Making and Modifying RLVs

[mickyduck]

I also have a lot of old tech that is eventually not able to be used or superseded.. the issue that I think we all have is we bought a license for TTS4 advances (two in my case) which may be useless if I have any issues with my hardware.

It not the lack of support its the removal of the license server without giving legitimate users the ability to use a product they purchased in good faith.

I know there are ways round this but we have films that were not activated as I copied them over to a new PC instead of installing them so TTS was uninstalled and reinstalled I just copied rather than installed the films. I can now rename the PC to the same as before but then TTS wont work ...

Anyway I suspect we wont get anywhere I have emailed Garmin support and had the 20% off new equipment offer but its not that simple for us, my very old tech carbon "Look" will only take an 8 speed cassette (Campag) so its more than the two Fortius trainers we have.. any way Im bitching for the sake of a couple of films.. ATM we have two fortius trainers and two TTS4 Adv both working

Thanks for the clarification re TTS depreciation before the Garmin sale

[mcorn]

Anyway I suspect we wont get anywhere I have emailed Garmin support and had the 20% off new equipment offer but its not that simple for us, my very old tech carbon "Look" will only take an 8 speed cassette (Campag) so its more than the two Fortius trainers we have.. any way Im bitching for the sake of a couple of films.. ATM we have two fortius trainers and two TTS4 Adv both working

There should be no issue using your 8 speed with the newer trainers. You just have to buy a compatible cassette and possibly an inexpensive spacer. 20% is a pretty good discount.

[mickyduck]

The 8 speed information is contrary to an earlier response that said * speed Shimano would be fine but not Campag as the spline pattern was changed.. TBH I dont know

[mcorn]

The 8 speed information is contrary to an earlier response that said * speed Shimano would be fine but not Campag as the spline pattern was changed.. TBH I dont know

Looks like I may have been incorrect, at least in part. The Garmin - Tacx website says the Campy body for the Neo is for 9 to 12 speed:

https://buy.garmin.com/en-US/US/p/707020/pn/T2805.51

But people do a fair amount of adapting between Shimano and Campy shift systems. My understand is a Campy 8 speed has 5 mm spacing gear, which is the same as a Shimano or SRAM 7 speed cassette, so it is possible either of those would be compatible with your shifter, although you might a lose a gear. There are a fair number of folks with a lot of expertise on adapting and interchangeability, so I am guessing it is doable with some effort.

[BerthaVision]

This is for rides prior to the DRM being added to the TTS files? Ive actually got a drive open with this ride on it right now, just checking out the original data. It looks like we were using at least barometric data back then, the grades should not be too bad, for sure for this ride they are smooth. The speed we averaged was about 22-23kph so about twice what we would do it at now. So your speed is roughly your actual frame rate!

Hi Bertha,

Folks have been telling me about all the features they used to use on the software with their fortius. Cool stuff. You guys had a cool team building this stuff. I bet theyve forgotten more than i learned.

Heres a thing i learned though... And your devs knew it too.

The reality and accuracy of the altitude data plays second fiddle to the user experience. I started by using altitude data, building a spline and deriving gradient from that. This is the path you get if you rode a smooth line through all the location points. My hope was that this 'true' data would be fine so i could simply re-ride my own recorded rides. I found of course that my phone's gps was too noisy in location, and especially altitude noise made them unridable - i had to smooth to remove noise and make slope reasonable.

I expected that the rlvs i had would do some great smoothing and at least those routes would be ridable, but what i find is that none of the rides have 'good enough' gps data. Instead some schmoe made a new field for 'slope' and that is what the user experiences for their virtual ride. The 'slope' is manually curated, nice and smooth and is lovely to ride. In the end i figured since slope is what users experienced with tacx, i will do the same so i derived altitude from slope when im building location data, and then i build a spline and derive a new slope from that while riding ghe route. In that way the altitude 'work' matches tacx yet we still get a continuous gradient. The alternative would be to have fixed slopes which i thought felt hokey.

