[Dalmonegrig]

You asked if it is the only possibility, and it isn’t, but it’s one of them and thinking you really hear a change when fully aware of the codec setting, sadly that does not remove the possibility of expectation bias. A diagnostic about sound change should start by making sure it exists.

I would also add, you can set a change on your source but the headphone does something else. Like some sources have more codec options than the headphone but won’t tell you when it’s not accepted(they just fall back on a more universal codec while showing whatever you picked still highlighted in the selection).

And back again to bias, there could be a difference in delay when using a certain codec that gives your brain another evidence of change, or a sound when the internal dac changes sample rate, which in turn could be enough to experience other differences in sound that don’t have to exist.

It could be that the conversion between your file and some codecs are audible in some cases and not others.

This scenario could involve sound differences you’re really hearing because while fairly transparent, few codecs(if any) are entirely faultless under all conditions with all files.

Interesting and useful information for me.
You can be very persuasive.
Thank you, I am very grateful to you.

 

[Dalmonegrig]

That doesn't mean there can't be a reason codecs change audio quality because bit rate can be bottlenecked by your gear for whatever hardware or software reasons,

Yes, it may be related to the software/hardware.

 

[bigshot]

Yes, it may be related to the software/hardware.

Highly unlikely. Bluetooth had some problems in the early days, but it's had plenty of time to be implemented properly by now. If you are using equipment that is more than a decade old, perhaps it is due to the hardware, but current components should be able to work transparently.

The most likely reason you are hearing a difference is because your listening comparison is too casual and it's allowing error to skew your results. You need to apply controls and do a better test.

 

[Dalmonegrig]

You would make an effort to eliminate bias and perceptual error by doing a line level matched, direct A/B switched, blind listening test with multiple trials averaged,.

What do you mean by:
1) doing a line level matched
2) direct A/B switched

The most likely reason you are hearing a difference is because your listening comparison is too casual and it's allowing error to skew your results. You need to apply controls and do a better test.

Will the test be valid if a number of conditions below are met:
1) the same file is tested
2) the same equipment and its settings are used
3) a blind listening test is carried out (several times)
4) before each audition, a codec is selected, and the subject is not told about it

 

[bigshot]

Line level matched means that you measure the volume output and make them the same. Different components have different output levels. If one is a little louder than the other, human ears will interpret that as being better sound quality.

Direct A/B switched means that you can flip a switch to instantly listen to one sample and then the other. You can flick back and forth and directly compare. Human ears can only remember a sound for a second or two. If there is too much time between one sample and another, you can forget and resort to guessing. The more similar the two samples, the less time it takes for your auditory memory to fail.

If you take these two things into account as well as the ones you mention, you are more likely to get more accurate results.

But remember. You are doing listening tests to find out for yourself, not to prove anything to us. If you want to have your results be meaningful to other people, your testing procedures have to be documented by a third party and submitted for review. But that isn’t necessary to figure things out for yourself. Just do as tight a test as you can, and if you discover you can do it better, do it again. After you’ve done a few blind tests you get good at it, and it becomes a useful tool.

 

[angelom]

What do you mean by:
1) doing a line level matched
2) direct A/B switched

Will the test be valid if a number of conditions below are met:
1) the same file is tested
2) the same equipment and its settings are used
3) a blind listening test is carried out (several times)
4) before each audition, a codec is selected, and the subject is not told about it

The science is already telling you that LDAC is not really a superior codec than AAC (for the human ear), but is always good that you are honest with yourself and try doing as rigorous tests as possible so you can seeing/corroborating if this is true.

And one maybe relevant or, for me, at least very interesting question (that I indirectly mentioning in a post before in this thread) is this: Why, when we ALL know that marketing is VERY effective in our societies, ALL recent premium ANC/BT headphones (Solitaire T, iO-12, Focal Bathys, H95 and PX8), except the ML 5909, don't have LDAC? (Two of this non-LDAC headphones, the iO-12 and Solitaire T, are for people that also wanting very good/excellent performance in wired-only (fully passive) mode, a mode where the majority of ANC headphones performing badly)

Human ears can only remember a sound for a second or two. If there is too much time between one sample and another, you can forget and resort to guessing. The more similar the two samples, the less time it takes for your auditory memory to fail.

This is something that is completely ignored by even supposed audiophiles!! (that our auditory memory is extremely poor)

 

[Bahnhof Zoo]

This thread prompted me to do a test with a friend of mine today actually. I was playing a song, multiple actually (flac files). Gave my friend my phone and had them flip from SBC to AptX adaptive as I tried to guess which was which. I was trying to pick up on something...high frequency roll offs etc. well I kept striking out as I guessed SBC when they were playing adaptive multiple times. Lol. I couldn't tell a difference and it turns out that Both codecs sounded really good because I like the sound that the headphones and the recordings reproduced. Nothing more to it than that.

