Is bluetooth aptx perfect for 320kbps mp3?
post-15386038
Post #16 of 62

castleofargh

Sound Science Forum Moderator
Joined
Jul 2, 2011
Messages
8,885
Reaction score
4,243
Joined
Jul 2, 2011
Posts
8,885
Likes
4,243
TBH, it wasn't a strange choice. AptX was originally designed as an intermediate codec for professional use, initially in the broadcast industry. Lossy codecs such as MP3, AAC, etc., are "perceptually" lossy codecs, meaning they mainly rely on auditory masking to reduce file size, IE.They remove audio frequencies that would not be heard/perceived. Perceptually lossy codecs are therefore particularly good as an end use codec but not so good as an intermediate codec. For example, let's say there is some frequency based processing downstream of the codec, say EQ or multi-band compression or limiting; we potentially have a problem with perceptually lossy codecs because we obviously can't process frequencies that have been removed. In other words, an EQ setting could cause certain masked freqs to no longer fall within the masking threshold (and therefore be audible) but if a perceptual codec has been used those masked freqs have been removed and can't be EQ'ed (and potentially unmasked). Compression/Limiting processes may also be impacted as the energy contained in various freq bands is altered by removing the masked freqs. This potential problem is lessened if we don't use a perceptual codec, if instead we just use a time domain codec which maintains all the freq content (albeit at less resolution). In addition, a time domain codec such as aptX is computationally far quicker to execute than perceptual codecs (which require a transform to the freq domain, the application of a psychoacoustic model and a transform back again) and therefore has lower latency, a consideration for certain applications such as gaming or audio/video over IP for example.
That's indeed a compelling argument. Thanks.

I have just read elsewhere that even if you have an AAC source (Tidal high Q) and a AAC codec in your Bluetooth device, Android and Apple iOS as well will decode the stream to pcm first so they can mix notifications in the music played.
That's a bummer.
I guess you're talking about this https://habr.com/en/post/456182/ or someone discussing it. I have not seen evidence that the AAC signal was decoded and then re-encoded to AAC, but a lot of what is said in that article seems legit so I'm tempted to trust that information. Plus AAC has a bunch of encoders, so there is always the possibility that the clever chip in a BT headphone only "knows" one decoder for various reasons? IDK.
 
     Share This Post       
post-15387038
Post #17 of 62

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Messages
3,988
Reaction score
2,024
Joined
Feb 14, 2008
Posts
3,988
Likes
2,024
[1] I've encoded back and forth from AAC 256 to AIFF and back to AAC 256 ten times and there is little or no change in the file.
[2] I've been told that transcoding from AAC to MP3 to Ogg Vorbis to whatever else would introduce more signal degradation than just re-encoding AAC over and over, but I haven't tried that myself. It would be an interesting experiment if someone wanted to try it. The way it was explained to me was that once AAC throws out the masked sound, if you bounce to AIFF and encode to AAC again, it's just trying to throw out the same stuff again, and it isn't there to throw out, so there is very little change to the file. If you change codecs, it throws out something different every time and you end up with what's left.
[3] In practice, I haven't found any audible degradation of the signal at all using bluetooth. I did back when it first started, but the newer ways of handling it are audibly transparent for me and my purposes. No reason to worry about it though because most serious listening is probably done at home wired. I think worrying about this is just another reason for audiophiles to encourage their OCD.
1. I've encoded back and forth from AAC 256 to wav and back again and have found a difference, confirmed objectively with a null test. Am I therefore calling you a liar? No, it's possible that with some (or maybe many/the vast majority of) input tracks there would be no difference.

2. That explanation is broadly true but is a gross oversimplification and therefore isn't entirely true. It relies on some assumptions (for simplicity's sake) that aren't actually true. For example: If we take say AAC, the assumption is that the AAC encoder applies a psychoacoustic model, which throws stuff out, so if we apply that same psychoacoustic model a second time, it will try to throw out exactly the same stuff (which isn't there), so we'll get exactly the same result, IE. No generation loss. Modern lossy codecs are very sophisticated, they are adaptive/predictive and, they don't really "throw stuff out" as such. They work on the principle of numerous small freq bands and reducing the number of bits employed to encode a particular band depending on how masked that band is calculated/predicted to be (which depends on the freq content of neighbouring bands). In other words, stuff isn't really thrown out, it's resolution is lowered or conversely, it's quantisation noise is increased (which will be inaudible/masked). A second pass is therefore not necessarily trying to throw out something that's already been thrown out, there may still be some signal there, there will be some more quantisation noise and therefore the encoder will analyse it differently and process it differently. So, there will be a change although in the vast majority of cases it should be inaudible.

