http://pic.gsmarena.com/vv/reviewsimg/htc-hero/gsmarena_a001.png - this frequency responce graph doesn't look promising. I know that there are ways to correct it a bit on the Dream (where it is a little bit better to start off), but how about the Hero? I'm really considering of buying this phone, and the sound quality stuff could be a real deal breaker for me.
So I would appreciate any feedback on this, thanks.
honestly, ive never really noticed anything bad about the but ill check against a mates iphone 3g and report on the results, but i may not see him till next week
If you are not an audiophile like me, you could miss the points. What I am interested in, is the highs - close to 20KHz.
I would like to believe that the frequency responce graph I mentioned earlier does not represent how Heroes sound like. It's so choppy and lacks highs - they cut of at like 15KHz or so. Also, why is that graph so choppy? iPhones are just invincible in this aspect..
So I was just wondering if someone had managed to fix this with software modifications, especially after seeing threads like this.
Connected my hero to a decent separates system and it sounded great. It blew my friend's N95 away comparing the same tracks!
1. Check out the vertical axis on that graph. Considering that 3dB is about the smallest increment/decrement in intensity that the human ear can readily distinguish, that frequency response is actually impressively flat for a consumer device.
2. Check out the frequency response of your earphones. Or, if you hook it up to your expensive hi-fi sound system, your speakers. Your frequency response is NOT going to be limited by this device but by the speakers/headphones. To see what I mean, compare that graph, which for the most part stays within +/- 0.5 dB, with this one of a pair of good quality earphones, or this one of some excellent speakers. Note the peaks and troughs in excess of +/- 7dB in the earphones, and +/- 3dB or so in the speakers.
3. From the look of the graphs, I doubt they've used the same method to test both. I don't think you'd find them that different. Maybe one is a manufacturer's stated frequency response and the other isn't, or they were tested with different types of equipment.
Although not an audiophile myself I am pretty keen on getting the best sound quality I possibly can. Before I purchased the Hero I plugged in my own cans to check frequency response. I was actually pleasantly suprised how much depth was in the music.
Not immediately noticable however was a strange noise audible at the start and end of tracks, and in quiet sections - kind of like a hiss. Although this would probably only be audible in a decent set of cans (with a wide frequency response), as an audiophile where only the best sound will do this would surely annoy the hell out of you.
The other point to note is the lack of EQ on the Hero's music player, but as an audiophile I'm sure you would frown upon using EQ anyway.
g00nerz said:
Not immediately noticable however was a strange noise audible at the start and end of tracks, and in quiet sections - kind of like a hiss. Although this would probably only be audible in a decent set of cans (with a wide frequency response), as an audiophile where only the best sound will do this would surely annoy the hell out of you.
Click to expand...
Click to collapse
I noticed that on my previous S60 Nokia but don't notice it on my new Hero. It's possible that differences in impedance or sensitivity between headphones might make this affect more noticeable on some headphones than others.
The other point to note is the lack of EQ on the Hero's music player, but as an audiophile I'm sure you would frown upon using EQ anyway.
Click to expand...
Click to collapse
Absolutely right!
I wouldn't call myself an audiophile (who would?) but I do have a keen interest in the technical aspects of audio and equipment.
this is the only phone where i can say the audio is really bad. i have yet to try more than one set of wired headphones but the bt headphones i have make it sound much worse. i have had a few htc phones and this is the worst by far. mono actually sounds better than stereo. Maybe im just used to my e71's sound quality.
Oh Lol.
Audiophiles listening to music on their phones. There's absolutely no way anyone who genuinely has a critical ear would expect anything like high end (as in quality) reproduction of audio on a phone.
Either way, the Hero's got a crappy signal to noise ratio so best you move on before you get offended.
I wonder how accurate the test was for the 3GS because I have never seen a frequency response graph that flat (even my HiFi tube amp has more variations).
And if you examine the hero graph more closely, you can notice that it's really very good. It has minimal gain loss in low freq. and it spans almost linearly up to 16kHz. And that's like very good. I noticed myself that the sound quality was far better then on the iPhone. But that is just my subjective opinion. But you can take in consideration that I used some HiQuality AKG headphones to test them both.
