Yes, but that’s not the relevant issue. You don’t justify removing something simply because it takes up space. You justify it by finding something that makes better use of the space. After all, the screen also takes up space, but removing it would be stupid because nothing would make better use of the space. And in the case of the iPhone X, I think they have not justified removing the 3.5 mm jack. FaceID notwithstanding. Furthermore, the original justifications by Apple (waterproofing, larger batteries) are clearly not incompatible with a 3.5 mm jack.
Finally, maximizing width/length/screen percentage while minimizing thickness/weight is an advantage, not a disadvantage. It’s actually what distinguishes the iPhone X from the cheaper iPhone models.
Sorry man, I just don’t think this is true, at all.
If you are hearing YOUR OWN voice, then you may (and I stress the may), be able to notice the delay. But for any other operation, there’s just no way you notice anything, because you run into this constantly throughout your normal operations in life. You constantly talk to people who are 10 feet away from you. At no point do you EVER notice the 10ms delay that it takes for the sound to travel those 10 feet. This just isn’t a thing that happens, to anyone, ever.
No dude, you can’t. Again, this happens all the time. Hell, I routinely talk to people in my office that are 20 feet away. There’s no noticable delay in how long it takes for stuff to get to your ears compared to what you are seeing. And you can clearly see someone’s lips at 20 feet.
On basically any iOS application, there is far more of a delay between you touching the screen and you getting auditory feedback, even if you are using a wired headset… just because of how the software works.
20ms is an essentially trivial amount of time.
Now, if you do some specific tests to try and establish what the human ear’s auditory framerate is? The amount of distance between two tones where they perceive them as distinct tones instead of just a single one? THEN you might start to be onto something, but even there I’ve seen stuff suggesting that the average human can’t perceive a difference in tones that is less than 30ms.
Dude, this isn’t even remotely true.
This isn’t like your brain works with steroscopic vision. In that case, yes, inconsistent signals will create massively disconcerting effects, because your eyes are specifically designed to provide fairly precise distance estimations.
Your auditory system does not function like that. It’s not possible for your auditory system to function like that. Due to the way your ears are positioned, and the way that sound works including differences in volume, you are not able to do a steroscopic range test like bats do. Humans just don’t have that ability, or if they do, I have never seen any scientific literature about it, ever.
Your ability to judge the distance of sound is far more tied to relative volumes and other acoustic effects that are almost entirely based upon the specific environment the sound is taking place in. The stereoscopic nature of your auditory system provides bearing, not range information.
Yeah, it says you can’t lead by more than 15ms, but you can trail by up to 45ms… because trailing by 45ms is largely imperceptible at all. And we’re talking specifically about lip syncing audio at this point.
It’s faster on the leading edge, because it’s far more likely that you will notice things like someone’s voice while they aren’t moving their lips at all… but once they are actually talking? Virtually none of the normal population can detect a sync that is off by less than 45ms. Because you have very little ability to perfectly match the phonemes of sound to lip movement for an arbitrary person. Hell, the lip syncing in any videogame is going to be orders of magnitude worse in this regard than anything you’re going to run into in a 45ms delay.
That’s why the standard is that. And that’s why a delay of less than 45ms largely doesn’t matter much. Hell, it used to be that they intentionally added a fixed 2.5 frame delay to audio on television, always. That’s about 100ms.
Again, you’re constantly encountering audio delays in your daily life, in everything from using basically any piece of software on your phone, to simply hearing sounds. It doesn’t significantly impact you.
I’m not saying that the human ear can’t technically perceive a very small difference in audio delay… but such a thing is essentially just ignored by your brain, because you constantly encounter other natural effects that result in more significant delays anyway.
They did that in the early-digital-TV era to match the video more closely to the audio, because of frame sync. Without the 2.5-frame delay, audio was ~60ms ahead of the video, which was noticeable and problematic.
Wow this was a long discussion about an inconsequential point. It still remains the case that wireless headphones introduce enough variable lag that you can’t play rhythm games effectively. And it still remains the case that in some cases (e.g. multiple wireless relays like Roku->phone->headphones) there is enough lag that lip syncing issues become intolerable. Why fixate on this 45ms thing?
Also, yes you can detect a delay of 45ms. Though, as you said, you don’t use interaural time delay to determine distance to a sound, its bearing accuracy is up to 1°, and can detect differences in arrival time of 10 microseconds. Range determination is much less precise, but part of that relies on distinguishing reflected sound from direct arrival, which can again be done with a precision of tens of μs. I operate a laser system that pulses at 50Hz, which is 20ms per pulse. When the flashlamps fire (at 50Hz), it makes a clicking sound as the spark gaps in the Marx banks arc over. I can easily detect a single misfire in the 50Hz pulse train. It’s entirely possible that it simply doesn’t cause lip sync annoyance at that threshold as you say, but it’s certainly detectable. That said, as I pointed out above, it often IS above the threshold that causes lip sync annoyance, whatever that threshold is.