In this particular case we also have another incorrect assumption to deal with, the "if we apply that same psychoacoustic model a second time" assumption. Neither MP3 nor AAC specify a particular psychoacoustic model, therefore an iTunes encoded AAC probably uses a different psychoacoustic model than say a Foobar encoded AAC and even an iTunes encoded AAC may employ different psychoacoustic models depending on the iTunes version. In this case, where apparently a second AAC conversion is occurring in or before the bluetooth device, it's unlikely we are "applying the same acoustic model a second time". In light of this, while your experiment might be somewhat interesting, it wouldn't necessarily be useful because the results may only apply to specific encoder versions.

Caveat: Even my more detailed explanation is still an oversimplification. Psychoacoustic models and perceptually lossy codec design is a complex, specialist field, a field which I don't claim expertise in!

3. Broadly, I would agree. However, I can't completely rule out that degradation of the signal will always be inaudible. There are reasonable, though very unlikely, conditions under which it could be.

G
 
     Share This Post       
post-15387537
Post #18 of 62

Harold999

Head-Fier
Joined
Mar 14, 2018
Messages
55
Reaction score
10
Location
Rotterdam
Joined
Mar 14, 2018
Location
Rotterdam
Posts
55
Likes
10
That's indeed a compelling argument. Thanks.


I guess you're talking about this https://habr.com/en/post/456182/ or someone discussing it. I have not seen evidence that the AAC signal was decoded and then re-encoded to AAC
Hearing incoming email notifications through the music looks like evidence.
This won't happen though if you put 'do not disturb' on.
 
     Share This Post       
post-15387603
Post #19 of 62

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Messages
21,049
Reaction score
3,323
Location
Hollywood USA
Joined
Nov 16, 2004
Location
Hollywood USA
Posts
21,049
Likes
3,323
Website
www.facebook.com
As always, I’m talking real world, not absolutes. A null test may reveal some slight difference, but it doesn’t mean jack diddley to how it sounds. If you can transcode back and forth ten times and it takes a null test to discern a difference, one extra transcode for Bluetooth is a foolish thing to lay in bed worrying about. It flat out doesn’t matter. People should focus on things that makes their system better, not noodley crap that doesn’t matter.

I’m answering the real questions here. I don’t care how many angels can dance on the head of a pin. The problem with audiophile groups really isn’t that they don’t know the science. It’s that they don’t know what really matters. Crossing every T and dotting every I just adds to that obfuscation.
 
Last edited:
     Share This Post       
  • Like
Reactions: Sgt. Ear Ache
post-15392085
Post #20 of 62

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Messages
3,988
Reaction score
2,024
Joined
Feb 14, 2008
Posts
3,988
Likes
2,024
[1] As always, I’m talking real world, not absolutes.
[2] A null test may reveal some slight difference, but it doesn’t mean jack diddley to how it sounds. If you can transcode back and forth ten times and it takes a null test to discern a difference,
[3] one extra transcode for Bluetooth is a foolish thing to lay in bed worrying about.
[3a] It flat out doesn’t matter. People should focus on things that makes their system better, not noodley crap that doesn’t matter.
[4] I’m answering the real questions here. I don’t care how many angels can dance on the head of a pin.
[5] The problem with audiophile groups really isn’t that they don’t know the science. It’s that they don’t know what really matters. Crossing every T and dotting every I just adds to that obfuscation.
1. No, you're not! The real world INCLUDES absolutes, sometimes those absolutes are way beyond audibility and we can dismiss them as far as real world consumer listening is concerned but sometimes those absolutes can be within or just within audibility. In the latter case and if you personally don't encounter the conditions that reveal the "just within audibility", then you are ONLY talking about your personal world, NOT the actual real world and you are in danger of doing exactly what you accuse audiophiles of doing!!

2. You're not reading what I wrote bigshot! Firstly, I did NOT say "it takes a null test to discern a difference", I actually said I could discern with a blind listening test and that a null test just "confirmed" a difference. Secondly (and again!), perceptually lossy codecs do NOT just blindly apply the same processing regardless of what you throw at them (unlike for example a DAC). A perceptual codec analyses the input data and applies DIFFERENT processing depending on the characteristics of the input signal. Therefore, "if you can transcode back and forth ten times" and the result is audibly transparent, all you've demonstrated is that particular input signal/song is audibly unaffected but you have NOT demonstrated that another track, with different characteristics wouldn't be!