LP
B
flexte said:
this is the only phone where i can say the audio is really bad. i have yet to try more than one set of wired headphones but the bt headphones i have make it sound much worse. i have had a few htc phones and this is the worst by far. mono actually sounds better than stereo. Maybe im just used to my e71's sound quality.
Click to expand...
Click to collapse
Any audio quality problems with bluetooth headphones/headsets have nothing to do with the hero, because the Hero is not actually producing the sound you hear, it's just following the Bluetooth spec, so it's not relevant to the graphs posted at the top of this thread.
Just so we're clear. You may already have known this.
oblika said:
And if you examine the hero graph more closely, you can notice that it's really very good.
Click to expand...
Click to collapse
Just as I said above. I think the 3GS graph there is bogus; there's no way that was measured using the same equipment. Most likely it is the 'theoretical' frequency response and the Hero's is 'measured'. And when you look at the Hero graph and pay attention to the Y axis, it is actually a very good, flat frequency response.
I have tried configuring Adapt Sound, but I can barely hear the bips: is this normal?
thegios said:
I have tried configuring Adapt Sound, but I can barely hear the bips: is this normal?
Click to expand...
Click to collapse
Yes that's the whole point.
It will adjust sound based on the tone frequency wavelengths you can and can not hear. Hence it asks can you hear this...
Sent from my SM-G928F using Tapatalk
Yes this I understand but they are really barely audible...
They are meant to be barely audible. Imagine if everyone could hear them all... it kinds defeats the issue, as there would be no need to adjust the sound.
The bips vary by frequency, volume and channel. This way, adjustments can be made based upon your personal experience. I was surprised to see how the results curve differed between my left and right ears. If for example you can not hear many of the higher frequencies, Adapt Sound can make adjustments to boost this range to compensate.
Another factor to consider is how as we get older, our ability to hear higher frequencies diminishes. Younger people hear a broader range than older people.
All of the bips are quiet, as if they were loud, like I said before, no adjustments would be needed. The more honest you are in registering the sounds, the better job Adapt Sound can do to improve your listening experience. If you cheat, you are only cheating yourself.
Ok I will give it a try in a very quite place.
Does this work overall (google play music for example) or only in some samsung apps?
How does this affect SoundAlive?
For example I can access SoundAlive from Play Music but not from Settings, and I can access AdaptSound from Settings but not from Play Music.
Could you please explain what works with what?
nobnut said:
They are meant to be barely audible. Imagine if everyone could hear them all... it kinds defeats the issue, as there would be no need to adjust the sound.
The bips vary by frequency, volume and channel. This way, adjustments can be made based upon your personal experience. I was surprised to see how the results curve differed between my left and right ears. If for example you can not hear many of the higher frequencies, Adapt Sound can make adjustments to boost this range to compensate.
Another factor to consider is how as we get older, our ability to hear higher frequencies diminishes. Younger people hear a broader range than older people.
All of the bips are quiet, as if they were loud, like I said before, no adjustments would be needed. The more honest you are in registering the sounds, the better job Adapt Sound can do to improve your listening experience. If you cheat, you are only cheating yourself.
Click to expand...
Click to collapse
omg look who is here !!!
:victory:
didn't saw you 100 years
Introduction
I created this thread to post and discuss measurements and data related to audio fidelity of the headphone output. Note that audio fidelity does not automatically correlate with audio quality. Audio quality is a very subjective measure. Some people like their audio to be very bass heavy while others prefer accentuated mids and/or highs. Yet others (like myself) enjoy a very "neutral" sound signature. Audio fidelity -- at least the way I define it -- on the other hand can be measured: Audio data is fed to the DAC/amp and measured at the output. For perfect fidelity, the recorded output is identical to the input. Of course, this ideal cannot be achieved. Deviations from the ideal can be measured, documented, and discussed. Particularly, not all deviations from the ideal are actually audible. I predict some heated discussions on audibility in this thread going forward...
Test Equipment
The device under test (DUT) is a Pixel XL that sports a Qualcomm WCD9335 codec/amp discrete audio chip. The headphone output is split with a Y-wire. One signal path goes into the measurement rig, a Rohde & Schwarz UPV Audio Analyzer. Depending on the measurement, the other end either stays open (not connected or N/C) or connected to my Sennheiser IE800 IEMs. Those IEMs have a virtually frequency-independent impedance of 16 Ohm, a sensitivity of 125 dB/Vrms at 1 kHz, and a ruler-flat phase response. These IEMs are not only one of the best "sounding" headphones I know, but actually very easy to drive as the specs that I've listed suggest.
