In/Out Device

amalguy

New User
the devices records at 720p with 30fps ; thats the only way how the device can replay so many shots, the replay cam Shows only black and White colour to save space
what else should it have?

120fps? do you know how much storage capacitiy is needed for 120fps Videos in a high resolution?
The device use differents framerate for recording and stats/line calling :
  • For recording :
    • 30fps at 1920x1440 for entire court
    • 46fps at 1080p for one half court (46fps is get from my tests)
    • 120fps in slow-motion at 1540x504 for one half court : Real measured framerate rather 95-105fps
  • For Linecall/stats :
    • 120fps if my reasoning is good. This video stream is not stored
 

RayF

New User
The device use differents framerate for recording and stats/line calling :
  • For recording :
    • 30fps at 1920x1440 for entire court
    • 46fps at 1080p for one half court (46fps is get from my tests)
    • 120fps in slow-motion at 1540x504 for one half court : Real measured framerate rather 95-105fps
  • For Linecall/stats :
    • 120fps if my reasoning is good. This video stream is not stored
firmware 06/06 just came out - no comment about improved line calling accuracy - since you are in contact with Gregoire, do you know if any improvement has been made in the algorhythm? - thanks
 

amalguy

New User
firmware 06/06 just came out - no comment about improved line calling accuracy - since you are in contact with Gregoire, do you know if any improvement has been made in the algorhythm? - thanks
I don't think there is any improvement in accuracy but there is some in the detection part :
  • Improvment of the analyzed wiew close to the net. From Gregoire : "In 2018-05-05, for the serve, the detection is much closer to the net. And it has been generalized to all bounces in 2018-06-06. The analyzed view doesn't still go up to the net but it's close now." - Very important for me
  • From Gregoire "Important fix coming in 2018-06-06 that enhances the reliability of detection". But I don't know what is it. I have to ask Gregoire
 

RayF

New User
I don't have high speed camera but clay courts ;) and it's better for doing accuracy test.
I made such tests with the 2018-02-02 firmware and with the latest 2018-05-05.
And there is great improvements between them.

With the 2018-05-05 firmware, the average difference, in cm, between the position computed from the device and real position is 6.6cm in the length of the court and 3.8cm in the width. Even better if the bounce is in the half court near the device : 4.5cm in the length and 2.2cm in the width (against 15cm and 18.3cm with 2018-02-02 firmware).

The results are detailed in the two following pictures :

  • test with 2018-05-05 firmware
  • test with 2018-02-02 firmware
And my process is detailed in the following picture:


The accuracy is starting to be very interesting. And don't forget that there is another process for linecall : pure computer vision to "see" if the ball is "before" or "after" the line.
Almalguy, have you thought about doing a detailed analysis of the accuracy near the lines? As you mention there is a separate algorithm that determines in or out based on position vs line and I'm wondering if the accuracy there is materially better than the measurements you took - which seem to be mainly for balls landing well inside the court. To do this experiment I assume you'd need a precise ball machine capable of aiming at the line (perhaps Gregoire is using something similar to refine his device?). After all that is really the most important function of this device - thanks
 

amalguy

New User
Almalguy, have you thought about doing a detailed analysis of the accuracy near the lines? As you mention there is a separate algorithm that determines in or out based on position vs line and I'm wondering if the accuracy there is materially better than the measurements you took - which seem to be mainly for balls landing well inside the court. To do this experiment I assume you'd need a precise ball machine capable of aiming at the line (perhaps Gregoire is using something similar to refine his device?). After all that is really the most important function of this device - thanks
No I didn't think about a such test because, for me :
  1. We can't yet measure the accuracy of a close call based on the "computer vision" algorithm because this program only determine if the ball is in or out and don't measure the distance between the ball and the line. In fact it does the job but only in pixels (viewable with debug button in the bounce's replay). Gregoire told me that "centimeters will be provided soon" 3 months ago. So I will ask him again.
  2. My first goal is to use the In/Out device to get stats of traning/real matches of my son or myself. Then I try to figure out if if can trust the device in this context.
For me the more important is to have reliable devices to be sure, when I use them on court, that I will get, at least each time I have good calibration court score, reliable stats (no crash/stop of the session, a maximum of bounces recognized, not too many false positive, ...) with known accuracy and an exploitable video (no broken video, acceptable resolution, no bug like camera switch, big flickering, part of video occulted, ...)
 

