Related
Hello,
I was wondering if anyone knew what the output impedance of the Desire Z is? Or if someone with a voltmeter could maybe test it? The only source i've managed to find online was some other random forum post which cites 64ohms which just seems way to high for a mobile device.
I'm planning on buying a pair of low impedance headphones and am trying to decide if I'll need to buy an amplifier.
Thanks
worldestroyer said:
Hello,
I was wondering if anyone knew what the output impedance of the Desire Z is? Or if someone with a voltmeter could maybe test it? The only source i've managed to find online was some other random forum post which cites 64ohms which just seems way to high for a mobile device.
I'm planning on buying a pair of low impedance headphones and am trying to decide if I'll need to buy an amplifier.
Thanks
Click to expand...
Click to collapse
the lower the impedance of your headphones, the less power is needed to drive them
I have a normal multimeter, so I cant get the impedance with accruacy.
In my case I read alot on mp3 players and audio cards in them. on headfi.com forums I found some frequency graphics and the htc desire worth it. Its not a cowon player but its pretty good. way better than a iphone.
So I decided to bought audiotechnica ath-m50 studio closed cans headphones. they are always in the top 10 lists of audiophile forums.
They are 38 ohms impedance headphone, and its pretty impressive on the DZ. I can't barely see any difference on a real amp like yamaha or audioresearch one.
on audiophile forums lot say that youll need a amp, buy with poweramp app its just allright.
Just had a go at measuring the output impedance of my Desire Z.
R Load = 19.9 ohm resistor
V = 171.3mV
V load = 76.7mV
Zo = 24.5 ohms
So output impedance is quite high, ideally use headphones with impedance of 300 ohms or above, low impedance headphones will work but will draw more current and won't sound as manufacturer intended but may well sound fine for day to day use.
My Sony broadcast headphones are 32 ohm whereas my Beyer DT100's are 400 ohm so a much better match for the DZ, both play loud, the DT's therefore are more efficient at converting input power into sound pressure.
measured again with 180 ohm load
Zo = 24.9 ohms
Just got a i9505. A few months I had an i9500 that was lost :crying:
I remembered people complaining of distortion in the i9505's output, which was eventually solved by a firmware upgrade. (can anyone still find that thread?) However what I'm hearing seems to be just the background electronic noise of the headphone out. It fades in and out in the quiet parts of classical music. I'm using IEMs, but dynamic ones that aren't particularly sensitive. I have a pair of custom IEMs on order that employ balanced-armature drivers and in my testing were much more sensitive than what I'm using right now--I'm afraid the noise will really get on my nerves when those arrive.
I note that the the headphone out has plenty of gain--I'm only at 25% playing quiet classical music to loud levels. Is there a mod / tweak that turns down the hardware gain of the i9505's headphone amplifier to hopefully reduce the amp noise?
Ironically the fake S4 I have been using these few months, a "Bluebo B9500", may have the i9505 beat for audio quality :silly:
You can install Faux Kernel, he has implemented a sound control, you can alter the headphone gain with his app or scripts.
On further experimentation I discovered that the "noise" I hear is signal dependent, I am pretty sure it is an artifact of a sample rate conversion from 44.1kHz (the sampling rate of almost all music files) to 48kHz (the output sample rate of the i9505,as reported by Viper4Android). When I play a 0-22050Hz sine sweep I can hear and record one big alias mirroring the original tone from 22050Hz back down to 0Hz
Does anybody else know of this problem and has there been any custom ROMs that address this?
I've been reading mixed reviews on the SQ of these units. I've read about the mod and it sounded like the sound just got louder. But is the sound fine and just flat sounding or is it something else? I was thinking a line driver would help or maybe it would just amplify the noise. I have some decent audio equipment and I don't want to feel I'm not getting the full potential out of them. I do use some WAV and FLAC files. I guess I can't tell if people are commenting on the on board audio or from their amps hooked up to the RCAs.
I have a 2016 Jeep Wrangler Unlimited. I had the factory 430N head unit, with the base 8 speaker system. I replaced 4 of the 8 speakers with Kickers and add a Kicker sub and amp (amp is just for the sub), and it sounded better but still not great.
THis weekend I added a Joying Jeep specific unit, and I feel like the sound of the overall system is vastly improved. I think it sounds pretty good, I primarily use Google Play Music and Pandora.