Test Signals
For the tests conducted so far, I have used two signals that I created with Matlab:
stereo wave file with full-scale (0 dBFS) 1 kHz sine wave in the left channel and silence (zeros) in the right channel (16 bit, 48 kHz sampling rate)
stereo wave file with white noise mastered to take advantage of the full dynamic range (16 bit, 48 kHz sampling rate)
These signals were played with the stock Android music player, with all audio effects disabled.
In some cases I swapped the Pixel XL with an iPhone 5s for comparison purposes. All other hardware stayed the same.
Measurement Results
This section summarizes the measurements detailed in the second post.
Output Impedance
The Pixel's output impedance was measured to be 4.8 Ohm. To compare, the output impedance of the iPhone 5s' amp is 2.0 Ohm.
Comments: The Pixel's output impedance violates the often-quoted 1/8 rule in some cases. The rule says that most amps work best with headphones that have an impedance that is at least eight times its own output impedance. According to this rule, headphones should ideally have an impedance of more than 40 Ohm. Nevertheless, the amp seems to work with my 16 Ohm IEMs without any issues (other than the level drop, see below). The reason, I think, is that my IEMs make use of a single dynamic driver rather than multiple balanced armature ones like many other IEMs. Multiple balanced armature drivers are much harder to drive and may cause minor problems (slight modification of the frequency response) with the amp in the Pixel
Frequency Response (see plots in the second post)
The frequency response is ruler flat, whose shape does not change at all once my IEMs are plugged in as a load. The gain dropped by 2.3 dB after plugging in the headphones. The reason for this drop is the interaction of the headphone's impedances with the output impedance of the amp. Higher impedance headphones will see a smaller drop in volume when plugged into the Pixel.
The channel imbalance is a negligible 0.1 dB at 1 kHz.
Comments: As discussed above, the only potential issue I can see here is that the combination of this amp with low-impedance multiple armature-based IEMs may lead to slight modifications in the overall frequency response, maybe on the order of +/- 2 dB max.
Maximum output level
When the volume on the Pixel is turned up all the way, the maximum attainable output voltage with a full-scale 1 kHz sine wave is 0.427 Vrms, without any load attached. I have found that the output volume was capped digitally to -7 dBFS in the mixer by Google. The maximum theoretically attainable output voltage of virtually any DAC/amp combo devices is 1 Vrms. Once a custom recovery has been made available for the Pixel, it should be trivial to boost its maximum output voltage to 1 Vrms.
Comments: No issues here.
Volume level steps with volume rocker
As discussed above, at max volume the Pixel delivers 0.427 Vrms, which is volume step 15.
Lowering the output volume from the maximum setting, 15, with the volume rocker results in the following steps (rounded):
Volume setting: 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0
Change in gain: 0dB, -3dB, -3dB,-3dB,-3dB,-3dB,-3dB,-4dB,-4dB,-3dB,-4dB,-4dB,-4dB,-7dB,-7dB,-54dB
Noise Floor
The noise floor was measured to be 115 dB, which gives a theoretical dynamic range of (currently) 108 dB.
Comments: No issues here.
Total Harmonic Distortion + Noise (THD+N, see plots in the second post)
Measurements of THD+N were done on both iPhone 5s and Pixel at their respective maximum volume as well at a volume setting that corresponds to an output voltage of around 20 mV, with which the IE800 would produce around 90 dB SPL:
Pixel at 20 mV: THD+N is -71.4 dB
Pixel @ max volume: THD+N is -81.3 dB
iPhone 5s at 20 mV: THD+N is -78.7 dB
iPhone 5s @ max volume: THD+N is -84.1 dB
Comments: By comparing the plots in the second post, it can be seen that the Pixel has higher non-linear distortions than the iPhone 5s, especially at low output voltages. The reason for this is that Google and I believe any other Android manufacturer opts to maximize analog gain while controlling overall level with digital gain only. This is not ideal. It would be better to hit the codec with the highest digital gain possible (but no more than 0 dB), keep the analog gain low initially and control overall loudness with the analog gain only. This way the full dynamic range of the DAC can be utilized. This is the paradigm that virtually any home/car stereo systems utilizes. Based on the plots below, I'm guessing that Apple is doing it, too.