 

[gregorio]

I also adequately perceive information, but in this case, they are trying to convince me that I cannot hear what I really hear. This is already more like Orwell's story Animal Farm.

Yes, of course we are trying to convince you that you “cannot hear what you really hear” because you are a human being and you are listening to music recordings. You “really hear” instruments/sounds in the centre position where there isn’t a speaker, right? So either there is some magic, invisible speaker (in a two channel stereo system) or you cannot actually be hearing what you believe you “really hear”. It should be obvious that the correct answer is the latter. In addition, western music is built on the use of humans’ “expectation bias”, the expectation of melodic and harmonic “resolutions”. This fact has been extremely well studied, proven and exploited by pretty much every western music composer for over 600 years and if you did not suffer from expectation bias, music would make no sense and you wouldn’t be listening to it. Just as you wouldn’t be listening to stereo recordings if you didn’t suffer from the stereo illusion.

And yes, this is a bit like Orwell’s “Animal Farm” because what they end up with (a ruling elite) is no better than what they started with, despite some of them being falsely convinced that it is!

I understand correctly, the codecs don't cause dynamic compression, and therefore cannot influence the deterioration of the dynamic range of the recording.
And it is precisely because of this that “spears break” in our discussion?

No, reporting a symptom which does not correlate with the claimed cause is only one of the fallacies that indicate a typical false audiophile belief “that spears break in our discussion”. There are several others in addition, which further support/confirm that indication, EG: A lack of knowledge/understanding of how lossy codecs actually work, a lack of understanding of the methodology, reliability and quantity of science/engineering that has gone into the testing and development of lossy codecs and therefore an irrational dismissal of it all, and also, the refusal to accept human biases/perceptual errors, that have been proven beyond doubt for centuries and without which consumers/audiophiles wouldn’t be listening to stereo music recordings in the first place!

Will the test be valid if a number of conditions below are met:
1) the same file is tested
2) the same equipment and its settings are used
3) a blind listening test is carried out (several times)
4) before each audition, a codec is selected, and the subject is not told about it

That’s a good start and might give you a valid result but there are some further conditions/concerns that need to be addressed in order to be sure the test is valid. Using your points:
1) You need to be sure they really are the same. You mentioned previously you tested using different services, are you certain they are exactly the same master and if so: Are you certain of the provenance, that the supplied bitrate isn’t stepped down and that the pertinent metadata (such as loudness normalisation) is the same.
2) Using the same settings should not necessarily be the goal. Achieving a volume match to within 0.1dB is the goal and that will probably require a slightly different amp volume setting for some codecs.
3) A minimum of 16 rather than just “several” times is required, to lower to probability of just statistical chance causing the result.
4) In addition to the subject not knowing which codec is selected, the tester should not know either. IE. A double blind test rather than a single blind test.

A good way of addressing points 1, 3 and 4 is to use a lossless original file, create the different codec versions yourself and then use free software (such as foobar) to do an automated ABX, which has the added advantage of not needing a separate tester. Another option is Deltawave, which is also free and can automatically address all 4 points above and also provides an ABX test. In both cases, some care/understanding of the software is required to make sure you don’t accidentally introduce a potentially audible error when creating the versions yourself.

G

 

[bigshot]

extra post

 

[Dalmonegrig]

Yes, of course we are trying to convince you that you “cannot hear what you really hear” because you are a human being and you are listening to music recordings.

In addition, western music is built on the use of humans’ “expectation bias”, the expectation of melodic and harmonic “resolutions”. This fact has been extremely well studied, proven and exploited by pretty much every western music composer for over 600 years and if you did not suffer from expectation bias, music would make no sense and you wouldn’t be listening to it. Just as you wouldn’t be listening to stereo recordings if you didn’t suffer from the stereo illusion.

And yes, this is a bit like Orwell’s “Animal Farm” because what they end up with (a ruling elite) is no better than what they started with, despite some of them being falsely convinced that it is!

GREAT!

I have cognitive dissonance.

There are many interesting thoughts.
I need time to think about it.

I have a question for you. Do you listen to stereo music?

And yes, I am now satisfied with our discussion and grateful to all who are involved.

 

[MiffyRabbit]

GREAT!

I have cognitive dissonance.

There are many interesting thoughts.
I need time to think about it.

I have a question for you. Do you listen to stereo music?

And yes, I am now satisfied with our discussion and grateful to all who are involved.

I actually appreciate everyone who has posted in this thread.

I think another way of interpreting Gregorio's response to you about the lossy codecs. I want to share some thoughts, everyone here has been amazing so far.