What i found was that the tacx altitude data is often usable but it also sometimes truely horrific. See start of mortirolo... I spent a lot of time figuring this out by debugging the stelvio ride and its altitude data is not ridable either (note thats it_stelvio08, with the cool old switchbacks on top, not the current lame modern reroute...) There is noise everywhere but i focused on a segment in the lower valley before the first tunnel where tacx altitude has a bunch of meter steps.

Downside of honoring slope over altitude is that the climb total doesnt match reality, upside is that it is much easier to handle video gaps (the tacx stelvio ride has a few) I'm not happy and feel cheated that the virtual climb is short of the actual pass, but im ok now because the virtual ride is its own thing, i'm riding the classic tacx route from 2008, not the stelvio...

I was really impressed by the quality of the tacx rides once i started using slope. Somebody talented went through and made the slope match the screen really well, even if the gps doesnt. High end product and i think the choices were solid. I really appreciate the craft of the rides i have from tacx even if i think abandoning your old customers is loathsome.

Hi Eri

The data is all my domain, you looking at rides that are 12 years old now though, back then we simply did not use the Alt data at all really outside of calculating the slope which could have been modified by me if I thought it was an error, over the years accuracy has improved, its not the accuracy of the bearometric data but also that of the recorded speed & that sample rate you get from filming ever slower. Ive no idea why we had alt data even in the tts files back then , it was just surplus data.. You also need to combine this with the camera we used back then which was actually cutting edge sensor wise, but incredibly unreliable. We would have sections of film missing, the camera would just stop recording & we would be stuck halfway up some mountain trying find to the issue. It was temperamental, lots of cables & connectors that were not designed for the abuse, it was all linked to a laptop & external hard drives, the various power cables etc...a bit of a mess. Since then we made a lot of progress, right now we actually use inclinometer data & filming from a special EBike at cycling speed we capture what must be the most accurate data you can possibly get. Becouse of the slow recording speed when climbing the sample rate is really high also. If a bend flattens out or take the steeper inside line you will feel it. The customer base have no idea Im sure of just how good our data is. Obviously we moved onto cinema cameras, these are massively more reliable & built for a rough life. We also record everything in log & do a full colour edit on each ride... I would say the product your looking at is the Genesis of what we have now..we did not just stay in 2008

[eri]

Along with a few others I was playing at making my own virtual rides. The old tools to make rlv and tts files seem to be gone from internet, or whatever is there is really broken. I was looking at constructing new tools and am able to round trip rlv but the tts format was just too much typing. And that doesn't include writing compatible encryption.

Instead I made it so any reasonable file type can be used to both describe the route/workout and to act as videosync. Gpx will work for this but it isnt great because it doesnt always hold an explicit distance, distance on gpx is determined from the location values, and the distance determination is arbitrary. Some people use euclidian distance but truth is probably some curvy thing that accounts for location noise, no convention. The good news is that speed is the gps' most accurate data because it can use doppler, which means distance can be pretty accurate too even though its independent of location change.

The tcx and fit formats are really flexible about what streams they hold, are text based, public formats, round-trippable. The main benefit though is that they both hold distance and time so can be used as both workout and videosync, no triple of files to get out of sync, now its just workout and video.

With new file support its pretty easy to create a virtual ride:

start video recording (video implicitly holds time)
start gps recording (location, time, maybe speed)
do ride
stop gps recording
stop video recording

At this point you have two streams of data. The gps stream's time brackets the video stream.

Open your gps stream in golden cheetah (fit file, gpx, whatever)
Edit the activity to crop out the time before and after the video (remove the rows from the activity edit window)
Edit the activity so the slope and route are reasonably smooth (smoothing in fixgps)
Probably delete all the data columns that aren't relevent to a workout. It doesn't help or hurt to have your power or heart rate there, but no point either.
Export the activity to your workout dir. Export as some reasonable format: tcx is good
Copy video to your workout dir
Go to train mode,
Scan your workout dir for workouts, you'll see the new video and workout file.
Select training device, workout file, video, and select workout file again in videosync

Boom. Done!