Yeah, being ahead of the video is very bad, so they erred on the delay side intentionally.
There actually are quantitative studies on this stuff. This study covers what’s probably the most hard core test case where audio latency will matter more than anything, which is live monitoring while playing an instrument.
The variation in effects tend to be largely driven by instrument, but in most cases, even in this super hard core case, latency times of 35ms still result in an experience subjectively classified as fair to good.
In normal situations where you are merely an observer of the sound? it’s unlikely it could impact your experience much.
Ya, sorry, I just started digging into it as I read some of this stuff, and got interested in the scientific aspects.
Oh, absolutely. Your ability to detect a delay difference between each of your ears is SUPER precise, because that’s what they’ve evolved to do.
With regards to distance, I wasn’t talking about stereoscopic cues. There are other cues the brain uses to judge distance of a sound, including visual estimates. In normal settings, the audio and visual cues are concordant. But if those cues are not concordant due to an artificial audio delay, then your brain can report a mismatch between the audio and visual stimuli.
dot clusters presented with a sound delay were judged to be more distant than dot clusters paired with equivalent sound leads. In the second experiment, we confirmed that the presence of a sound delay was sufficient to cause stimuli to appear as more distant. Additionally, we found that ecologically congruent pairing of more distant events with a sound delay resulted in an increase in the precision of distance judgments
In a simultaneity judgment task we presented a large range of stimulus onset asynchronies corresponding to distances of up to 35 meters. We found an effect of distance over the simultaneity estimates, with greater distances requiring larger stimulus onset asynchronies… These findings reveal that there should be an internal mechanism to compensate for audiovisual delays, which critically depends on the depth information available.
In other words, the fact that you don’t notice a delay when talking to someone across a room is not evidence that it is impossible to perceive that delay, because your brain compensates for perceived distance when deciding whether sound localized from a particular object.
The acceptable limit depends on the medium. For film, the limit is 22 ms.
Modem Android phones and good headphones support AptX Bluetooth which actually has lower latency (45ms) than wired headphones (60ms).
OK, this is clearly nonsense. You’re telling me that a cable connected to a driver connected directly to the phone’s audio hardware can somehow have more latency than the exact same thing, but with an additional digital wireless protocol interposed between them? Bullshit. The fact that you could even make this claim renders anything else you say on the topic suspect.
Bluetooth is mostly tolerable and I haven’t used wired headphones for ages, but honestly? Bluetooth skips, and degrades when you hold the phone wrong, and has significant lag and has done so for me even with AptX, for years. It’s kind of garbage and people put up with it because it does the job just well enough to prevent anything else from replacing it.
I still use the headphone jack occasionally and when I do it’s necessary and useful not having to carry a stupid bloody dongle. I hate that I’m forced to buy Samsung to do so, but given so many other manufacturers also think that I don’t need an ability to expand my storage, i wouldn’t gain that many more options if I stopped caring about it.
And the fact is, the headphone jack is a robust and well-developed standard extending back decades. It’s still good enough for professional audio, so I find the claims that it’s obsolete laughable. And people complaining about the wasted space - phones have been too thin to comfortably grip for several years now, so if they just added 2mm of thickness, I’m sure they could manage to fit a poor little headphone jack in, as well as more battery!
I’m not questioning whether your auditory system can perceive time slices that small. When detecting bearings to sounds, it’s based one differences that are smaller.
But the fact that it’s not noticeable in tons of normal situations means that it’s not going to be disruptive in situations that we’re talking about. All of those normal cues that your brain uses to do stuff are already fucked up at that point… you’re looking at pictures of tiny people on a flat screen a few feet from your face. There are already so many things which are conflicting with any biologically evolved mechanism that it demonstrates that your brain doesn’t have any problem just dealing with it. Because in the real world, the way that sound is routinely affected by the environment, volume disparities, distance, etc… while there is actual information in the audio signal that reaches your brain that includes all of that stuff, there are so many possible variations of the input that could result in that signal, that your brain doesn’t freak out about any of it.
It’s only when YOU are producing the sound that you start to run into significant problems, because your brain is very specifically listening for audio feedback that is directly related to your own actions. When that sort of delay exists, you start to have major problems. But again, as that study above presents, even in that case a delay of 35ms isn’t going to cause a major problem. And anything you are doing other than live audio monitoring of your own performance is going to have more room for latency without affecting the experience.