3. Probably.
3a. It flat out obviously doesn't matter to YOU but then by definition your assertion is false! Doesn't matter to you personally is incompatible with "flat out doesn't matter", unless it doesn't matter to everyone else as well but you haven't established that is the case. This is the science sub-forum, not the "What doesn't matter to bigshot" forum!

4. That's not true bigshot! Firstly, we are obviously NOT dealing with "angels", angels don't exist but we're dealing with real, objectively measured differences that do exist and could fall within the range of audibility. Secondly, as is your want, you're answering the question with your personal experience/impressions. The vast majority of the time that isn't an issue in this sub-forum because your impressions agree with the science/reliable evidence. For example issues related to jitter, cables, hi-res, etc., where the measured differences are well below audibility, there's no rational mechanism explaining how it could be audible and there's a significant amount of reliable empirical evidence that demonstrates it is inaudible. However, that not the case here, so all we're left with is your personal impression; fairly reliable testing of certain musical tracks, with your system and your hearing. I did a fairly reliable test too and I could (just) detect a difference and there is a rational mechanism that explains how it could be audible. So should I therefore state that it IS audible and that it "flat out DOES matter"? No, this is no more the "gregorio's impressions" forum than it is the "bigshot's impressions" forum, this is the Science forum! There is very little direct reliable evidence/science (due to the interplay of a large number of variables) that I'm aware of, and extrapolating from the science/evidence (that I am aware of), I'm confident in saying it's very unlikely to matter. However, I'm open to reliable evidence that it does matter some of the time or a body of reliable evidence that it "flat out" never matters (under reasonable listening conditions), do you have any?

5. Again, how do you know that it doesn't matter? Your test demonstrates it doesn't matter to you with your test file/s but where's the supporting evidence that it "flat out" doesn't matter with other files and is inaudible to others under other reasonable listening conditions? And, "crossing every T and dotting every last I" is a requirement of science! It's also what defines the difference between say vinyl and CD so in that case you would agree that it's valid but in this case you think it's just "obfuscation" because it disagrees with your impressions, doesn't this make you guilty of exactly what you accuse misguided audiophiles of??!

G
 
     Share This Post       
post-15392664
Post #21 of 62

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Messages
21,049
Reaction score
3,323
Location
Hollywood USA
Joined
Nov 16, 2004
Location
Hollywood USA
Posts
21,049
Likes
3,323
Website
www.facebook.com
Minute differences and rare worst case scenarios are only useful for the fraction of a percent of people to whom they apply. The vast majority of people fit within the fat part of the bell curve. Spending five paragraphs outlining one-in-a-million trivia in detail doesn't help very many people at all. And burying the single sentence that says "But of course to most people none of this matters." deep down amid reams if irrelevancies is doing a disservice to the majority of the public who just wants a direct and simple answer to their question. I usually start with the simple answer first and if the questioning goes further, that's the time to go in depth. But it's your style to do it the other way around and that's fine. I'm here to follow up and weed out the simple answer that they can easily put into practice to solve their problem. It all works out.

The road to clarity is proportion in all things. Balance matters. "Sometimes a cigar is just a cigar." -Siegmund Freud
 
Last edited:
     Share This Post       
post-15392753
Post #22 of 62

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Messages
21,049
Reaction score
3,323
Location
Hollywood USA
Joined
Nov 16, 2004
Location
Hollywood USA
Posts
21,049
Likes
3,323
Website
www.facebook.com
One illustration that might explain why I take the approach I do...

I have a good friend who has OCD. It goes beyond "Did I lock the front door this morning?' and it can send him into loops of double think that completely stymies him. Since we're friends, he trusts me. If I tell him he doesn't need to worry, he can take my word for it and stop worrying. But how I say that to him makes a big difference.... For instance, if I say, "In 99% of cases, there is nothing to worry about, but if any of these rare worse case scenarios exist, then it could be very serious. But I would bet that odds are you are fine not worrying." that will send him into a feedback loop of OCD that never ends. He'll think about the 1%, not the 99%, and he will try to think of reasons why his situation might be one of those rare worse case scenarios. He'll think about the odds and betting and imagine himself on the wrong end of the stick. However, if I say, "In your case, there is nothing to worry about. Let me know if you have problems, but I'm sure you are fine." then he calms down and remains alert to problems but he's able to move forward.

