEC. So the music goes through two sets of lossy encoding and decoding. Transcoding. Each CODEC is going to have a different data rate, so they cannot be sent directly. This transcoding process is not ideal. Each system throws away some different data. The only way to reduce it to one is to use FLAC or another lossless format on the source.

The point of lossy formats is that they presume that past a certain data rate (which depends on the codec and the actual quality of the encoder's algorithms) the artefacts introduced are inaudible.
Which has already been measured and which has been supported by blind tests.

 

[jagwap]

The point of lossy formats is that they presume that past a certain data rate (which depends on the codec and the actual quality of the encoder's algorithms) the artefacts introduced are inaudible.
Which has already been measured and which has been supported by blind tests.

Yes, but transcoding is doing this twice, at different data rates, using different CODEC engines. Add to that most android phones work at 48kHz so have to convert 44.1kHz music through an indifferent ASCR, and then maybe again to whatever the BT is using, it can get messy.

 

[MayaTlab]

Yes, but transcoding is doing this twice, at different data rates, using different CODEC engines. Add to that most android phones work at 48kHz so have to convert 44.1kHz music through an indifferent ASCR, and then maybe again to whatever the BT is using, it can get messy.

if done well, no. We can measure that.
If not done well, yes. If it isn’t done well in 2020, well someone didn’t care - and it’s quite often the case that SQ isn’t something that a lot of companies care about (and B&O in general is a good example of a company that manages to sell audio products regardless of their SQ which can range from dog sh...t to decent).

A lot of people on audio boards spend a lot of time being concerned about lower priority details and miss, at least as far as headphones are concerned, the big thing : frequency response at your eardrums. It’s quite pointless to be concerned about multiple encode / decode process and then listen to music on a pair of H9i.

 

[jagwap]

if done well, no. We can measure that.
If not done well, yes. If it isn’t done well in 2020, well someone didn’t care - and it’s quite often the case that SQ isn’t something that a lot of companies care about (and B&O in general is a good example of a company that manages to sell audio products regardless of their SQ which can range from dog sh...t to decent).

A lot of people on audio boards spend a lot of time being concerned about lower priority details and miss, at least as far as headphones are concerned, the big thing : frequency response at your eardrums. It’s quite pointless to be concerned about multiple encode / decode process and then listen to music on a pair of H9i.

Agreed. But I would differentiate between B&O and B&O Play. B&O Play has been absorbed into the main brand, but there is no way they are the same quality as the original. The original company cares deeply about quality, user experience and sound. Are they the best sounding in the world? Hell no, not at the price certainly, but you will not be let down by their products, and they sound fine. B&O Play was set up as a "let a chinese factory design it to our industrial design, and we'll tweak the DSP a bit" operation. Initially they had real B&O engineers help design it. But they all left when they fundimentally desagreed with the philosophy of the numpty in charge. That and huge redundancies. That's why the original H6 are quite good, but subsiquent models got worse, as the marketing guys kept turning the bass up because they thought they knew better than the audiomeisters.

 

[angelom]

B&O Play was set up as a "let a chinese factory design it to our industrial design, and we'll tweak the DSP a bit" operation. Initially they had real B&O engineers help design it. But they all left when they fundimentally desagreed with the philosophy of the numpty in charge. That and huge redundancies. That's why the original H6 are quite good, but subsiquent models got worse, as the marketing guys kept turning the bass up because they thought they knew better than the audiomeisters.

The "real B&O engineers" and/or "audiomeisters" don't having very good inspiration from the real good audio Gods when creating the h6. A 20dB dipping from 1 khz to 5,5 khz and a 20dB peak from 5,5 khz to 9 khz isn't giving good sound at all.

 

[jagwap]

The "real B&O engineers" and/or "audiomeisters" don't having very good inspiration from the real good audio Gods when creating the h6. A 20dB dipping from 1 khz to 5,5 khz and a 20dB peak from 5,5 khz to 9 khz isn't giving good sound at all.

Nobody's perfect...

You are basing those numbers on the Harman curve correction. If I remember correctly it hadn't been released by then, or at least certainly not acepted as an industry standard. So they were working from the bottom curve to their own standard. The real engineers didn't design it. Probably given a low cost headset and told to make it better (no DSP). My point was they didn't ruin it by trying to copy Beat's 12 dB bass boost.

 

[angelom]

You are basing those numbers on the Harman curve correction.

No, I'm not.

So they were working from the bottom curve to their own standard.

I don't think this is true. The bottom curve is only the raw measurements. The important results are the compensated and averaged (top curve). Even the graph in the bottom is showing big dipping of 14 db and peak of 12-13 dB, AND other dipping of 13 dB from 3 kHz down to 300 Hz.

My point was they didn't ruin it by trying to copy Beat's 12 dB bass boost.

You saying before that things are worse in last few years, but for example all the more recent Beats headphones in last few years measuring better, and sounding more correct, than the old bassy models.

 

[jagwap]

No, I'm not.

I'm not looking for an argument, but there isn't 20dB difference from 1kHz to 5.5kHz on the raw data, more like 8, but there is on the compensated curve, so this is why I assumed this.