It’s not the same as the kind of instantaneous, massive problem that is caused when, say, you screw up the z-draw order of things rendered in VR space. Because the difference in perception between your eyes translates to a very specific distance, and that distance correlates to very specific properties in relation to obscuring other objects or being obscured. When those two factors are in conflict, it creates major problems for your brain. (for me, it causes an almost instantaneous headache)
For an auditory effect, to achieve this kind of problem, you would need to somehow delay the signals reaching each ear, differently. This would result in audio hitting you that suggested a source at one location, an then potentially have a visual indicator of the source being somewhere else.
Although, in practice, when you have exactly this happen (it’s fairly easy to do, through misconfiguring of surround sound systems), the effect is nowhere near as disruptive as the visual issues described above. Even with significant disparities in the binaural signal from the visual source, the effect ends up being somewhat subtle. It’s easily recognized, but it doesn’t cause the violent reaction that screwed up z-order in VR does.
Could you provide a source that explains the basis of this value? I’d be interested to see what resulted in that number.
I have B&O h9i’s, a beoplay A1 and Sony SRS X88. For phones, I have an lg v30 with a quad dac and a Sony Xperia xz1 compact.
The H9i and beoplay A1 only support AAC, the SRS-X88 supports aptx hd and LDAC.
The H9i also accepts a 3.5mm connection.
The H9i with the LG v30 tricked into high impedance mode blows away playing on Bluetooth via AAC, the beoplay A1 with aac sounds better than the srs-x88 playing with either LDAC or AptX HD.
As for headphone poets, any explanation I’ve read for removing them was about reducing the depth of the phone, not freeing up space for more battery or whatever. I mean, they tend to be in a part of the phone that in teardowns isn’t necessarily utilised for much else!
(Despite being a lower aitrate than AptX, aac is a higher quality due to compression - much higher latency though)
That did not measure the latency of playing the sound. It measured the latency of the touch screen detecting the touch, the event being passed through the full software stack (OS and application), and then the latency in the reverse order of the application initiating playing a sound, that action passing through the whole software stack, and then the hardware actually playing the sound. The latency for the first two parts of that chain is going to be on the order of 40-80ms on a modern phone.
Sorry, I don’t buy your “Bluetooth headphones with less latency than wired” theory.
These are iPhone tear downs? Apple took out the port with the 7, if I recall correctly. That’s the same model that had the taptic engine component introduced in that spot. At least with respect to iPhones, I haven’t seen a ton of extra internal space in tear downs.
Android has had terrible UI->audio latency for a long time, as I discovered when I was working on a Morse code input method. Vibration is just about responsive enough to work with, but the beeps threw me off entirely.
My wife and I have been using a pair of Samsung Galaxy S8’s for almost 2 years and generally like them. However, within the last couple of weeks she managed to break the screen on her phone and the USB port on the bottom of my phone is starting to fail. Local shops seem to want about $300 to fix both of them (mostly for that curved screen). Fortunately both phones are completely paid for so no worries about having to pay a balance on a broken device. So for the last week or so I have been doing some research into whether to fix or replace our phones and pulled the trigger on 2 Moto G6 phones from Google Fi this morning. Here was my logic:
Cost of 2 new Moto G6 = $199 each plus tax = $430
Estimated trade-in value of both phones in their current state = $304 ($186 for mine and $118 for the one with a broken screen)
Google is offering $100 in bill credits when you activate a G6 on Fi (I think I only get one of these credits despite buying 2 phones as the terms say “only one per person”)
So assuming the trade-in and bill credits work out I will get 2 new phones for $430 - $304 - $100 = $26. Granted the G6 is a bit of a downgrade from the S8 in terms of specs, but I think the G6 offers some definite advantages of it’s own such as:
No Samsung bloat-ware
Cheaper to fix/replace in the future
No curved screen means cases that can have a lip around the entire edge of the screen and screen protectors fit better and are easier to apply
Same sized battery (3000mAH), but lower-end specs = possibly better battery life
Based on our data usage Google Fi will be at most roughly the same price as our current t-mobile plan and possibly much cheaper if we curb our data usage somewhat
Neither of us are really demanding on our phones. We don’t play any demanding games. We don’t take a lot of pictures or video. As long as it can do texts, emails, web browsing, play music, and stream Netflix/YouTube we are happy. While there is certainly plenty of room to grow as far as the cutting edge of smart phone performance, I think we are swiftly approaching the territory of “good enough for most people” that desktop PC’s have been stuck in for about a decade. The Moto G6 seems like it will fit our most common use-cases perfectly for significantly cheaper than even fixing our existing flagship phones. Forget about paying for new ones!