These are my thoughts after doing lots of reading and I am not trying to disqualify what you are hearing but wanted to explain some things.
AAC has been developed for decades.
I've posted some screenshots below.
Greg has mentioned to you about international standards. Sony owns record studios and has a huge incentive to pour resources into AAC, Fraunhofer, Dolby, Nokia, Panasonic are some big names. These companies/organizations have a huge incentive to making AAC have a level of quality and perform very well.
ISO is an international organization that publishes standards.

Why are standards important? In networking world, network devices routers, switches, cables are all standards set by international organizations so that they can interoperate and work with each other regardless of brand because the underlying technology follows a standard.

While I am not a sound engineer or an expert in developing psychoacoustic models. LAME MP3 and AAC have gone through tremendous development in 1990s and 2000s and beyond. Public Listening tests on HydrogenAudio show QuickTime Apple AAC doing very well at bitrates like 128 and 160 kbps. FhG (fraunhofers version of AAC) and Nero AAC do pretty good but from a consensus qaac/Apple QuickTime AAC seems to be very good for most problem samples or artifacts. Bigshot himself mentioned it being audibly transparent for them at 192 or 256 VBR.

One more thing I forgot to note! About international Standards, AAC low complexity can be played back by any device/software that supports AAC-LC, whether it's encoded by fraunhofer, Nero AAC, Apple aac. The aac-lc should play back, even on Android it should playback. So that's a good thing about standards.

Recently Xiph Opus is the hot new lossy audio codec where people have been finding transparency at 64 and 128 kbps.

I inquired Gregorio and they let me know back in 90s they listened to LAME mp3 and found it transparent at 160 kbps even. A lot of work has been done developing these psychoacoustic models to achieve a certain level of quality that hopefully everyone can enjoy.

Also what Angelom mentioned as well and others about LDAC not being a lossless codec is right on the dot. LDAC is still compressing audio data but there is a very high level of confidence that it would sound audibly transparent from the audio source (CD, FLAC ripped from CD, etc.)

However, from the dicussions about online store transcoding, different masterings, same mastering, I just recognized also basically putting faith that the online stores where I bought flacs from either gotten it from a lossless audio source or the record company/publishing provided a lossless source, there could be that they could have converted mp3 to flac. (Can't know for sure unless we RIP from the cd ourselves). I wasn't there to watch where these stores get their lossless audio files from.

Some other thoughts are since I wasn't at the music studio at the time of recording, I have no idea what it sounded like in the studio.

But anyhow, I don't want to discourage you from utilizing LDAC. A lot on here resonate that pretty much all BT audio codecs should sound pretty good to a degree at minimum or audibly transparent with the exception that AAC could not play well on Android devices due to either not following a standard or some sort out of my knowledge. I plan to encode opus or lame mp3 to my phone for listening and have it utilized either by Bluetooth headphones or iems or something.

From what I know, as well as looking at Sony's battery life utilizing LDAC significantly reduces the battery life of earbuds.

 

[gregorio]

I have cognitive dissonance.

It would seem so. However, you can’t really be entirely blamed for that. The audiophile world largely exists on a belief system based on audiophile marketing, both direct and indirect (incentivised reviews, testimonials, etc.) and that marketing is commonly designed to mislead/misrepresent the actual facts in order to sell audiophile products. For the majority of audiophiles, these forms of marketing are their only source of information, so they have little choice but to accept it. While this audiophile marketing is internally somewhat inconsistent (if one digs deep enough), it’s typically only when encountering remote areas of the audiophile world (like this subforum) or other areas of the wider audio world which are based on the actual facts/science, that cognitive dissonance becomes virtually unavoidable. That’s why the audiophile world is such an isolated niche community within the audio world and why public audiophile forums like this one commonly ban the discussion of science (or banish it to a remote subforum)!

I have a question for you. Do you listen to stereo music?

Yes, of course. And yes, I am a human and therefore I also suffer from the stereo illusion and from expectation and other biases. That’s why I use objective measurements and sometimes controlled listening tests that eliminate or restrict the influence of certain biases.

Why are standards important? In networking world, network devices routers, switches, cables are all standards set by international organizations so that they can interoperate and work with each other regardless of brand because the underlying technology follows a standard.

It goes a lot further than that. The modern (and even to an extent the ancient) world simply could not function without international standards. Imagine a world where every company or country had its own definition of: The duration of a second or of time in general, a kilogram, longitude and latitude, a meter or a nautical mile, etc. International trade, travel and communications would be almost impossible.

[1] I inquired Gregorio and they let me know back in 90s they listened to LAME mp3 and found it transparent at 160 kbps even. [2] A lot of work has been done developing these psychoacoustic models to achieve a certain level of quality that hopefully everyone can enjoy.