Intermodulation Distortion
Preliminary inspection of SMPTE IMD measurements suggest no problems either unloaded or loaded with my IEMs. Measurements to follow sometime next week.
Stereo Crosstalk
Stereo crosstalk very much depends on the interaction of amp, headphone, and analog circuit design of the phone. Here, at least on the surface, the Pixel does not seem to be showing strong numbers. With my headphones attached and throughout the volume range offered by the Pixel, I measured the stereo crosstalk to be 45 dB. I compared this number with the iPhone 5s. Interestingly, stereo crosstalk on the iPhone shows a strong dependence on the playback level. Again with the IEMs attached the crosstalk ranged from 43 dB at the lower volume settings up to 60 dB at the higher end. I repeated the measurements with full-size cans, the Sennheiser HD 540 (300 Ohm). With those headphones attached, the stereo crosstalk of both the iPhone 5s and the Pixel are north of 80 dB.
Comments: To put things into perspective, though, at the playback levels that in the long run do not cause permanent hearing loss (less than 90 dB SPL at the ear drum) with the IEMs, both iPhone 5s and Pixel have very similar stereo crosstalk performance, i.e. around 45 dB. Carefully controlled double-blind tests should be considered to determine at what point crosstalk actually becomes audible.
Comments on the Pixel mixer
There is one thing that strikes me as odd after having examined the Pixel mixer (/system/etc/mixer_paths.xml): The amplifier is set up as "CLS_H_LOHIFI" and not as "CLS_H_HIFI", which is the default for the WCD9335 in general and the HTC10 in particular. While I don't know what exactly the differences between those two settings are, I will play around with them as soon as a custom recovery becomes available. I'll also look into enabling hardware IIR filtering via the Qualcomm codec and update my biQuads app.
Supporting plots done with stock mixer can be found here.
Pixel "frequency response" at 20 mV output voltage. Note that the 2.3 dB level drop to to the presence of the IE800 has been compensated for (with this output voltage my IE800 delivers around 90 dB SPL):
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Pixel "frequency response" at maximum output voltage (0.427 Vrms). Note that the 2.3 dB level drop to to the presence of the IE800 has been compensated for (with this output voltage my IE800 delivers around 115 dB SPL):
Pixel THD+N at 20 mV output voltage (only the left graph is relevant):
Pixel THD+N at maximum output voltage (only the left graph is relevant):
iPhone 5s THD+N at 20 mV output voltage (only the left graph is relevant):
iPhone 5s THD+N at maximum output voltage (only the left graph is relevant):
reserved (for future measurements based on a modified mixer)
Damn. Thanks for doing a deep dive on this.
Greatly appreciate you taking the time to do the measurements.
I'm assuming Normal Pixel would have mirrors measurements as my listening observations are in line with what you have presented.
Although on my Pixel it seems like the Single loud speaker also has the same 'Boomy' bass boost applied.
I have Campfire Jupiters and Shure SE535 with Brown Knowles dampeners and the Shures just pair better with the stock Pixel.
Looking forward to future Audio tweaks on this device.
With both headphones I have been using and brothers Shure SE215: I have noticed that the lowest volume level is still what I would consider listening level. (Where as on Note 7 sound was not audible until about 25% volume from off)
Thanks for providing such quality measurements and interpretation.
Since the physical structure of the plain Pixel is different, it is possible the crosstalk is not the same. Would be interesting to compare if anyone has one to test.
mariano3113 said:
I'm assuming Normal Pixel would have mirrors measurements as my listening observations are in line with what you have presented.
Click to expand...
Click to collapse
bjrmd said:
Since the physical structure of the plain Pixel is different, it is possible the crosstalk is not the same. Would be interesting to compare if anyone has one to test.
Click to expand...