I see this same sort of OCD among audiophiles. They wonder if they can hear a difference between high data rate lossy and lossless, and even though they don't think they can hear a difference, they worry about whether they are losing "potential sound quality". If someone tells them that one kind of bluetooth involves two encodings instead of just one, they start worrying that they might be degrading their signal and focus on that even though the degradation probably falls well below the threshold of audibility. They look at the specs for their amp that has .01% THD and see another amp advertised that has .001% THD. They start worrying that their amp might be inferior. Is a noise floor of -96dB low enough or do I need a higher bit rate? The marketing of home audio is directly targeted at this kind of OCD. They are using this as a wedge to sell things to people that they don't actually need. The salesmen won't flat out tell them, "You can't hear this. It doesn't matter." They tell them "You may not be able to hear this BUT... Listeners who consider themselves experienced have reported hearing this, and they say it matters." That is factually correct, but the skew on it is all wrong. Do you see how they are trying to avoid a useful answer and planting a feedback loop of thought there?

So someone stumbles into Sound Science after marinating in "what ifs" and "maybes" in the commercial audio forums. What are they looking for? They are looking for science to be applied to clearly address the seeds of nagging doubt implanted in them by high end audio salesmen. They aren't looking for a dissertation on all of the exceptions to the rule. They just want to know if their stereo is OK or not. Once they're given the simple and straightforward answer to that question, they can either 1) move forward and apply that information secure in the knowledge that they don't need to worry; or 2) they can ask for more details if they are interested in figuring out how the nuts and bolts of it works. Only a few people will fall into that latter category, and they are the most interesting to speak with, but you can't make a person who just wants a simple answer into that kind of person. If you try, the only thing you will accomplish is to make them even more confused and worried. You might be thinking you are exercising "scientific diligence" by crossing every T and dotting every I, but all you end up doing is accomplishing exactly what the high end audio salesmen are trying to do.

I may be attuned to people's psychology more than other people. The reason is because I read between the lines to see what the emotions and desires are underlying the question (kind of like what the Dale Carnegie course preaches). Sometimes I see people here being thrown for a loop by answers that are factually correct, but don't address the real point of the question. That's when I speak up and say, "In your case, there is nothing to worry about. Let me know if you have problems, but I'm sure you are fine." If they continue to fret, I say, "If you're worried, do a simple listening test. It doesn't need to be clinically perfect. It just needs to be relatively free of bias and perceptual error. If you do the test and you have to strain to hear any difference at all, it doesn't matter. You can safely forget it and move on to things that might actually have a bigger impact on improving the sound of your system." That is advice that might help them focus on things that count instead of getting bogged down with things that don't. Facts without clarity of purpose don't help.
 
Last edited:
     Share This Post       
post-15393519
Post #23 of 62

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Messages
3,988
Reaction score
2,024
Joined
Feb 14, 2008
Posts
3,988
Likes
2,024
[1] Spending five paragraphs outlining one-in-a-million trivia in detail doesn't help very many people at all.
[2] And burying the single sentence that says "But of course to most people none of this matters."
[2a] deep down amid reams if irrelevancies is doing a disservice to the majority of the public who just wants a direct and simple answer to their question.
[3] I usually start with the simple answer first and if the questioning goes further, that's the time to go in depth.
[3a] But it's your style to do it the other way around and that's fine.
[3b] I'm here to follow up and weed out the simple answer that they can easily put into practice to solve their problem.
1. And you have evidence to support the assertion that's it's "one-in-a-million" trivia do you, or did you just make that figure up to defend your position?

2. You didn't say "to most people none of this matters", you said it "flat out doesn't matter" and you don't have reliable evidence for that assertion. It's gross hypocrisy to require reliable evidence to support an assertion made by some audiophile but you can assert whatever you want without reliable evidence!!
2a. The majority of the public do not come to head-fi in the first place and far fewer still come to this sub-forum. Some/Many of those who do come to this sub-forum maybe are looking for a "simple answer" but this is not the "Simple Answer" sub-forum, it's the Sound Science sub-forum.

