Any hints / tips for using the following a GPX feature?

[yellowbarchetta]

Tried it out this morning following a 12 mile route I ran a couple of weeks ago. The gpx file is fine, but the "instructions" that the file sent to the treadmill were inconsistent, and after just over a mile it slowed things down. During the first mile the pace wavered a lot more than the gpx file did.
I suspect user error on my part.
Any hints or tips on how to use it properly? And how could I see the map at the same time on the phone (android), or is this not possible? (Need to keep the laptop free so I can watch movies etc and I don't have admin rights on it either).

[cagnulein]

are you talking about the video+gpx feature? if so share a debug log. also @Bepo7012 could be interested

[yellowbarchetta]

For some reason the debug log didn't get created (or I deleted it!) so next time I'll try the same file and see what happens. I'm just talking about the "Open GPX" feature on the main menu, no idea about the video+gpx.

I did also try to simulate it with the "fake treadmill" but that doesn't seem to work with the "Open GPX" feature (though again, this could be user error!)

[cagnulein]

For some reason the debug log didn't get created (or I deleted it!) so next time I'll try the same file and see what happens. I'm just talking about the "Open GPX" feature on the main menu, no idea about the video+gpx.

i was talking about the ability to use qz as rouvy or kinomap to sync a video with the gpx file!

I did also try to simulate it with the "fake treadmill" but that doesn't seem to work with the "Open GPX" feature (though again, this could be user error!)

did you enable the speed forcing in the treadmill settings? disable it and it will work because they are gpx for bike

[cagnulein]

for the map, yes, you can see on the same phone. There is the map icon in the top right corner.

[yellowbarchetta]

Without speed forcing the app didn't appear to control the speed of the treadmill - but I need to disable it for the fake treadmill? OK, will try that. Again, maybe user error - I will disable that and see what happens.

[cagnulein]

yes disable it and then the inclination at least will change following the gpx or at least it should :) Roberto Viola Software engineer and open source enthusiast http://robertoviola.cloud Il giorno dom 17 dic 2023 alle 22:21 yellowbarchetta < ***@***.***> ha scritto:

…

Without speed forcing the app didn't appear to control the speed of the treadmill - but I need to disable it for the fake treadmill? OK, will try that. Again, maybe user error - I will disable that and see what happens. — Reply to this email directly, view it on GitHub <#1893 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWEOYDXJ2VTRGEXMDITYJ5O6NAVCNFSM6AAAAABAYTMAPOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TQNZZGU4DE> . You are receiving this because you commented.Message ID: ***@***.*** com>

[yellowbarchetta]

That might be what I'm looking for to help me out. Though the .fit file leapt from the UK to Australia (hence "crazy stuff"), but more boringly the speed dropped to 0.5mph from about 7mph for no discernible reason.

I posted it here in discussions rather than bugs as i'm sure it's something I am doing wrong or not understanding about how the feature works. My assumption was that you simply uploaded a gpx and then the treadmill would both run at the speed that the gpx had previously recorded, and simulate inclination as the file had recorded too.

[cagnulein]

the fit file generated has the right location

are you talking about what you see on the map on QZ in realtime?

but more boringly the speed dropped to 0.5mph from about 7mph for no discernible reason.

with the speed force disabled, you're controlling the speed. am I missing something? Sorry probably I'm not understanding the issue completely

I posted it here in discussions rather than bugs as i'm sure it's something I am doing wrong or not understanding about how the feature works.

sure no problem

My assumption was that you simply uploaded a gpx and then the treadmill would both run at the speed that the gpx had previously recorded, and simulate inclination as the file had recorded too.

exactly, it's the purpose. I used it with the same scenario several times last year. If i didn't add some new bugs, it should still work.
Let me know

[yellowbarchetta]

That's not the map I get (though the first part is correct), and I can't use fitfileviewer to show maps like you can (is that a dev option?)

But if you show the rows especially for the 2nd and 3rd records you can see it leap from lat/long +53 / -2.7 to -50.4 / 167.9 and then back into the Irish Sea at 53.7/ -3.4

Here's the .fit file when uploaded to GPS visualiser overall showing that being mapped.