CadillacMike said:
I have a 2016 Jeep Wrangler Unlimited. I had the factory 430N head unit, with the base 8 speaker system. I replaced 4 of the 8 speakers with Kickers and add a Kicker sub and amp (amp is just for the sub), and it sounded better but still not great.
THis weekend I added a Joying Jeep specific unit, and I feel like the sound of the overall system is vastly improved. I think it sounds pretty good, I primarily use Google Play Music and Pandora.
Click to expand...
Click to collapse
Did you add a line driver at all to bump up the voltage on the pre-outs?
splxtreme said:
Did you add a line driver at all to bump up the voltage on the pre-outs?
Click to expand...
Click to collapse
I did not add anything. The sub and amp are built for Jeep, so it might be a little weird:
http://www.kicker.com/SWRA411
have any of you measure the pre-out voltage on the joying? On my tonghai create 2nd gen unit, pre-out voltage is only 1V. What this means is you will have a high noise floor going into your amplifier. Then your amp will amplify any noise your pre-outs have picked up. So if you use a line driver to raise the pre-out voltage prior to going into your amp, the higher voltage actually LOWERS the noise floor, hence eliminating noise that your RCAs may pick up.
My advice is you MUST have a line driver if you plan to use these units w/ a external amp. The difference will be night and day, particularly in the higher frequencies, where amplified noise is more pronounced.
They are dirt cheap too. I bought one on amazon for ~$30. It is actually a 3-way cross-over/line driver combo. It feels cheap but does it's job as a line driver very well.
edit: I say the above from personal experience, as I literally spent months trying to figure out why my tweeters always sounded so terrible. Line driver did the trick. Now I have clean highs and more clear, vibrant mids. I don't seem to notice much difference in my bass, so apparently noise interference isn't audible in the lower frequencies.
explain "line driver" to a stupid dumb dumb person, not me of course, just so other people understand
CadillacMike said:
explain "line driver" to a stupid dumb dumb person, not me of course, just so other people understand
Click to expand...
Click to collapse
line driver is just another word for a voltage step-up device. It's simply a way to increase the voltage of the RCA pre-outs coming out of your HU. Most amplifiers can read all the way down to about .5V pre-outs no problem. So a low pre-out voltage isn't much of a problem in theory.
But, an amplifier is also a voltage step-up device. It's going to take the pre-out voltage and multiply it by a pre-determined amount. You can adjust that amount by modifying gain. Your speakers are rated for a certain RMS power (80W for instance). So, if your speakers are 4 ohms rated for 80W RMS, the formula for the voltage it needs from your amp is sqrt(80 * 4) = 18V. So for your speaker to get the full rated power, it needs at least 18V from your amp. Higher pre-amp voltage = easier to reach the rated output to your speakers, since less gain is required.
Also, I mentioned noise floor. It's explained well here -
http://www.glasswolf.net/papers/preouts.html
It's not so easy to explain in laymans terms, but just know that higher voltage pre-outs = higher signal to noise ratio (SNR), meaning more signal & less noise = lower noise floor.
The article does mention however that having too high of a pre-out voltage is not good either, as it can lead to clipping. Best to buy an adjustable line driver (the $30 amazon model I bought had this feature... this one -
https://www.amazon.com/SX310-Pre-Am...70255121&sr=8-1&keywords=electronic+crossover). Most devices that accept pre-outs max out at around 6V, some as low as 4V. So don't crank up the line driver to max... I leave mine around 75-80%.
Hope that helps.
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!
Important update 11/9/18:
Android Pie seems to have fixed the digital gain defaulting to max. This means that the stock implementation is now optimal and using the gain control app is not needed (unless you want more voltage from a Dragonfly Red).
This was not announced anywhere to my knowledge, but is very pleasant surprise. I found it accidentally while testing the new dongle (above link).
Updated: Edit--I did some further tests and unfortunately the gain optimization only seems to work with line in mode, not with a "forced" headset scenario. So with a headset inserted, the distortion bands are still present.
As of the Nov patch, Pie on the Pixel has fixed both the line out and headset mode distortion banding when using the Google DAC.
When looking at the tracings, assume that stock will look as good as the optimized.
Made a simple spreadsheet with data so far:
{
"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"
}
The above numbers are using my gain control app, android volume at 25/25 to produce the best possible output signal
Note, the voltage at distortion threshold is the point where harmonic banding becomes >5 dB.