Click to collapse
I can imagine that the Pixel would measure exactly the same as the Pixel XL. It would make a lot of sense for Google/HTC to use the same board layout as a cost savings measure, even though the two devices have different physical dimensions.
chdloc said:
I can imagine that the Pixel would measure exactly the same as the Pixel XL. It would make a lot of sense for Google/HTC to use the same board layout as a cost savings measure, even though the two devices have different physical dimensions.
Click to expand...
Click to collapse
What I meant (and it was just a possibility) is illustrated here
If you look at the sony z5 regular vs compact, the crosstalk is 8 dB different. As you have mentioned, it's the grounding circuitry that determines crosstalk (not the dac) and perhaps a different physical layout makes a difference.
bjrmd said:
What I meant (and it was just a possibility) is illustrated here
If you look at the sony z5 regular vs compact, the crosstalk is 8 dB different. As you have mentioned, it's the grounding circuitry that determines crosstalk (not the dac) and perhaps a different physical layout makes a difference.
Click to expand...
Click to collapse
Oh, I apologize for being so vague in my response. As reported, Google/HTC spent only 9 months in designing the Pixels. On such a short development timeframe, it would make a lot of sense to clone as much of the hardware design as possible. Provided the circuit board carrying the codec physically fits into the regular Pixel's body (a big "if" as the regular Pixel is 6.2 mm narrower), I'm willing to bet that the two phones are identical from a headphone audio perspective. Otherwise, of course, all bets are off.
chdloc said:
Introduction
Output Impedance
The Pixel's output impedance was measured to be 4.8 Ohm. To compare, the output impedance of the iPhone 5s' amp is 2.0 Ohm.
Comments: The Pixel's output impedance violates the often-quoted 1/8 rule in some cases. The rule says that most amps work best with headphones that have an impedance that is at least eight times its own output impedance. According to this rule, headphones should ideally have an impedance of more than 40 Ohm. Nevertheless, the amp seems to work with my 16 Ohm IEMs without any issues (other than the level drop, see below). The reason, I think, is that my IEMs make use of a single dynamic driver rather than multiple balanced armature ones like many other IEMs. Multiple balanced armature drivers are much harder to drive and may cause minor problems (slight modification of the frequency response) with the amp in the Pixel
Frequency Response (see plots in the second post)
The frequency response is ruler flat, whose shape does not change at all once my IEMs are plugged in as a load. The gain dropped by 2.3 dB after plugging in the headphones. The reason for this drop is the interaction of the headphone's impedances with the output impedance of the amp. Higher impedance headphones will see a smaller drop in volume when plugged into the Pixel.
Click to expand...
Click to collapse
Just to be clear, the output impedance is a problem. Here is the frequency response driving Unique Melody Merlins:
More than a minor problem there.
Here it is with a 32Ω load.
For comparison, here is a Sony ZX1 with the same two headphones used as load:
With the Merlins, the ZX1 between 30Hz-15KHz is +.5/-1.25dB with the heavy roll starting at around 7KHz, not ideal, but not much music content up there.
The Pixel XL is around +1dB/-1.5dB, but with a ton of rolling in the middle. And that's what you will see with balanced armature drivers.
Also note that the hump at around 80Hz with the PS1000 load is around double what it is on the ZX1, but the PS1000 impedance goes a little nuts right there: http://www.innerfidelity.com/images/GradoPS1000.pdf
And a comparison of the Astell&Kern AK300, Sony ZX1, and Nexus 6 with the Merlins as a load. N6 was actually the best performer of the lot!
Reginalb124 said:
Just to be clear, the output impedance is a problem. Here is the frequency response driving Unique Melody Merlins:
Click to expand...
Click to collapse
Thanks for sharing your measurements!
For the sake of others looking at this, however, I think it is important to point out that the output impedance may be a problem and is by no means guaranteed.
Not all folks using this phone or looking to buying one use such high-demanding IEMs. Remember that my IEMs, the Sennheiser IE800, do not cause any measurable change in the overall frequency response, other than a frequency-independent level drop of 2.3 dB. I'm very satisfied how my IE800 sound with the Pixel.
As your own plots show, your IEMs cause "issues" in all configurations, even with the Astrell&Kern dedicated music player! Note that you are arguing over frequency response swing differences of less than 1 dB. Keep in mind that the amplifier-headphone system is largely linear so you can take the additional small frequency (and likely phase) swings due to the headphone-amp interaction and fold them into the already existing large swings of the headphones themselves (linear system theory).