3. But you haven't started with the simple answer first, you've started with a made-up assertion that has no reliable supporting evidence and does not align with the actual science/facts.
3a. No, it's my style to try and present a layman's version of the actual science/facts but sometimes that layman's version has to be more detailed because the facts/science are not simple and sometimes don't lend themselves to a simplification (or I can't think of a valid simplification).
3b. And there is the difference between us! You'll "weed out" a simple answer even if it involves simply making-up assertions that are NOT supported by reliable evidence. That's NOT science and, it's exactly what you criticize others for! Using your own example: "They wonder if they can hear a difference between high data rate lossy and lossless, and even though they don't think they can hear a difference, they worry about whether they are losing "potential sound quality"." - We CAN "weed out a simple answer" here and state it "flat out doesn't matter" because we have a mechanism that suggests and explains transparency; a great deal of development and evolution of perceptually transparent codecs AND a significant body of reliable evidence that under reasonable listening conditions they are in fact transparent. Same with a noise floor at -96dB, which is below the noise floor of the recordings themselves, below the noise floor of even very good systems/listening environments, is outside the range of reasonable listening conditions, is typically NOT at -96dB anyway (recommended/standard practice puts it at -120dB) and there is a significant body of reliable evidence that supports the inaudibility of even a -96dB noise floor. And continuing ...
[1] The marketing of home audio is directly targeted at this kind of OCD.
[2] So someone stumbles into Sound Science after marinating in "what ifs" and "maybes" in the commercial audio forums. What are they looking for? They are looking for science to be applied to clearly address the seeds of nagging doubt implanted in them by high end audio salesmen. They aren't looking for a dissertation on all of the exceptions to the rule.
[3] I may be attuned to people's psychology more than other people. The reason is because I read between the lines to see what the emotions and desires are underlying the question.
[4] If someone tells them that one kind of bluetooth involves two encodings instead of just one,
[4a] they start worrying that they might be degrading their signal and focus on that even though the degradation probably falls well below the threshold of audibility.
1. It doesn't matter whether someone has OCD or not, they're not going to hear the digital noise floor of 16bit or differentiate a max bit rate modern perceptually lossy codec under reasonable listening conditions, even with OCD. Much of the marketing of home audio is therefore NOT "directly targeted at this kind of OCD" (and if it were I wouldn't have a problem with it), the reason I have a problem with it here in this sub-forum is because it directly contradicts the science/facts/reliable evidence!

2. And here is the nub of the problem. Sure, I agree they are probably NOT looking for a detailed/complex response that doesn't provide an absolute answer but I'll tell you what else they are NOT looking for: A simple absolute answer that's based on someone's opinions/made-up assertions that are NOT supported by science/a body of reliable evidence!!! Again, fortunately a lot of the time we can give a simple absolute answer that is based on science/reliable evidence, the fidelity of vinyl vs digital, the noise floor of 16bit, etc., BUT IF, as is the case here, the ONLY choice is between simple absolute answer that is NOT supported by science/reliable evidence OR a detailed response that doesn't provide an absolute answer then as this is the Sound Science sub-forum there is ONLY one choice. You said it yourself, "they are looking for science to be applied", not for personal impressions/opinions devoid of science. So please present the science/reliable evidence that you are applying (I'm certainly open to it), otherwise you're being a hypocrite and are guilty of exactly what you accuse others of!!!

3. This isn't the "bigshot will diagnose your OCD" forum or "The answers you need according to bigshot's assessment of your emotional state" forum, this is the Sound Science forum and the WHOLE POINT of science in the first place is to separate the actual facts from people's emotions and desires! No one is saying you can't provide your own opinions/impressions if you think it maybe beneficial, just that you don't present them as absolute facts (such as "flat out doesn't matter") unless supported by science/reliable evidence. And you, of all people, should know this, or does it only apply to everyone else?!!!

4. Two or more encodings using what codecs? You seem to be ignorant of the process actually occurring and therefore extrapolating false equivalences, that's not science, it's essentially the opposite of science!
4a. I agree that the degradation probably falls below the threshold of audibility BUT, that's just my personal opinion/impression, it is NOT based on reliable evidence (the application of science) and I cannot claim it "flat out doesn't matter" unless I want to be a hypocrite! In fact, my understanding provides a rational explanation for why it may sometimes fall WITHIN the threshold of audibility, which is why I'd be interested in science/reliable evidence either way.