The THD is the harmonic distortion at 250 mV, this would be about 90 dB using full size cans like the Senn 650/800 that have a sensitivity of 103 dB/V.
Will work on uploading the exact images if anyone is interested.
Stepped sine tests at higher output (750mV) - An attempt to see how high impedance, lower sensitivity full size cans will effect distortion.
Edit: I updated the measurements and comments
Edit: Tested Apple USB C Dac
Edit: Tested the HTC dongle
Edit: Tested the Razer Dac
Edit: Tested the Realtec Dac
Edit: Tested the Geekria Dac
Edit: Tested the Audiolab USB device
Edit: Brief test of Dragonfly red. I would consider this to be the gold standard, very low impedance (<.1), minimal crosstalk, high volt out (not on stock), very low distortion.
Edit: There appears to be 3 (at least) "modes" that the Google dongle will do. Depending on the impedance of the headset/line out device attached, the output voltage varies quite a bit.
Low impedance - .36 Volt (high efficiency iem's)
High impedance - .9 Volt (full size headphones)
Very high impedance- 1.8 Volt (line out)
What gets interesting is the effects on testing. The "reference" sites like GSMarena, are testing with a Y cable setup (as I was) and will see high voltage out even with headphones attached since the use case was taken from the line out initially. What made me aware of this was that my app was not reading the Spl correctly, as if the max voltage was about 7 dB down (which it was). Also of note, the reduction in output is not due to a change in the Dac digital gain (the one that adjust from 1 to 175), so the distortion bands are just as bad since they kept that gain at 175.
RMAA:
Used both a Senn 800S and Shure 846 as load, as well as no load.
Note the flat FR on the SE 846 which verifies the low impedance above.
So superb impedance, great voltage out, pretty good crosstalk (with low impedance iems).
But, significant distortion using stock gain settings:
And it has 25 volume steps (each tick is 2.3 dB).
going to subscribe for those that have invested in 3.5mm portable audiophile amps/iems.
Nice work. Subscribed!
By the way, very nice headphones. I'm looking forward to testing out LDAC on my MDR-1000Xs when my Pixel arrives.
Hi bjmrd,
So I was looking at the Phone arena review in regards to the audio quality through the dongle.
It showed this
Not sure if this means the 2XL is weaker than the 2?
What does your test conclude? Would this device provide a nice audio experience for the user ?
This is important to me as I use 3.5mm headphones.
Thanks in advance,
Previous devices are the V20 (superb) and the non DAC G6.
I double checked my voltage measurements and they seem correct (compared to a dragonfly red).
The dongle may contain a DAC (like apple lightening). It is possible that with root, ALSA, some of the distortion can be reduced if so. Certainly the voltage is good, impedance great.
Oh and each volume tick is about 2.5 dB.
Addendum - The dongle definitely has a DAC
I have ordered a USB audio pass through cable to see if onboard analog is enabled
bjrmd said:
I double checked my voltage measurements and they seem correct (compared to a dragonfly red).
The dongle may contain a DAC (like apple lightening). It is possible that with root, ALSA, some of the distortion can be reduced if so. Certainly the voltage is good, impedance great.
Oh and each volume tick is about 2.5 dB.
Addendum - The dongle definitely has a DAC
I have ordered a USB audio pass through cable to see if onboard analog is enabled
Click to expand...
Click to collapse
bjrmd when you have a minute can you explain some of the numbers and what they might mean to the layperson. Why is high voltage good and low impedance good.
jawmail1 said:
bjrmd when you have a minute can you explain some of the numbers and what they might mean to the layperson. Why is high voltage good and low impedance good.
Click to expand...
Click to collapse
The max voltage is what limits the overall volume, so if you are driving over the ear cans and need power, this is helpful. Not important for sensitive iem's.
Crosstalk--a mixing of the R and L channels- not intentional. Better crosstalk--more stereo separation.
Impedance- internal resistance of the amplifier--if this is too high, causes frequency response changes depending on the type of iem and the iem impedance.
Thanks for the definitions bjrmd!
A couple of key points:
The dongle is a USB DAC, so basically audio out as supplied by Google is determined by that cheap little item. You could get a better USB DAC like a dragonfly, but it is bigger.
The distortion is based on some sort of driver parameter issue. I think the reason will end up being that the dongle DAC is set for max digital gain which is not optimal (unless needed). The following test shows why:
I used the Dongle on my Pixel orig and Pixel 2 with the Neutron player usb driver and the signal was clean.