heat010

Rookie
Interesting testing today. I was one of the first batch adopters/purchasers of the device. In fact I hit with Gregoire when he was doing some indoor testing awhile ago. Well I really hadn’t pulled it out and used it for quite a while since I got it since I wasn’t pleased with how it ran. I installed the most current software update and played about 2.5 sets of singles. It was a cloudy overcast day with light drizzle. The device picked up the court on the first try which surprised me. The battery level indicator was all over the place, but I ran it in line call mode only and it kept running for over and hour with not much issues. I would say it was about 90% accurate during our sets. It missed calls on some sideline shots and some serves that were near the back of the service line. When they were missed we just replayed the point. It was a fairly satisfying round of testing. Will keep using it and see how the results are in brighter conditions
 

Nostradamus

Bionic Poster
The device use differents framerate for recording and stats/line calling :
  • For recording :
    • 30fps at 1920x1440 for entire court
    • 46fps at 1080p for one half court (46fps is get from my tests)
    • 120fps in slow-motion at 1540x504 for one half court : Real measured framerate rather 95-105fps
  • For Linecall/stats :
    • 120fps if my reasoning is good. This video stream is not stored
Are they ever going to come out with even better more accurate version sometime ?? with slight improvement, this device can be used in All USTA tournaments and league matches to improve the line calls. Too many cheating going on in important points in USTA leagues especially if your opponents think you don't have a clean view.
 

tennisnut33

New User
Has anyone had any issues with the link to download the firmware for version 2018-07-07? The link on support.inout.tennis doesn't seem to work.
 
Are they ever going to come out with even better more accurate version sometime ?? with slight improvement, this device can be used in All USTA tournaments and league matches to improve the line calls. Too many cheating going on in important points in USTA leagues especially if your opponents think you don't have a clean view.
Their road map says they will not work on accuracy any more.

Do you need 2 machines for doubles? I cannot get mine to call doubles.
 
With the latest update, it seems like the detection is quite good. Court detection works on the first try almost every time and with serving practice, it detects 95% of my serves.

Note that I did keep it on a tri-pod rather than above the net post when I recently tested it. The problem with using the net post is that any good shake of the net ruins its positioning and accuracy heavily deteriorates.
 
With the latest update, it seems like the detection is quite good. Court detection works on the first try almost every time and with serving practice, it detects 95% of my serves.

Note that I did keep it on a tri-pod rather than above the net post when I recently tested it. The problem with using the net post is that any good shake of the net ruins its positioning and accuracy heavily deteriorates.
I started using a strap across the base under the device to keep it stable and prevent it from moving .the device moves too easily on the base.
 
One of my other frustrations is that the fairly new battery will not run the device. I need to plug in an external battery to the device battery to keep it running for more than a hand full of seconds. Again, I wish a company with hardware experience was involved in the project. It is flimsy and poorly designed.
 

Dartagnan64

G.O.A.T.
So it sounds like this product never lived up to the hype. Thanks to all you alpha and beta testers that put it through its hoops. Admittedly a frustrating experience I'm sure but a lot of great products needed some failures along the way. Hopefully the next iteration of a product like this will be more useful. I'm sure hawkeye level sensors will be the future.
 