G
 
     Share This Post       
  • Like
Reactions: bfreedma
post-15393674
Post #24 of 62

Sgt. Ear Ache

100+ Head-Fier
Joined
Apr 28, 2018
Messages
347
Reaction score
229
Location
canada
Joined
Apr 28, 2018
Location
canada
Posts
347
Likes
229
I've lost track. Has this now turned into a debate over whether or not files degrade (to an audible degree) over multiple compression iterations? Because that isn't what the thread was originally about...
 
     Share This Post       
post-15394129
Post #25 of 62

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Messages
21,049
Reaction score
3,323
Location
Hollywood USA
Joined
Nov 16, 2004
Location
Hollywood USA
Posts
21,049
Likes
3,323
Website
www.facebook.com
I've lost track. Has this now turned into a debate over whether or not files degrade (to an audible degree) over multiple compression iterations? Because that isn't what the thread was originally about...
I believe there was some sort of discussion over whether APTX encoded twice as opposed to AAC passing across intact, decoding on the other end. It's been my experience that stuff like that doesn't make any appreciable difference. Modern bluetooth implementations are perfectly fine for their purpose.

This isn't the "bigshot will diagnose your OCD" forum
You get that as a bonus at no added charge! I reserve the right to put two and two together and figure out whether people here have personality quirks. It would be difficult to not notice it a lot of the time!
 
Last edited:
     Share This Post       
post-15395504
Post #26 of 62

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Messages
3,988
Reaction score
2,024
Joined
Feb 14, 2008
Posts
3,988
Likes
2,024
I've lost track. Has this now turned into a debate over whether or not files degrade (to an audible degree) over multiple compression iterations? Because that isn't what the thread was originally about...
Essentially, that is indeed what the thread was originally about. AptX is a codec and MP3 is a codec, so if you stream an MP3 over bluetooth using aptX the MP3 is decoded, then encoded into aptX, transmitted and then aptX decoded by the receiver, so we have "multiple compression iterations" or more precisely, two compression iterations of different codecs (a transcode).

The OP then changed the conditions of his original post and subsequently brought in AAC as the original format (being converted to PCM, atpX encoded, transmitted and then decoded back to PCM) and then, the scenario of the original AAC being decoded to PCM, then a second encoding to AAC, transmitted in AAC format and then decoded back to PCM. So we now have two compression iterations of the same codec (AAC) and bigshot pointed out that multiple iterations of AAC 256VBR is transparent. There's two problems with that assertion: Firstly, it may not always be true and secondly, it's a false equivalency anyway! Bigshot is failing to consider the sophistication of the AAC format, that it contains all kinds of modules and options for different use scenarios and bluetooth presents a very different "use scenario" - For example, bluetooth requires far more robust error correction, low latency and low computational complexity (to conserve battery life). The second encoding to AAC (for bluetooth transmission) is therefore significantly different, more of the bitrate is used for error correction (and therefore the audio data bitrate is lower), the application of the psychoacoustic model is different to reduce latency (from several hundred milli-secs down to around 30ms) and the psychoacoustic model itself is less sophisticated to reduce computational requirements. These trade-offs must reduce fidelity (more than say two iterations of the same 256VBR encode) and plausibly within the range of audibility in some cases.

G
 
     Share This Post       
post-15395602
Post #27 of 62

Sgt. Ear Ache

100+ Head-Fier
Joined
Apr 28, 2018
Messages
347
Reaction score
229
Location
canada
Joined
Apr 28, 2018
Location
canada
Posts
347
Likes
229
Meh, for me it's all sort of a moot point as standard Bluetooth (SBC) is essentially operating at 320kbps and since nearly nobody can reliably tell the difference between 320kb mp3 and lossless even under controlled test conditions I don't even think the "improved" Bluetooth protocols are necessary given that normal everyday listening situations are far removed from controlled test conditions (which I think is basically the point Big Shot is making)...

I mean I'm reasonably certain that under normal real world listening conditions nobody can actually hear whatever degradation might occur by files being converted from mp3 to aptx or whatever anyway. When I've done controlled listening tests I find it difficult to hear any difference between compressed files above 192kbps and lossless so there's certainly a buffer there for a few bits to be lost converting from one compressed format to another if we're starting from 256kbps or above.
 
     Share This Post       
post-15395917
Post #28 of 62

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Messages
21,049
Reaction score
3,323
Location
Hollywood USA
Joined
Nov 16, 2004
Location
Hollywood USA
Posts
21,049
Likes
3,323
Website
www.facebook.com
^ Exactly. If you have a peculiar interest in trivia, you might want to say more than that; but for people listening to music in the home, it just doesn't matter. There are enough things that actually do make a difference to focus your energy on than to waste your time worrying about bluetooth encoding strategies.
 