1. I started testing lossy codecs in the 1990’s but I didn’t start to find them audibly transparent until the very end of the 90’s and early 2000’s. BTW, LAME is not an international standard, it’s proprietary, although it’s patents are now expired and of course it operates within the specifications of the MP3 format, which is an international standard.
2. A very common marketing tactic in the audiophile world is to raise some supposed problem/issue (which their audiophile products profess to solve) which was actually known and addressed by science/engineers years or decades earlier, long before some audiophile marketer realised they could get some marketing mileage out of it and the audiophile community even first heard of it. “These psychoacoustic models” were first proposed in the early 1970’s, so there’s been roughly half a century of work on developing them and, their application for the last ~30 years goes far beyond the tiny audiophile market and even the music recording industry as a whole. These models and the codecs that employ them therefore cover numerous international corporations and multiple national/international bodies. For example, AAC is incorporated into digital television and film, so although the international standard is administered by the ISO/IEC, it also impacts many other national/international bodies, off the top of my head; MPEG (Motion Picture Experts Group), SMPTE (Society of Motion Pictures and Television Engineers), ATSC (American Television Standards Committee), EBU (European Broadcast Union), AES (Audio Engineering Society), ETSI (European Telecommunications Standards Institute), IEEE, ITU and probably a few others, not to mention all the multinational companies it impacts. Some of these companies and organisations rely on research by other bodies/organisations but many do their own independent research. So that really is a huge amount of research/work over the last few decades!

However, from the dicussions about online store transcoding, different masterings, same mastering, I just recognized also basically putting faith that the online stores where I bought flacs from either gotten it from a lossless audio source or the record company/publishing provided a lossless source, there could be that they could have converted mp3 to flac.

This is more of an issue than many realise. Firstly, it is common for there to be several different masters for different use cases and the bigger the artist/band, the more famous and older the tracks or albums, the more masters there are likely to be, dozens in some cases. In addition, each of the different distributors/services will have different contracts with the record labels/copyright holders and will only be entitled to distribute certain masters as specified. So, the distributor may receive lossy compressed files created by the mastering engineer/label or it may use it’s own internal staff to create their lossy files and the bitrates/formats will be prescribed by the contract. It can all get rather messy and suboptimal masters, encodes or transcodes are unfortunately not an exceptional rarity, hence why I previously mentioned the importance of “provence”.

G

 

[angelom]

@gregorio

Is refreshing reading your very true comment saying:

"While this audiophile marketing is internally somewhat inconsistent (if one digs deep enough), it’s typically only when encountering remote areas of the audiophile world (like this subforum) or other areas of the wider audio world which are based on the actual facts/science, that cognitive dissonance becomes virtually unavoidable. That’s why the audiophile world is such an isolated niche community within the audio world and why public audiophile forums like this one commonly ban the discussion of science (or banish it to a remote subforum)!"

...AND your comments speaking about recording/master quality and the provenance of so many albums in streaming services, this companies not disclosing important information about this albums. Just post the words "hi-res" and the magic working automatically (for many years I'm saying that the whole 'hi-res' world is complete marketing BS).

Because in the last 6-7 years I 'residing' principally in the (more expensive) ANC/BT headphones threads --because this headphones are my primary source of listening to music all this time-- is becoming often a little lonely and frustrating seeing so, sooooooooooooo much BS that even people with apparent experience posting --- throw a few expensive headphone models in your comments and few audio jargon words and this 'type of marketing' is working so effectively with people with little or no experience but with sufficient money for wanting believing ALL the BS (oh, and if you having several thousands of posts is giving you even more deceiving credibility!). This happening A LOT less when I 'residing' in different (more serious) types of threads...but this threads aren't immune from plenty of misinformation and sufficient BS too.

More frustrating --but in last 2-3 years thankfully becoming more amusing, really-- is how this supposed experts throw false cliche after false cliche regarding codecs differences, BT vs wired performance, streaming services, DACs, amps, cables, etc. Is fascinating for me (not so much in recent years) how much SO many people will NEVER listening to facts when you presenting this facts to them.

 

[Dalmonegrig]

I have questions for the discussion participants.
You use to listen to HD Audio music | Hi-Res Audio?
What do you think about Direct Stream Digital (DSD)?

 

[angelom]

I have questions for the discussion participants.
You use to listen to HD Audio music | Hi-Res Audio?
What do you think about Direct Stream Digital (DSD)?

A CD mastering is ALL that you needing, either from a truly fully (1441kbps) lossless format or a good (but not necessarily LDAC) lossy format.

You have very rare instances where a better master is available only in SACD/'hi-res' format. 99.99999999% of the time the best master is available in CD format.