And showing the first part only

That leap across the world was there visually within QZ as the accumulated distance suddenly went up to 36000 miles.

[yellowbarchetta]

I'm going to have a look and see if there is any issue in the source data of the GPX file that makes these things go crazy (e.g. dodgy altitude data or something like that) which would explain things.

So, for me to do what I thought I wanted it to do with my Horizon treadmill, am I right in thinking these are the settings?

• Treadmill speed forcing - ON (this will make the TM run at the speed that I actually did the real world run at)
• Simulate incline with speed - OFF (incline will adjust as per the file source)
• Open GPX - use this to load the run
• Tiles - don't touch any of the tiles to adjust speed / inclination as QZ will read from the gpx file to control it

[yellowbarchetta]

OK, just ran it again this time with speed forcing off and simulate incline with speed off (on Fake Treadmill). What I can see visually happening at around the 1 mile (1609 metre) mark is that the "inclination" field suddenly starts showing extreme negative inclination (-38 I think was the biggest number). This correlates with what I've seen in the real world treadmill where the pace dropped (when I had speed and incline linked). When I was on the "real" treadmill I didn't see this because (I assume) the minimum inclination that could be displayed on QZ and on the TM itself was "0.0", but presumably behind the scenes the negative incline was being processed.
What's really odd is that I can't see that negative incline (not to that degree anyway!) in the source file.
Can you replicate this? Upload the gpx file I've shared, run the "fake treadmill" option as above and set speed to something reasonably quick so that after 5 or 6 minutes it gets to the problem section?
Debug attached for this session.
debug-Mon_Dec_18_11_30_39_2023.log

[cagnulein]

That's not the map I get (though the first part is correct), and I can't use fitfileviewer to show maps like you can (is that a dev option?)

nope user mode, check the button on the top

But if you show the rows especially for the 2nd and 3rd records you can see it leap from lat/long +53 / -2.7 to -50.4 / 167.9 and then back into the Irish Sea at 53.7/ -3.4

i guess we are using different fit file, because i don't see this in the fit file that you sent, am I going crazy? :D

So, for me to do what I thought I wanted it to do with my Horizon treadmill, am I right in thinking these are the settings?

• Treadmill speed forcing - ON (this will make the TM run at the speed that I actually did the real world run at)
• Simulate incline with speed - OFF (incline will adjust as per the file source)
• Open GPX - use this to load the run
• Tiles - don't touch any of the tiles to adjust speed / inclination as QZ will read from the gpx file to control it

yes but there is a missing step after the open gpx: you have to press pause/start to load the training program. I mean it applies only to a new session (or just press start if the treadmill was already on pause)

I'm checking now your last log, hold on

[yellowbarchetta]

I think the fundamental here is I used the app incorrectly! Even if it has highlighted a minor bug!!

[cagnulein]

got the issue, in the AVG function i wasn't put a "weight" for each sample, so i was assuming that all the samples had more or less the same distance, instead, in your gpx, there are some points in the middle with very short distance. this put crazyness in the algorithm, let me fix this, should be very simple

[cagnulein]

done, i'm testing it :)
then i will check the coordinates issue

[cagnulein]

@yellowbarchetta it's building https://github.com/cagnulein/qdomyos-zwift/actions/runs/7248766081 tested and now it works correctly, thanks for the feedback. now i'm focusing on the coordinates issue

[cagnulein]

the problem of the coordinates is the big power jump

23:27:59 | -50,416301 | 167,901207 | 0 | 0 | - | 52.427 | 0

let me check if i see the reason in the debug log

[cagnulein]

the issue start from here

Sun Dec 17 22:27:59 2023 1702852079899 Debug: trainprogram.cpp void trainprogram::scheduler() positionOffset 59326.525

[cagnulein]

it's a problem from the auto lap feature, let me check!

[yellowbarchetta]

Glad the oddities have highlighted something meaningful!

Odd that auto-lap didn't create the issue this morning when I tried it, perhaps I caught the screen and logged a manual lap which has impacted things?