Last edited:
     Share This Post       
post-15400798
Post #29 of 62

gregorio

Headphoneus Supremus
Joined
Feb 14, 2008
Messages
3,988
Reaction score
2,024
Joined
Feb 14, 2008
Posts
3,988
Likes
2,024
[1] Meh, for me it's all sort of a moot point as standard Bluetooth (SBC) is essentially operating at 320kbps and since nearly nobody can reliably tell the difference between 320kb mp3 and lossless even under controlled test conditions I don't even think the "improved" Bluetooth protocols are necessary given that normal everyday listening situations are far removed from controlled test conditions (which I think is basically the point Big Shot is making)...
[2] I mean I'm reasonably certain that under normal real world listening conditions nobody can actually hear whatever degradation might occur by files being converted from mp3 to aptx or whatever anyway.
[3] When I've done controlled listening tests I find it difficult to hear any difference between compressed files above 192kbps and lossless so there's certainly a buffer there for a few bits to be lost converting from one compressed format to another if we're starting from 256kbps or above.
1. That's the exact same false equivalency bigshot used! Bluetooth (SBC) at 320kbps and an MP3 at 320kbps are completely different codecs with completely different design goals. An MP3 at 320kbps is perceptually lossy codec, designed to be audibly transparent to human perception, SBC is NOT a perceptually lossy codec. It's like saying my car has a 1.6l engine, a Formula 1 car has a 1.6l engine and therefore performance should be similar. It's not just about the same 320kbps but how you get down to 320kbps.

2. Again though, this isn't the "What Sgt Ear Ache is reasonably certain of" forum, it's the Sound Science forum! We have to hold ourselves to the same standards to which we hold others and therefore, not only can I ask for reliable evidence to support your assertion but logically I must, because your conclusion/assertion appears at least partly based on a false equivalency.

3. Again, that may well be true when transcoding between perceptual lossy codecs but we're not dealing with transcoding between perceptual lossy codecs! And, how do you know you're only loosing "a few bits" (or a few bits equivalent)? A significant number of audio bits are lost due to far more robust error correction with SBC and aptX and in addition, a significant number of equivalent bits are lost to due to the requirement of reducing latency by a factor of 20 or more. So the question is: Is this enough of a loss of audio bits (and equivalent bits) to be audible? It's certainly plausible that it is (under certain reasonable listening conditions) but just as this isn't the "What Sgt Ear Ache is reasonably certain of" subforum, it's also NOT the "What's plausible to gregorio" subforum, it's the sound Science subforum and like you and bigshot, I have no reliable evidence!

[1] ^ Exactly. If you have a peculiar interest in trivia, you might want to say more than that; but for people listening to music in the home, it just doesn't matter.
[2] There are enough things that actually do make a difference to focus your energy on than to waste your time worrying about bluetooth encoding strategies.
1. Just repeating the same mantra/assertion without reliable evidence is exactly what many audiophiles do when they visit this subforum and exactly what you criticise them for, yet here you are doing exactly the same. How does that make you anything other than a hypocrite? A criticism I've seen levelled at this subforum is that it's not really a sound SCIENCE subforum, it's just another opinion based subforum, more or less the same as all the other subforums here, except with significantly different opinions. You seem determined to prove that criticism true!!

2. You do NOT know and have NOT produced any reliable evidence that it is only "trivia", doesn't "make a difference" and "just doesn't matter", all you know is that it doesn't matter to you personally. Sure, there are things that make a bigger difference but this isn't the "Only Mastering and Transducer" subforum or the "What bigshot thinks you shouldn't worry about" subforum, it's the sound Science subforum!!

G
 
     Share This Post       
post-15400917
Post #30 of 62

Sgt. Ear Ache

100+ Head-Fier
Joined
Apr 28, 2018
Messages
347
Reaction score
229
Location
canada
Joined
Apr 28, 2018
Location
canada
Posts
347
Likes
229
Yeah, that's all true. My answer to the original question posed in the title of this thread would be "scientifically-speaking, no aptx is not perfect for 320kb mp3. But it's highly unlikely you (or almost anyone else) would hear the imperfections under the conditions most of us listen to music."
 
     Share This Post       

Users Who Are Viewing This Thread (Users: 0, Guests: 1)

Top