Mika I

New User
So it sounds like this product never lived up to the hype. Thanks to all you alpha and beta testers that put it through its hoops. Admittedly a frustrating experience I'm sure but a lot of great products needed some failures along the way. Hopefully the next iteration of a product like this will be more useful. I'm sure hawkeye level sensors will be the future.
This was some interesting reading. Four cameras they had.

https://web.stanford.edu/class/cs231a/prev_projects_2016/final_report_v2.pdf
 
Well, my device will no longer power on and the battery just shows red when I charge. I suspect it is a battery issue.

Too bad, as I thought it was getting closer. It seems that he did not implement score. That will be tough. I was also looking forward to trying the watch functionality.

If someone is successful with doubles, I may buy another one with an extra battery. I still think the best thing to happen would be for Wilson or someone to buy him out.
 

CarterStC

New User
I guess I was the lucky one - owned it, struggled with it, and got a refund (mostly). If he keeps pushing sales with the current hardware, especially to organizations, he's going to end up with a class-action lawsuit on his hands. The website really makes the device look functional, but it is so flimsy and worthless, it is only the delusion of hope that keeps a customer's trying long enough to expire the warranty. However, California's implied warranty laws leave him heavily exposed.
 
Yes, the hardware seems to be poorly designed. I live in CA as well. I just submitted a ticket for my non functional device.

If Sony bought this and combined it with their sensor, we would really have something.
 
I know this device hasn't exactly gotten rave reviews on this message board but if anyone's interested in buying mine at a discount, let me know. From what I read, it still can't be trusted with line calls but maybe there are other features someone would find useful but doesn't want to pay full price.
 

Beacon Hill

Hall of Fame
I know this device hasn't exactly gotten rave reviews on this message board but if anyone's interested in buying mine at a discount, let me know. From what I read, it still can't be trusted with line calls but maybe there are other features someone would find useful but doesn't want to pay full price.
How have you found it? Does it work, and just not very accurate? How is the battery?
 
I feel the accuracy has improved but like others have said, it does suffer from the net vibrations and seems to be poorly made. Tripod is the way to go and I wish they would have a better way of determining lobs. Once the ball goes out of view, it seems to quit sensing it? @Buddy Koehler What do you want for it? I might be interested to test the double feature of it.
 

RayF

New User
No I didn't think about a such test because, for me :
  1. We can't yet measure the accuracy of a close call based on the "computer vision" algorithm because this program only determine if the ball is in or out and don't measure the distance between the ball and the line. In fact it does the job but only in pixels (viewable with debug button in the bounce's replay). Gregoire told me that "centimeters will be provided soon" 3 months ago. So I will ask him again.
  2. My first goal is to use the In/Out device to get stats of traning/real matches of my son or myself. Then I try to figure out if if can trust the device in this context.
For me the more important is to have reliable devices to be sure, when I use them on court, that I will get, at least each time I have good calibration court score, reliable stats (no crash/stop of the session, a maximum of bounces recognized, not too many false positive, ...) with known accuracy and an exploitable video (no broken video, acceptable resolution, no bug like camera switch, big flickering, part of video occulted, ...)
What I had in mind was a ball machine sending balls very close to the base line, then one could measure manually the distance between the ball mark and the line to determine the accuracy of an 'in our out' call. Anyway if you speak to Gregoire it would be interesting to see if he is doing this kind of test. Also interesting to see Mika1's post on Aug 23 where it seems like a university student achieved 8mm precision (much better than in/out) with a 4-camera system....and maybe a better algorythm?
 

West Coast Ace

G.O.A.T.
What I had in mind was a ball machine sending balls very close to the base line, then one could measure manually the distance between the ball mark and the line to determine the accuracy of an 'in our out' call.
I offered that up many pages ago - throw down talcum or baby powder so there’s a mark that can be measured. It would be a very quick, easy test.
 
Looks like essential tennis did a whole video on the in/out.

I think it says something that the inventor is still out there pushing his product. The video shows two in out devices being used for singles, and seems like he’s really pushing out the types of data the device can try to capture like average height over net etc...
 
I think he is struggling to sell product. I just had to buy a new battery, which is my third since getting the device. Poor hardware design on both the batteries and device.