My point is that those tiny frequency swings, +/- 1 dB, by themselves likely do not cause sufficient changes in the headphone's overall sound signature to be problematic. Unless the artificial bass boost causes the bass to become muddy.
The final word, however, will have to come after carefully conducted double-blind listening tests have been performed. I'm not going to making a lot of friends around here by saying that comparing the audio quality of two phones/devices without carefully matching the levels within a fraction of a dB and performing the tests blindly only result in subjective opinions without much merit.
My Shure SE535's sure love this phone, sounds a bit better to me than my Nexus 6.
My SE846 with an impedance of 9 ohms sound great, although I do use the built in EQ a bit to create a v-shaped sound signature which I prefer. Maybe I'll do a comparison with my Grace m9XX DAC/amp to see if I notice any major differences with a flat EQ on both.
The "problem" with low impedance headphones is not necessarily the low impedance as @chdloc has indicated. It's the variability with frequency (because each driver and crossover has different impedance effects) that can mess with the frequency. The two graphs below are from a Note 4 (1.4 ohm) with a Shure se535 (variable but higher impedance) and the se846 (variable low impedance).
Here is the impedance graph of the se846-note the variability:
As can be seen, the frequency shifts are higher with the se846--but if the headphone impedance was the same throughout, it would look flat.
The se846- more exaggerated:
The se535 (much higher impedance but still variable)-not too bad:
I have measured the same curve with my Dragonfly red (.02 ohm) and it is totally flat
I would guess the exaggeration would be more with the Pixel (5 ohm). This may sound "better" to some, but not to others-as noted above.
My plan (after I get my preorder) is to remeasure the frequency response and just design some biquads to "fix it" the way I like it. But, the average user who is not adjusting this may notice a different sound signature.
It is not a deal breaker, but I was a bit disappointed in Google/HTC for not doing a better job. My Note 2 was close to 1 ohm, S6 1.4 ohm and Iphone noted above.
bjrmd said:
The "problem" with low impedance headphones is not necessarily the low impedance as @chdloc has indicated. It's the variability with frequency (because each driver and crossover has different impedance effects) that can mess with the frequency.
Click to expand...
Click to collapse
It is actually a combination of a highly variable and low impedance. Higher impedance headphones with varying impedance will have less of an effect as the amplifier's damping factor -- which is a function of its output impedance -- will be able to "deal" with it.
But still, the number of users that have such high-demanding IEMs, like your SE846 or the AKG K3003, is likely very low. That's why I wanted to counter the general statement of doom and gloom expressed earlier in this thread. Having said that, you guys with demanding IEMs may want/need to resort to either utilizing correction filters, buying a different phone with a lower impedance amp, or investing in an external DAC and/or amp.
Thanks for the effort you put into this. I've got a couple sets of ciems from Noble Audio that I'm itching to try once my XL comes in.
GSMarena review up-no real surprises and basically agrees with above. Crosstalk -62 dB, not as good as HTC 10, Iphone but better than the LG V20 (special ESS dac).
All in all, with respect to audio a solid performer but certainly not spectacular.
Life is full of compromises.
This is from my Google Pixel (non XL)
Removed LO from LOHIFI (To my ears it seems to have reduced the 'Boomy’ bass)
About to test Elemental kernel.
Edit:
another XML edit: "RX HPH Mode" value="CLS_H_LP" compared to HTC 10 & Note 7 (Snapdragon) "RX HPH Mode" value="CLD_H_HIFI"
Before:
After:
mariano3113 said:
This is from my Google Pixel (non XL)
Removed LO from LOHIFI (To my ears it seems to have reduced the 'Boomyâ?? bass)
Click to expand...
Click to collapse
I've made that change, too, but I have a hard time hearing a difference (my headphones don't sound boomy to begin with).
Did you reboot after making the change? Measurements will have to wait until at least Monday.
Did you actually edit your mixer_paths.xml in /system or did you go the system-less route by binding a modified mixer in /su/etc?
BTW, biquad filtering, stereo recording, and a very first shot at dual loudspeaker playback seem to be working (the earpiece is a *lot* quieter than the main speaker, however), all done via mixer_paths.xml changes. I'm still having trouble, though, getting the biquads modification to work reliably when going the system-less route...