Of course stuff gets more complicated if you want to edit the video because everything you do to video must be done to the workout file. Deal with this a few times and you become very conscious how desirable it is to make the ride in a single take.

The new workout/videosync file support isn't checked in yet but it makes it super easy to make a virtual ride, and nobody needs to re-implement a tts writer.
Map widget isn't checked in yet either but man is that nice.

[mickyduck]

Read this post on the "other" forum... I may have a go at this as you say using the old tools was cumbersome. I have some video just need to find the GPX files that go with them.....

[Foobar]

I do not recall very much in the way of TTS tools as the TTS control files are encrypted. There was quite a bit around for the non-encrypted PGMF and RLV control files used for the old Fortius program, which only worked with a 32-bit OS. I remember a few programs like RLV Cockpit, RLV Studio, and Tacx Run Copy, but they are all very old.

Can someone post a copy of these programs. They are still useful for me but they can't be found!
Just managed to find this
https://web.archive.org/web/20160821093 ... ler-GB.pdf
https://web.archive.org/web/20160821093 ... taller.zip
but all the others are disappeared from the net and Internet Archive Wayback Machine is of no help
Cheers
Foobar

[nzuraw]

https://www.lunicus.org/download/RLV_Creation.rar

[Brave Lee Flea]

I had what seemed like a brilliant idea the other day. I decided to try using modern video AI upscaling on the avi file of my favourite TACX RLV rides. These are quite old, even by TACX RLV standards, and while the video quality is not every good I enjoy the rides.

I used the trial version of Topaz Video Enhance AI and I must say I am really pleased with the results! I increased the resolution to 150% of original (from 768x576 to 1152x864) and it has really sharpened up the image quality, beyond what I thought would be possible.

However the software only writes h264 MP4 files, so I used ffmpeg to repackage the video stream into an avi container and replaced the original avi with my new upscaled version.

I happen to have bought some after-market RLVs from www.reallifevideo.de and I can see that his avi files contain videos in h264 (x264? unsure of the difference) codec so I had thought that I would not need to re-encode.

This does not work.
It turns out that I am not as clever as I thought I was. Which is a shame as I thought I was really jolly clever.

I will try re-econding the video a) back to its original resolution (really don't want to do this as it is still less than the 1920x1080 res of my screen) and b) as an xvid rather than h264 but I wondered if anybody has any experience of doing something similar? If so, do you think it is possible to simply replace the existing video with a better quality one, and if so what are the likely restrictions? Is it possible to modify the TTS4 file appropriately?

[mcorn]

The TTS control file is encrypted.

[Brave Lee Flea]

The TTS control file is encrypted.

Do you understand how the tts control file relates to the avi? I had assumed it might contain time-offsets and/or frame numbers in which case I imagined that there wouldn't be an issue. If it contains information about the resolution for example then presumably all I would have to do is ensure that the resolution is the same as the original?
Is it possible to generate a new tts file if we have say the old pgmf file? Or is the whole thing a non-starter?

[mcorn]

My personal opinion, which may or may not be correct, is it is a non-starter.

If you had a working version of TTS1 or TTS2, which allowed older non-encrypted RLV's to be imported and created a new *.tts file, you might be able to do something. But I really do not know how much flexibility you would gain in terms of using different codecs.

[Brave Lee Flea]

My personal opinion, which may or may not be correct, is it is a non-starter.

If you had a working version of TTS1 or TTS2, which allowed older non-encrypted RLV's to be imported and created a new *.tts file, you might be able to do something. But I really do not know how much flexibility you would gain in terms of using different codecs.

That's a shame but thanks for your advice.
I can still ride the existing vids but they look sooo much nicer after being enhanced.
Maybe I will experiment with the fortius ant+ thing but I am dubious as to how well it will work. TBH I am quite happy with TTS4 and I don't want to incur a monthly cost for software that I would only use 4 or 5 months of the year and would forget to cancel for the rest of the year. I prefer to pay for stuff once.