This is a tracing, the same on both devices.
This is a dragonfly red, virtually the same:
So, I think the Google designers did not properly design the software, but the hardware is solid. In the meantime, use the Neutron player USB driver and you will have a very clean signal.
bjrmd said:
The max voltage is what limits the overall volume, so if you are driving over the ear cans and need power, this is helpful. Not important for sensitive iem's.
Crosstalk--a mixing of the R and L channels- not intentional. Better crosstalk--more stereo separation.
Impedance- internal resistance of the amplifier--if this is too high, causes frequency response changes depending on the type of iem and the iem impedance.
Click to expand...
Click to collapse
Thanks:good:
Dongle optimization on the old Pixel:
Alsa installed (need root, so can't do on the new Pixel yet).
Dongle card output:
numid=2,iface=MIXER,name='Headphone Playback Switch'
; type=BOOLEAN,access=rw------,values=1
: values=on
numid=3,iface=MIXER,name='Headphone Playback Volume'
; type=INTEGER,access=rw---R--,values=2,min=0,max=175,step=0
: values=18,18
| dBminmax-min=-175.00dB,max=0.00dB
numid=4,iface=MIXER,name='Extension Unit Switch'
; type=BOOLEAN,access=rw------,values=1
: values=off
numid=1,iface=PCM,name='Playback Channel Map'
; type=INTEGER,access=r----R--,values=2,min=0,max=36,step=0
: values=0,0
It seems that the android software defaults the dongle digital gain at max:
This leads to the distortion bands previously seen.
But, if we cut the dongle DAC gain down, and raise the android Volume to net the same(or even a bit higher) voltage:
The bands are gone.
In addition we can get 1 volt out without too much distortion as well:
This is not that difficult to do, but root is needed.
I will try to do some more detailed testing over the weekend, but I wanted to get this up so if someone was on the fence due to audio issues, if you are willing to wait for root and a few mods, this is a reasonable solution.
BTW for 9 bucks the dongle is pretty good with the correct gain settings
do you have test the ldac on google pixel 2 xl ?
is it working? or any change?
Does anyone else think it's a bit ridiculous how large the adapter is? It's huge compared to Apple's.
Thanks for posting this! Would you say this adapter is superior to Apple's lightning adapter? There are a few audiophile sites that have posted measurements for it and say it's great, but I'm not sure how to compare it to this.
Here is one such review: http://www.kenrockwell.com/apple/lightning-adapter-audio-quality.htm
yangyangyanglei said:
do you have test the ldac on google pixel 2 xl ?
is it working? or any change?
Click to expand...
Click to collapse
See first post.
In order to get digital gain to change in the dongle Dac, root is needed, hence tested the dongle on original Pixel in post 11.
The results should be identical .
Nitemare3219 said:
Does anyone else think it's a bit ridiculous how large the adapter is? It's huge compared to Apple's.
Thanks for posting this! Would you say this adapter is superior to Apple's lightning adapter? There are a few audiophile sites that have posted measurements for it and say it's great, but I'm not sure how to compare it to this.
Here is one such review: http://www.kenrockwell.com/apple/lightning-adapter-audio-quality.htm
Click to expand...
Click to collapse
The dongle is a reasonable USB DAC, but the stock implementation (at least on my setup) is poor. Google has the DAC digital gain at max, which introduces significant distortion.
To backtrack, the net powerout of this device is dependent on the Android system volume + the digital gain of the DAC itself. So you could have a very low android volume and very high DAC gain (as it is now) or a higher android volume and lower DAC gain (optimal for distortion).
My dragonfly red will behave this way as well in regards to distortion.
So unless you want maximum power out, we want android volume higher and DAC gain lower.
The impedance is lower than Apple's, power out may be higher as well. Not sure about crosstalk but both are probably very good.
Size wise- I couldn't even tell there was circuitry in the dongle-they did a good job IMO in the size.
Bottom line-- the audio out of the new Pixel is basically dependent on what USB DAC used. The stock is fine with the above noted. In fact for all you know, another USB DAC may also default to max gain and distort as well.
bjrmd said:
The dongle is a reasonable USB DAC, but the stock implementation (at least on my setup) is poor. Google has the DAC digital gain at max, which introduces significant distortion.
To backtrack, the net powerout of this device is dependent on the Android system volume + the digital gain of the DAC itself. So you could have a very low android volume and very high DAC gain (as it is now) or a higher android volume and lower DAC gain (optimal for distortion).