This is a great thread. I love getting the best out of my devices and sound has always been important to me. I'm no audiophile but I'll be keeping up with this. Thanks!
As an audio-lover coming from HTC's and later the Google Pixel devices, the Xperia 5 II got me very excited not only for retaining Sony's signature dual front-firing speaker setup but also for re-introducing the headphone jack. I understand that I am very much in the minority here with my wants and needs but, having access to a wealth of audio gear and competing audio-focused smartphones, I thought it would be helpful to draw some attention to one of the 5 II's best-performing features that hasn't been given justice yet online. Jump to the speaker quality section to skip the rambles.
To add context, I've owned the Xperia Z2 and loved it overall. However, while it ticked a lot of boxes, I was left dissatisfied with the performance of each feature - i.e. the cam was noisy and slow, the speakers very tinny, the headphone out lacking volume and I haven't returned to Sony until now. Until then, I was content with HTC's approach, their Boomsound stereo speakers were much richer and louder, the headphone jack implementations up to the U11 were also very good; the 10 especially had a dedicated audio section which delivered a very low noise floor, albeit a higher output impedance which made it a no-go for expensive multi-driver earphones.
Before the 5 II, I owned the Pixel 3 and 4, the Pixel 4 I found to offer the best speaker quality with good sound chamber design providing nice fullness and volume albeit at the cost of no headphone jack and abysmal battery size and life. So of course, audio is a priority for me. I have a wide collection of high-end earphones and headphones and like to travel light without an additional DAP, dongle or BT receiver. The 5 II is an audio lover's dream with regards to its speaker quality/volume/setup and headphone jack output. It also comes across as a substantially more refined Android experience than past Xperia's I've used to the extent that it rivals Pixel smartphones in fluidity.
Speaker Quality -
Though not the best on the market, of the devices I've owned, the Pixel 4 actually has the best speaker quality in terms of max volume, frequency balance and bass extension and I would say any user would be satisfied with the performance. The speaker setup on the 5 II is immediately more immersive and symmetrical - I'm not a huge fan of the popularisation of downward-firing woofer + earpiece tweeter as I find the volume discrepancy between the two distracting. The stereo setup is a big plus here for media consumption, and something I missed from my old HTC's and the PIxel 3. The 5 II, thankfully, also has far better speaker quality than previous Xperia's.
They are no longer tinny in the slightest with a nice body and balance to the midrange. With Dolby on, you even get a little bass punch in addition to a healthy jump in volume. There's more depth and fullness on the Pixel 4, but the difference was not as big as expected and I can personally forgive that given the real symmetry and stereo separation of the dual front-firing setup. Max volume is also excellent, easily as good as the Pixel and the audio reproduction is clearer and more balanced in the midrange. As such, vocals are easier to discern e.g. listening to videos while showering. I also don't find the 5 II to distort as much at max volume, where the Pixel 4 earpiece speaker became peaky and distorted on the top few notches. So though the Pixel may measure louder in some reviews, in use, the 5 II has the same useable volume range and both easily sufficient for listening to/sharing videos and music in loud environments.
It would be great to see Sony continue to work on the speaker quality in future models with a bit more bass extension, but this is definitely not an area where potential buyers should feel concerned that the 5 II is lacking.
Headphone Jack -
Onto the headphone jack, I have read that Sony have utilised custom amplifiers in their phones rather than the in-built audio from the Qualcomm chipset in previous models, and I wouldn't be surprised if this was also the case on the 5 II. In my eyes, a good audio source should provide 3 fundamental qualities:
- A low output impedance
- Ample driving power with low noise floor
- Linear frequency response
The 5 II ticks all boxes and it ticks them well for IEMs especially. It has a relatively low output impedance, just a few ohms at most. I tested by hooking up the Xperia and my desktop amp (THX789 with 1-ohm impedance) to the Campfire Audio Andromeda (5-driver BA earphone) with an in-line switcher. This revealed a slight bass drop-off, but pretty similar audio balance overall, a very good result. The Andro is notoriously sensitive to source impedance meaning that the vast majority of other IEMs, even high-end multi-drivers such as those from Empire Ears, Shure, etc, will sound accurate and faithful to their design on the 5 II - whereas, a higher OI will result in a shift in the sound signature.