GPX files with varying gaps between timestamps are not too uncommon, but I suspect different watches and different download services process things in different ways. The GPX file I used came from Runalyze which must have generated the file itself as it will originally have been synced from Garmin to Runalyse as a .fit file. So you're at the mercy of developers on various websites here!

[cagnulein]

argh the debug log is truncated, let me anyway try to check if i can see it as well

[yellowbarchetta]

No worries, that part wasn't the biggest real world issue, but if I see that discrepancy again I'll let you know - the inclination thing was the biggest issue.

…

On Mon, 18 Dec 2023 at 16:35, Roberto Viola ***@***.***> wrote: @yellowbarchetta <https://github.com/yellowbarchetta> i'm not able reproduce the auto lap issue again unfortunately. give it a try to my last build and let's see if it works ok. Let me know! — Reply to this email directly, view it on GitHub <#1893 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AOAEASNZAX5N5JV6PZ43ZWLYKBWG7AVCNFSM6AAAAABAYTMAPOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TQOBYGUYTI> . You are receiving this because you were mentioned.Message ID: ***@***.*** com>

[cagnulein]

@yellowbarchetta http://80.211.67.253:8081/training.htm?fbclid=IwAR07s1MGtk-xjNPZDj_vsE89PEUqv6D5ycffYmNxDUqax8rD0bbdMFpqDLg this one :)

i didn't update it since a lot of time, but the basic functions are there already :)

[cagnulein]

there is a web gui that i created in the past. i will bring it up again and i will share the link Roberto Viola Software engineer and open source enthusiast http://robertoviola.cloud Il giorno mer 20 dic 2023 alle 18:30 yellowbarchetta < ***@***.***> ha scritto:

…

Yes, I saw that bit; would just be nice to have little online GUI for it like those zwo creating websites to make it easier to create things. But 100% that sort of structure would be usable, just need to test a few out. — Reply to this email directly, view it on GitHub <#1893 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWHDYGRYS7J42JIZQC3YKMOCLAVCNFSM6AAAAABAYTMAPOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TSMJQG4ZDM> . You are receiving this because you commented.Message ID: ***@***.*** com>

[yellowbarchetta]

Probably the wrong place to ask this, but with all of the builds that you've done to address queries I've raised, how do I know if I'm running a version of the app which has all of the right bits in it? Is there something which indicates when a pull request has been blended in with other versions and developments?
It's probably basic stuff about how GitHub works that I'm missing fully understanding to be honest!
I guess I should revert to using the the play store version at some point, but I wouldn't want to do that and lose the developments that you've made (especially the pace smoothing one).

[cagnulein]

if you are downloading the master one (as you did since i have linked you this one the last time), you are good to go since all the patches are already merged. you can download the nightly release from the homepage of qz in github. the beta on the play store is aligned, it's only doesn't have the last fix about the inclination Roberto Viola Software engineer and open source enthusiast http://robertoviola.cloud Il giorno gio 21 dic 2023 alle 17:54 yellowbarchetta < ***@***.***> ha scritto:

…

Probably the wrong place to ask this, but with all of the builds that you've done to address queries I've raised, how do I know if I'm running a version of the app which has all of the right bits in it? Is there something which indicates when a pull request has been blended in with other versions and developments? It's probably basic stuff about how GitHub works that I'm missing fully understanding to be honest! I guess I should revert to using the the play store version at some point, but I wouldn't want to do that and lose the developments that you've made (especially the pace smoothing one). — Reply to this email directly, view it on GitHub <#1893 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWDWP4AZWDJN6PQPQHTYKRSWDAVCNFSM6AAAAABAYTMAPOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TSMRQGQ2TM> . You are receiving this because you commented.Message ID: ***@***.*** com>

[yellowbarchetta]

Sorry to post on this old thread, but just wanted to check my understanding. I use the "speed offset" at 0.975 in the app so that the pace reported in the final file better matches the treadmill (both in terms of measuring the speed of the belt and in terms of perceieved effort). That all works fine.
But when I load a GPX file to "follow", do I need to do one of the follow:-

• leave it as is
• adjust the speed offset factor back to zero so that if the gpx file says run at 7.0mph it will tell the treadmill to run at 7.0mph
• invert the speed offset (i.e. run at 1.025) so that whatever the gpx file says will cause the treadmill to "think" it is running slightly faster than that?