I don't think this is true. The bottom curve is only the raw measurements. The important results are the compensated and averaged (top curve). Even the graph in the bottom is showing big dipping of 14 db and peak of 12-13 dB, AND other dipping of 13 dB from 3 kHz down to 300 Hz.

Raw measurements are what everyone gets, before they apply any compensation. There is no industry standard back then. The Innerfidelity compensation curve is their own: a combination of Harman curve at the top end, and raw in the low frequencies. I am not sure if they had moved up to this at that time. Who knows what B&O where using as a reference?

You saying before that things are worse in last few years, but for example all the more recent Beats headphones in last few years measuring better, and sounding more correct, than the old bassy models.

No, I am saying that in the early years the terrible EQ of Beats encouraged marketing departments to follow suit. Beats then tempered their extremes as time went by, as I guess they realised they were not getting repeat business like they wanted. Now owned by Apple they have as you say reduced the bass boost to the Harman curve area. Note Apple Airpod Pros do not add the Harman base boost. Hopefully their new Airpod Studio also doesn't.

 

[angelom]

... but there isn't 20dB difference from 1kHz to 5.5kHz on the raw data, more like 8, but there is on the compensated curve, so this is why I assumed this.

I explaining in my previous post that in the bottom graph (the raw data) is showing a big dipping of 14 db, a peak of 12-13 dB, AND other dipping of 13 dB from 3 kHz down to 300 Hz. You saying "more like 8" and the isn't true.

Raw measurements are what everyone gets, before they apply any compensation. There is no industry standard back then. The Innerfidelity compensation curve is their own: a combination of Harman curve at the top end, and raw in the low frequencies. I am not sure if they had moved up to this at that time.

I don't know where you hearing or reading that the compensation curve is the combination you mentioning in the innerfidelity graph in the top. The compensation curve is only the average of how different ears really hearing the results of the raw data and isn't connected with Harman curve. Your comment that the bass is from raw data and the high frequencies is from Harman isn't correct. Other websites maybe having their own target curve, but that is different and they usually saying this and showing with a line with dots in different colour that is superimposing with the measurements of the headphone. Innerfidelity never doing this.

No, I am saying that in the early years the terrible EQ of Beats encouraged marketing departments to follow suit. Beats then tempered their extremes as time went by, as I guess they realised they were not getting repeat business like they wanted. Now owned by Apple they have as you say reduced the bass boost to the Harman curve area. Note Apple Airpod Pros do not add the Harman base boost. Hopefully their new Airpod Studio also doesn't.

You saying that initially Beats influencing b&o bad sound. This, again, isn't making sense and is a big generalisation. b&o headphones never sounding like the old Beats. Only the h8 of the 8 models I know (h7, h8, h9 1g, h9i, h9 3g, h95, e8 1g, e8 2g) having bassy sound.

If you love the h6, this is ok, this is your preference, but the reality is that even the raw measurements is showing clearly 3 bad valleys (peaks and dippings) in the frequency response.

 

[jagwap]

I explaining in my previous post that in the bottom graph (the raw data) is showing a big dipping of 14 db, a peak of 12-13 dB, AND other dipping of 13 dB from 3 kHz down to 300 Hz. You saying "more like 8" and the isn't true.

I don't know where you hearing or reading that the compensation curve is the combination you mentioning in the innerfidelity graph in the top. The compensation curve is only the average of how different ears really hearing the results of the raw data and isn't connected with Harman curve. Your comment that the bass is from raw data and the high frequencies is from Harman isn't correct. Other websites maybe having their own target curve, but that is different and they usually saying this and showing with a line with dots in different colour that is superimposing with the measurements of the headphone. Innerfidelity never doing this.

I read it on Innerfidelity, in the detailed debate on the subject. Perhaps it hadn't been finalised when this graph was plotted, but Tyll went into some detail as to why he favoured using the Harman curve as I said.

You saying that initially Beats influencing b&o bad sound. This, again, isn't making sense and is a big generalisation. b&o headphones never sounding like the old Beats. Only the h8 of the 8 models I know (h7, h8, h9 1g, h9i, h9 3g, h95, e8 1g, e8 2g) having bassy sound.

Beats influenced almost all headphone manufacturers. Not much at the H6 stage for B&O, and obviously heavily by H8 as you say. Sennheiser, Sony, AKG and to a point Beyer, never exaggerated the bass before Beats. Then Beats outsold all of them combined for a while, and this shock things up a bit. It also increased the sales of all of them as headphones were on the ascendance. Then all of them made models which added more bass, often with a bit of mid treble to match. Now market saturation has occurred, and most manufacturers are experiencing a decline.

If you love the h6, this is ok, this is your preference, but the reality is that even the raw measurements is showing clearly 3 bad valleys (peaks and dippings) in the frequency response.

I don't love the H6. I tried it before it came out, and twice since then. It is nice, reasonably well balanced, but ultimately a bit uninvolving, like most B&O kit. But until the H95 it is the best I have heard from B&O so far. I didn't try the H9 as I can see from the plots I won't like them.

I work in the industry, and I have worked with most of the companies I mention here.