I like the new custom zones functionality. Has anyone tried 2 devices?
 
Last edited:
I think he is struggling to sell product. I just had to buy a new battery, which is my third since getting the device. Poor hardware design on both the batteries and device.

I like the new custom zones functionality. Has anyone tried 2 devices?
Honestly he probably should have just made the in/out default order cost $500 for two cameras and two tripods and not give the option of just one dangling on the net.

I can’t speak for anyone else but I’ve had zero battery problems itself, just that the battery life display being inaccurate on the touch screen.

I wouldn’t mind buying another one because using just one plus a tripod for serving practice seems really useful. However with everyone here having product issues, I’m hesitant to commit.
 

Beacon Hill

Hall of Fame
With 2 devices, I believe so.

The saddest part of that video was when he said the hardware was good and he had no changes planned. I hope that is just because he has lots of inventory and is not his true assessment of the hardware.
That was disappointing.
The company today sent me an email that the device now supports doubles "very experimentally". I'm encouraged by the product, but understand its shortcomings.
 

bertrevert

Hall of Fame
How has its stats been helpful/useless to you?
I imagine ball height over net being helpful?
I don't know what else it does but are there stats you are gravitating towards that teach you about your game?
 

DarthKYS

New User
LOL my school courts are blue and white on the outside. It's newly resurfaced so it's freaking hard to see the lines on the other side. I'm pretty sure this man will have a hard time seeing base lines and doubles XD
 
The smart watch integration is very cool. Full mirroring of the device so you can see replays or where the shot landed. It also buzzes when out.
 
It will show you the video of the call. It even streams it to your connected watch.

I bought a second one and doubles is still not active. Only singles. I will set them up for some doubles tomorrow.
 

bertrevert

Hall of Fame
"Moment of Truth" technology footage in recent Chinese tournaments was interesting.

It combines hawkeye plus court level cameras (up at net on the ground).
http://www.foxtenn.com/in&out

and we had sml discussion here
https://tt.tennis-warehouse.com/index.php?threads/the-moment-of-truth.628667/

so for us recreational/comp/club players I think a few things are apparent

1. we have to be more generous with our linecalls
2. we have to bring technology to what our eyes see
3. ball flattens at impact and when it reforms to bounce it looks like it is out, it has brushed the line

so we prob all need some sort of electronics as standard on our courts...

now if only there was an app for that :p

oh if we could only mount our mobile phones on net posts and make it bleep when the ball was out!
 

Dartagnan64

G.O.A.T.
"Moment of Truth" technology footage in recent Chinese tournaments was interesting.

It combines hawkeye plus court level cameras (up at net on the ground).
http://www.foxtenn.com/in&out

and we had sml discussion here
https://tt.tennis-warehouse.com/index.php?threads/the-moment-of-truth.628667/

so for us recreational/comp/club players I think a few things are apparent

1. we have to be more generous with our linecalls
2. we have to bring technology to what our eyes see
3. ball flattens at impact and when it reforms to bounce it looks like it is out, it has brushed the line

so we prob all need some sort of electronics as standard on our courts...

now if only there was an app for that :p

oh if we could only mount our mobile phones on net posts and make it bleep when the ball was out!
Admittedly if everyone calls those balls that skid on the line as "out" then there is no problem as long as consistency exists. Consistency amongst tennis players is more important than accuracy. If everyone has the same bias, things will even out. Problems arise when someone clearly makes biased calls that his opponents wouldn't.
 

bertrevert

Hall of Fame
as long as consistency exists
Hang on we're talking about players here, players in motion.
Varying ages and abilities, under varying light and conditions.
There is no consistency in most matches, we just muddle through.
Certain players acquire a reputation for error, others are generous.
Sometimes I have to reflect - we play an odd game that requires more umpires than players on crt.
And most of us play our whole lives without those requisite umpires.
We need to add the electronics... for any hope of consistency.
 
Top