Similarly, it has a flat response when measured using right RightMark Audio Analyser (Dolby off) and a black background with zero hiss even on the sensitive Andro. Of course, the detail retrieval and soundstage are not as strong as the THX desktop amplifier, but this is one of the best phones I've used in terms of audio output; it is simply clean and balanced. Even the LG V30 with its ESS audio hardware is not vastly more resolving and its output impedance is only slightly lower (measured at 1.5ohms) so both are equally good choices for the audio conscious. Power output is on the lesser side but easily adequate. I required 60% volume to get the Focal Clear to a comfortable listening volume leaving plenty of headroom for listeners preferring higher volumes. The THX amp obviously had a tighter, more extended bass but surprisingly, the Sony was not too bad, delivering a balanced sound and good soundstage. There was no bass drop-off and good detail retrieval. Impressive for a phone driving a high-end full-size headphone.
TLDR -
I think this is a pretty impressive smartphone all around and it's good to see that being recognised by critics and users. I am not personally hugely enamoured by the camera performance coming from the Pixel but there's definitely potential there especially if you like to tinker in post. The screen with warm colour temp in creator mode is a sight to behold with great accuracy to boot, and the battery life and audio performance are both sensationally good. In fact, 2-day battery life is very achievable with 3hrs SOT per day, not something I could say about any phone that I've owned prior. So hope this write up helps hobbyists on the fence. If you're looking for a good all-rounder smartphone with a clean headphone out, this is a great option!
Thanks a lot for this comprehensive writeup!
I'm in a similar situation to you, except I jumped ship from my previous Xperia's a bit later, ending up with a HTC U11+. Unfortunately the USB port has became faulty, meaning that I can't use the Usonic earphones that came bundled with the device. Flashing Android 10 to the device using a GSI image also means I can't even use a dongle anymore. Rough.
I had already planned to get the Xperia 5 last year, but managed to keep using my U11+ because it is still lasting me all day and I was afraid of having to face a worse camera in the X5 compared to the U11+.
With this post I think I'm pretty sure I'll love the X5 ii, it confirms everything I was thinking and have seen, from the display quality, the audio chipset, the speaker performance, etc.
I'm curious, on your profile it says you're a Physiotherapist but your level of understanding of tech is quite advanced for someone without any formal design/engineering background. I'm studying Electrical/Electronic Engineering at the moment, but I can say that a lot of my knowledge is from before I started my degree studies.
Shrenade514 said:
Thanks a lot for this comprehensive writeup!
I'm in a similar situation to you, except I jumped ship from my previous Xperia's a bit later, ending up with a HTC U11+. Unfortunately the USB port has became faulty, meaning that I can't use the Usonic earphones that came bundled with the device. Flashing Android 10 to the device using a GSI image also means I can't even use a dongle anymore. Rough.
I had already planned to get the Xperia 5 last year, but managed to keep using my U11+ because it is still lasting me all day and I was afraid of having to face a worse camera in the X5 compared to the U11+.
With this post I think I'm pretty sure I'll love the X5 ii, it confirms everything I was thinking and have seen, from the display quality, the audio chipset, the speaker performance, etc.
I'm curious, on your profile it says you're a Physiotherapist but your level of understanding of tech is quite advanced for someone without any formal design/engineering background. I'm studying Electrical/Electronic Engineering at the moment, but I can say that a lot of my knowledge is from before I started my degree studies.
Click to expand...
Click to collapse
Thanks! HTC made some great smartphones, wish they'd done better after the original One, such a goodie.
And yep, I studied Physio in my undergrad but tech/audio is my passion. I mostly lurk on forums reading the work of far more intelligent writers. On the side, I run an audio review website and through both I've picked up some knowledge - about audio especially.
I definitely am not a professional here in any way and love the perspective true pros bring to the industry, always a fascinating read. Hope you're enjoying your studies!
As a lurking audiophile, I wholeheartedly second this! I find that LG headphone jack still wins by virtue of having more volume steps. Sound quality wise, it is truly barely discernible given my portable usage of these two in my daily commutes.