Or am I overthinking this and there is no connection between the speed offset and the gpx file following?

[cagnulein]

you can increase the target speed tile yes. But if you want i can handle the speed offset and the speed gain in the general scope

[yellowbarchetta]

I've not done this for a while, but I think these are the outputs from the last time I did it. First set of splits is what the app reported in the .fit file and the second is the original splits from the source file. I think this shows that the resulting file reported a slower speed than the original, which I think is where my original "concern" about this came from. The gpx has made the TM properly run at the speed in the .gpx file. but then when the .fit file has received the speed data back from the treadmill, it has applied my 0.975 factor and as a result the actual run I did was slower than the outdoor one. e.g. mile 1 - 535s vs 524s (97.9%), mile 2 518s vs 505s (97.5%) etc. It's not entirely precise, probably due to roundings, but the difference is there.

1) 1mi - 8:55(8:56/mi)
2) 1mi - 8:38(8:38/mi)
3) 1mi - 8:33(8:32/mi)
4) 1mi - 8:35(8:34/mi)
5) 1mi - 8:36(8:35/mi)
6) 1mi - 8:25(8:25/mi)
7) 1mi - 8:25(8:25/mi)
8) 0.06mi - 27(8:01/mi)
9) 0.04mi - 20(8:48/mi)

1) 1m - 8:44(8:44/m) 137/149bpm [1196b/mi] 111cal 6.88/7.24mph
2) 1m - 8:25(8:25/m) 147/154bpm [1238b/mi] 117cal 7.12/7.78mph
3) 1m - 8:17(8:17/m) 143/148bpm [1185b/mi] 106cal 7.24/7.95mph
4) 1m - 8:21(8:21/m) 145/149bpm [1211b/mi] 107cal 7.19/7.62mph
5) 1m - 8:21(8:21/m) 144/147bpm [1203b/mi] 102cal 7.18/7.58mph
6) 1m - 8:12(8:12/m) 147/152bpm [1206b/mi] 104cal 7.31/7.78mph
7) 1m - 8:11(8:11/m) 145/149bpm [1188b/mi] 96cal 7.33/7.89mph
8) 0.04m - 17(7:43/m) 144/145bpm [1112b/mi] 3cal 7.77/7.41mph

[yellowbarchetta]

So yes, I think the "fix" would be that if there is a speed offset, apply that to the instruction to the treadmill, and then that would fix the resulting .fit file being correct.

[cagnulein]

@yellowbarchetta Good news! I just checked and the code was already there! So it's already working on this way! Which version do you have of qz?

[yellowbarchetta]

I'm on 2.16.46

[cagnulein]

Ok so you have already this patch, just added :) Roberto Viola Software engineer and open source enthusiast http://robertoviola.cloud Il giorno lun 11 mar 2024 alle 20:36 yellowbarchetta < ***@***.***> ha scritto:

…

I'm on 2.16.46 — Reply to this email directly, view it on GitHub <#1893 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWDJUKPA2YMSZKZLWXDYXYBSTAVCNFSM6AAAAABAYTMAPOVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DONJQHAZTO> . You are receiving this because you commented.Message ID: ***@***.*** com>

[yellowbarchetta]

Great - so just so that I understand, is this saying that if there is a figure other than 1.0 in the speed offset setting, then this offset figure is applied to the instruction from the app to the treadmill without the user having to do anything else to speed things up? I can test this out tonight maybe, set the TM to follow a gpx file, and then after 1 mile check that the app display shows me the right lap pace.

[cagnulein]

1.0 in the speed offset setting

other than 1.0 on the speed gain and 0.0 from the speed offset

is applied to the instruction from the app to the treadmill without the user having to do anything else to speed things up?

exactly!

I can test this out tonight maybe, set the TM to follow a gpx file, and then after 1 mile check that the app display shows me the right lap pace.

oky!