 

[angelom]

I read it on Innerfidelity, in the detailed debate on the subject. Perhaps it hadn't been finalised when this graph was plotted, but Tyll went into some detail as to why he favoured using the Harman curve as I said.

Yes, Tyll discussing Harman experiments and the different types of target curves, but his measurements having nothing to do with Harman. Like I explaining before, the compensation curve is having no ralation with the Harman curve.

Beats influenced almost all headphone manufacturers. Not much at the H6 stage for B&O, and obviously heavily by H8 as you say. Sennheiser, Sony, AKG and to a point Beyer, never exaggerated the bass before Beats. Then Beats outsold all of them combined for a while, and this shock things up a bit. It also increased the sales of all of them as headphones were on the ascendance. Then all of them made models which added more bass, often with a bit of mid treble to match. Now market saturation has occurred, and most manufacturers are experiencing a decline.

A lot of the success of Beats is in relation with Beats being a statement of fashion (the design being very important) and very expensive product at the time (today prices can being crazy in many brands), with endorsements and celebrities wearing this headphones. The bass emphasis of early Beats only making this headphone producing big 'wow' effect, specially with music with a lot of bass emphasis (rap/hip-hop and genres with this influence) is so popular. V-shaped headphones, very common today is the equivalent of turning on the old 'loudness' setting (increasing bass and treble) that making the sound very exciting and making a big false artificial soundstage.

Because people using headphones and earphones outside (without anc) in the street and in public transport more and more, more bass and treble is 'helping' this headphones sounding 'better'. This is too one reason the loudness wars (bad dynamic range) becoming worse and worse. This increase of bad DR is strongly in connection with portable earbuds (not in-ears or headphones) where the outside noise is coming inside the ears very easily because earbuds having shallow insertion, sitting outside the ear canal.

I don't love the H6. I tried it before it came out, and twice since then. It is nice, reasonably well balanced, but ultimately a bit uninvolving, like most B&O kit

The graphs I seeing of h6 telling me that I will not liking this headphone. I don't think is very balanced at all.

 

[matteventu]

P2

LMAO nice way to say sorry gifting you an End Of Life product

How can I tell if aptX Adaptive is working? My android device says it's using aptX HD. I'm curious weather I can get the 80ms advertised low-latency for casual gaming (e.g. Sonic 2 for android, or beat stomper).

You need a phone (or other Tx device) that supports aptX Adaptive. List (not full) here: https://www.aptx.com/product-listing?aptx_type=336

 

[plakat]

Yes, Tyll discussing Harman experiments and the different types of target curves, but his measurements having nothing to do with Harman. Like I explaining before, the compensation curve is having no ralation with the Harman curve.



A lot of the success of Beats is in relation with Beats being a statement of fashion (the design being very important) and very expensive product at the time (today prices can being crazy in many brands), with endorsements and celebrities wearing this headphones. The bass emphasis of early Beats only making this headphone producing big 'wow' effect, specially with music with a lot of bass emphasis (rap/hip-hop and genres with this influence) is so popular. V-shaped headphones, very common today is the equivalent of turning on the old 'loudness' setting (increasing bass and treble) that making the sound very exciting and making a big false artificial soundstage.

Because people using headphones and earphones outside (without anc) in the street and in public transport more and more, more bass and treble is 'helping' this headphones sounding 'better'. This is too one reason the loudness wars (bad dynamic range) becoming worse and worse. This increase of bad DR is strongly in connection with portable earbuds (not in-ears or headphones) where the outside noise is coming inside the ears very easily because earbuds having shallow insertion, sitting outside the ear canal.



The graphs I seeing of h6 telling me that I will not liking this headphone. I don't think is very balanced at all.

Tyll started experimenting with the Harman target curve in 2014:
https://web.archive.org/web/2019032...m/content/first-peak-head-measurements-harman

And here is another post that sketches "a crude rendering of what compensated headphone frequency response plots might look like in future":
https://web.archive.org/web/20160422114714/https://www.innerfidelity.com/comment/492722

Later he had his measurement head measured at Harman to be able to derive corrections for the target curve:
https://web.archive.org/web/2019032...m/content/first-peak-head-measurements-harman

So I think the target curve is well documented, with the main difference being less emphasis on the bass (0dB with Tylls original curve vs. +5dB with the 2014 preliminary Harman curve -- that part was changed later as far as I know).

Unfortunately there are no measurements of the H95 from innerfidelity (and there won't ever be...). But however they are tuned, they will have to live up to the expectations of their target demography: if that's pop listening young people that use their headphones over a beanie The tuning will be different from a model targeting a more audiophile-oriented audience: the mastering of the source material may be different, the use cases etc.

I can well imagine that B&O targets different audiences with different models, i.e. there are models with more V-shaped curves and more linear models, where the Harman target curve is a different definition of linear: not linear FR but rather linear perception akin to a good speaker system.

Going by the marketing text and photos I don't think B&O has an audiophile audience in mind with the H95... which is not to imply that they sound bad, but that they may just not play the audiophile game. The other model mentioned (H6?) may simply target another audience and I don't think it tells anything about the H95.