If anybody has a working copy of TTS1 or 2 and would be willing to help me generate a new tts file please pm me.

[mcorn]

If anybody has a working copy of TTS1 or 2 and would be willing to help me generate a new tts file please pm me.

There is some theoretical possibility that Garmin Tacx could take the position that this is a violation of its intellectual property rights. Given the obsolete status of TTS , this may be remote, but you should at least be aware.

Also, I'm not sure which videos files you have reprocessed. Some of the early TTS RLV's were pretty rough by current standards and were similar to what was used in the previous Fortius software. But the later ones were far better.

[Brave Lee Flea]

There is some theoretical possibility that Garmin Tacx could take the position that this is a violation of its intellectual property rights. Given the obsolete status of TTS , this may be remote, but you should at least be aware.

Also, I'm not sure which videos files you have reprocessed. Some of the early TTS RLV's were pretty rough by current standards and were similar to what was used in the previous Fortius software. But the later ones were far better.

I guess it is a possibility but given that I still have the original DVDs and I only want to use them for personal use in the manner they were intended with the software and hardware they were intended to be used with, albeit with improved video quality, they would have to be extraordinarily determined and entirely unreasonable.

I think the two videos in question are relatively early, specifically T1956.11 Mallorca Tour - Spain (which is a poor 768x576 res) and T1956.50 The Argus Tour 2010 - South Africa (1600x642 res). Both avis are much improved by the image processing.

I like these vids because they are at a nice level for my fitness. They are my "Goldilocks" rides. Neither too hard nor too easy. They are varied with some climbs and some long flat sections. I haven't managed to find any RLV's from other vendors which are similar in nature,

[mcorn]

Those videos are very early and so-so quality at best. They were fine for their time. Argus was letterbox format.

Frankly I am surprised that reprocessing helps much.

[Brave Lee Flea]

Those videos are very early and so-so quality at best. They were fine for their time. Argus was letterbox format.

Frankly I am surprised that reprocessing helps much.

lol, well beauty is in the eye of the beholder I guess!

I was pleasantly surprised at how much better the images became. I mean, it's not perfect, not up to modern standards obviously, but it's noticeably better than it was. Given that I spend quite a lot of hours riding these videos I thought it worth the effort trying to improve them. (Unless you or anybody else can recommend RLVs which are still available of a similar nature?) Besides, I like to tinker and I'm currently in the UK, in full COVID lockdown so it gives me something to do!!

The difference seems more pronounced when the image is moving but here are a couple of shots from Mallorca, you might need to download the images to see them better.

Before
https://drive.google.com/file/d/1g_kK9r ... sp=sharing

After
https://drive.google.com/file/d/1n2JshW ... sp=sharing

It's all sharper and more defined.

This is a side by side comparison of the Argus Tour (which was actually far better quality than the Mallorca vid and has less improvement). Ignore the red line the image is actually a preview of the encoding as it is being processed.

https://drive.google.com/file/d/1lyIuYd ... sp=sharing

[mickyduck]

That is impressive maybe I should do my old Ventoux ... but I feel the sweat in my eyes would blur it quite a lot anyway....

[Brave Lee Flea]

Well, I kept mucking around with this.

Today I found that if I re-encode the x264 mp4 file using ffmpeg specifying that the code should be XVID (which was what the original avi was) then it works just fine, even with greater resolution.

The video is improved but I think it can still be better, the software is cropping the video because my screen is a modern aspect ratio (1920x1080) and the original video is not (768x576). Having cropped the video, it then enlarges what's left, which it doesn't do nicely.

I'm going to experiment with the settings and resolutions until I get the best outcome but it seems that this can work, which is nice.

I'm also going to try Golden Cheetah as Micky suggested, but I haven't gotten round to this yet.