Pixel 6 logs in GPStest database
瀏覽次數：926 次
跳到第一則未讀訊息
Knut
2021年11月16日 清晨5:50:062021/11/16
收件者：GPSTest
Hello all,
I'm thinking about buying a Pixel 6 (not Pro) and I'm slightly puzzled about the 7 entries currently shown in the database, for the following reasons:
1. All entries show A12, as expected. Also all show the same App version name and version code (3.9.16), but still the first one shows a different number under "Android Software (incremental) build", 7738411 rather than 7805805 for all others. This first entry shows US as country location, and there are two more US entries further down in the list showing different properties
2. Entry 2021-10-28T06:08:12Z (US) has L1 and GPS only in terms of reported GNSS support, while the next one (2021-10-28T15:27:02Z, also US) has GPS, Glonass and Galileo on L1 and L5. Maybe here the "full measurements option" is at play?
3. Entries 2021-10-28T17:16:26Z and 2021-10-28T17:18:42Z (both DE as country) again show L1 only, and their only difference is w.r.t support of Carrier Phase, the latter being the only one out of 7 not to support Carrier Phase measurements
4. The first of the aforementioned DE entries and the third one (2021-10-26T17:37:07Z) have exactly one difference, with respect to Dual Freq.
I can't figure out how those differences would be related to different model restrictions per country. Other differences like only the GB model flagging support for "Delete assist" are at least consistent with a per-country differentiation.
Any ideas, anyone? Maybe some of these are grey imports?
Best regards and thanks in advance for your ideas, observations,
Knut
Sean Barbeau
2021年11月16日 下午2:00:062021/11/16
收件者：Knut、GPSTest
Knut,
I just sorted the sheet which pulled a few more Pixel 6 records up - here's all of them:
| Manufacturer | Model | Name | Android version | API level | GNSS hardware year | GNSS hardware model name | Dual-frequency | Supported GNSS | GNSS CFs | Supported SBAS | SBAS CFs | Raw measurements | Carrier phase (getAccumulatedDeltaRangeMeters()) | Navigation messages | NMEA | Inject Predicted Satellite Data Service (PSDS) | Inject time | Delete assist | RESERVED | RESERVED | Automatic gain control | GnssAntennaInfo | App version name | App version code | App build flavor | User location country code | Date | Date (ISO 8601 UTC format) | Android software (Incremental) build | Android software codename | Number GNSS antennas | GNSS antenna CFs | Device |
| SUPPORTED | GPS, GLONASS, GALILEO, QZSS | E1, E5a, L1, L5 | SUPPORTED | SUPPORTED | SUPPORTED | SUPPORTED | SUPPORTED | SUPPORTED | SUPPORTED | NOT_SUPPORTED | 3.9.16 | 18093 | US | 10/28/2021 05:28:28 AM (GMT) | 2021-10-28T05:28:28Z | 7738411 | REL | oriole | |||||||||||||||
I haven't had a hands-on with the Pixel 6 or 6 Pro yet, but from the Pixel 5 and other devices L5/E5a can be tricky because typically the device will lock on L1/E1 first, and then sometime after that (best case a few seconds) L5 and E5a will show up in the list. So it's very possible the records without L5/E5a were because of quick submissions before L5/E5a signals showed up. I plan to add another database field to track TTFF and/or uptime to help address this.
Based on my experience with the Pixel 5, my *guess* is that all variants of the Pixel 6 do indeed support L5/E5a and it's not region locked or hidden behind "force full measurements" (an aside, I plan to add "force full measurements" setting to the database for Android 12 and higher). I would expect the only region locking to be that Beidou wouldn't be included in the U.S. due to FCC restrictions.
I'm not sure why some of these devices are reporting Carrier phase as NOT_SUPPORTED - I'd expect all of them to support it if one does. On Android 12 GPSTest by default triggers "force full measurements", so it's possible the user switched off that setting for these values? Again, tracking this setting in the database will help in the future. My guess is that all devices do indeed support it.
Delete assist is a destructive operation, so it's not automatically performed by GPSTest before the upload to test support, which is why most records have it blank. If the user has manually invoked that function prior to submitting a record to the database, then the column will have a value for SUPPORTED or NOT_SUPPORTED (i.e., there is no way to know if deleting assist data is supported without actually triggering the deletion). My guess is that given one record supports that function, and all devices support injecting PSDS, that all devices will support deleting assist data too.
Sean
--
You received this message because you are subscribed to the Google Groups "GPSTest" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gpstest_andro...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gpstest_android/7b2f7e3c-ecbd-4e81-8045-535f01f46fa5n%40googlegroups.com.
Knut
2021年11月16日 下午2:04:342021/11/16
收件者：GPSTest
Sean,
thanks a lot for the quick and detailed answer. I guess my approach in the coming weeks will be to purchase a P6 in a brick and mortar store, run outside, install GPStest and in case dual F support and carrier phase measurements don't show up, return it right away ;-)
Thanks again, GPStest is really great!