My dragonfly red will behave this way as well in regards to distortion.
So unless you want maximum power out, we want android volume higher and DAC gain lower.
The impedance is lower than Apple's, power out may be higher as well. Not sure about crosstalk but both are probably very good.
Size wise- I couldn't even tell there was circuitry in the dongle-they did a good job IMO in the size.
Bottom line-- the audio out of the new Pixel is basically dependent on what USB DAC used. The stock is fine with the above noted. In fact for all you know, another USB DAC may also default to max gain and distort as well.
Click to expand...
Click to collapse
Here's the size comparison.
bjrmd said:
I double checked my voltage measurements and they seem correct (compared to a dragonfly red).
The dongle may contain a DAC (like apple lightening). It is possible that with root, ALSA, some of the distortion can be reduced if so. Certainly the voltage is good, impedance great.
Oh and each volume tick is about 2.5 dB.
Addendum - The dongle definitely has a DAC
I have ordered a USB audio pass through cable to see if onboard analog is enabled
Click to expand...
Click to collapse
Can you post a link of the pass through cable you are using or reccommend?
---------- Post added at 01:57 AM ---------- Previous post was at 01:50 AM ----------
I may be misunderstanding the tech involved here but what is the best way to get the best wired audio here? Root tweaking the existing dongle DAC or doing something like a pixel rooted with viper4audio with a pass through dongle?
Jayme Helgerud said:
Can you post a link of the pass through cable you are using or reccommend?
I may be misunderstanding the tech involved here but what is the best way to get the best wired audio here? Root tweaking the existing dongle DAC or doing something like a pixel rooted with viper4audio with a pass through dongle?
Click to expand...
Click to collapse
A pass through cable would be using the phones DAC/SOC (qualcomm) and analog audio out, just like with a headset jack. Unfortunately, @chdloc did a brief test with a Pixel 2 (not XL) with that type of cable and it did not play. Also, I bet the analog out would have higher impedance , worse crosstalk than a USB device. Either way you need an adapter, so a dongle or pass through.
So far the only way to get wired audio is through using a USB DAC. The dongle is a full fledged USB DAC. It has a fixed sample rate of 48K but otherwise excellent specs(by itself). Even if you wanted a high sample rate DAC, the android driver/audioflinger only puts out 48K.
As mentioned above, the stock gain settings cause distortion (at least in my case). After rooting this will be easy to fix. Right now, Neutron player has a custom USB driver that prevents that distortion (by also keeping the DAC gain lower) - that is what I am using.
Maybe Google will read this and fix the issue in the next update.
I can't give an opinion on Viper.
bjrmd said:
A pass through cable would be using the phones DAC/SOC (qualcomm) and analog audio out, just like with a headset jack. Unfortunately, @chdloc did a brief test with a Pixel 2 (not XL) with that type of cable and it did not play. Also, I bet the analog out would have higher impedance , worse crosstalk than a USB device. Either way you need an adapter, so a dongle or pass through.
So far the only way to get wired audio is through using a USB DAC. The dongle is a full fledged USB DAC. It has a fixed sample rate of 48K but otherwise excellent specs(by itself). Even if you wanted a high sample rate DAC, the android driver/audioflinger only puts out 48K.
As mentioned above, the stock gain settings cause distortion (at least in my case). After rooting this will be easy to fix. Right now, Neutron player has a custom USB driver that prevents that distortion (by also keeping the DAC gain lower) - that is what I am using.
Maybe Google will read this and fix the issue in the next update.
I can't give an opinion on Viper.
Click to expand...
Click to collapse
Will the neutron driver affect music played with other apps? I assume not...only music you've downloaded and are playing in the neutron app. It sounds like (haha) I'm going to be looking forward to some system level driver mods someone (hopefully) develops for a rooted pixel. Thanks for the insights!
Jayme Helgerud said:
Will the neutron driver affect music played with other apps? I assume not...only music you've downloaded and are playing in the neutron app. It sounds like (haha) I'm going to be looking forward to some system level driver mods someone (hopefully) develops for a rooted pixel. Thanks for the insights!
Click to expand...
Click to collapse
Sorry, no - only works in Neutron player.
Once root comes I plan on making something like this but will be much easier since only USB audio needs control. I would like some type of Spl level info as well to save our ears.