Knut
Sean Barbeau
2021年11月16日 下午2:18:392021/11/16
收件者：Knut、GPSTest
Thanks, let me know what you find!
I've actually installed GPSTest in a brick-and-mortar store demo unit before in the store (enabling installing from untrusted sources, then downloading the APK from GitHub to avoid logging into Google Play), although since it's a demo unit there is no guarantee that what you get in the box will match :).
Sean
To view this discussion on the web visit https://groups.google.com/d/msgid/gpstest_android/af9b187d-8066-458a-a176-3972b5c86d58n%40googlegroups.com.
Knut
2021年12月8日 中午12:13:342021/12/8
收件者：GPSTest
Hi Sean,
here are my findings:
Brick&Mortar store had the Pixel 6 locked up in a "demo" mode with no access to Settings, so no chance to enable installing from untrusted sources and testing in-store. So I ended up ordering online from a seller with a decent return policy and ran the tests from home.
Everything except GNSS_Antenna_Info is supported, but this is in line with all other Pixel 6 entries. Also I can confirm that force full measurements was enabled by default.
Now it has to go back into the box for a merry dual-frequency Xmas holiday ;-)
Cheers,
Knut
Sean Barbeau
2021年12月9日 上午11:37:202021/12/9
收件者：Knut、GPSTest
Thanks for sharing!
Sean
--
You received this message because you are subscribed to the Google Groups "GPSTest" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gpstest_andro...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gpstest_android/c44c2434-9b30-49ba-b0b8-1d7606e2ffb6n%40googlegroups.com.
gpsfan
2022年4月5日 下午1:56:562022/4/5
收件者：GPSTest
Hi Knut,
How are you getting along with your Pixel 6 ? I think you had the Mi 8 and were quite happy with it. Me too and It's still my main phone but I damaged the screen recently and with the slight price decrease of the P6 I've been looking into getting one.
Have you compared the accuracy of the Mi 8 and P6 ? TBH I think any current phone is going to provide enough accuracy for most anyone's needs, even for geeks like us ! It's nice to be able to record raw data too to possibly post-process it but the smartphone antennas are just not there (yet ?) to hope to get any easy RTK "fix" solution as some experts have found : https://rtklibexplorer.wordpress.com/2022/01/10/google-smartphone-decimeter-challenge/
gpsfan
2022年4月20日 下午2:24:272022/4/20
收件者：GPSTest
So I got a hold of a Pixel 6 and I'm...disappointed. It's just an "awkward" device, very heavy, no face unlock, buggy fingerprint unlock, good camera though...
Anyway as far as GPS goes it does have a separate GPS chip that seems to pick up things where my trusty Mi 8 left them with the BCM4776 (instead of the BCM4775) and raw measurements are supported (I also see ADR data in the GNSS Logger logs) but...the actual Android apps that leverage that data, i.e. PPP Wizlite and 3PGo and that work fine on the Mi 8 are "stumped" here and can't resolve a PPP solution. Not sure why but it's a problem, anyone else notice that ?
Sean Barbeau
2022年4月20日 下午2:36:382022/4/20
收件者：GPSTest
>the actual Android apps that leverage that data, i.e. PPP Wizlite and 3PGo and that work fine on the Mi 8 are "stumped" here and can't resolve a PPP solution. Not sure why but it's a problem, anyone else notice that ?
The Android API for raw measurements and ADR evolved since the Mi 8 was released, and the initial assumption made in the GNSSLogger (and I'm assuming other apps) that all valid ADRs will have GnssMeasurement.getAccumulatedDeltaRangeState() = 1 no longer holds true on newer devices (>= API level 28 i.e. Android 9).
getAccumulatedDeltaRangeState() now returns a bitmask, so you can have valid states where the returned value isn't equal to 1.
So if apps haven't been updated to take this into account, then they would think the device wasn't returning any valid ADR data.
I'd suggest reaching out to the developers of the apps and pointing this out (including the above link) and see what they say.
Sean
gpsfan
2022年4月20日 下午3:24:572022/4/20
收件者：GPSTest
Thanks for the quick reply, I suppose that could be a problem indeed. Thing is these apps that leverage raw data are few and far between and they got about zero traction and don't get mentioned anywhere, apart from here occasionally ;-) The (sad) reality is that smartphones are not able to provide measurements of a sufficient quality as rtklibexplorer found (see my previous post about that). I suppose that the Trimble RTK partnership does provide some hope but the limitations of the design are likely to stay, not to mention the fact that the few people who need extreme accuracy will not be relying on a consumer smartphone.
I was hoping to log some data with the Pixel 6 in a dense urban area to see how efficient Google's "Machine Learning" algorithms are (no recent mention of them unfortunately) but I'm not sure I'll have the time for that before returning it.
回覆所有人
回覆作者
轉寄
0 則新訊息