Mana Games

[manutoo]

TEM2 v1.0b !

Pfff, this is a big one done here ! I feel more heart lighted already !

Indeed, now TEM2 can be seen as complete, although I'll likely continue to improve it over the years, alongside TE4.

And the main things done in the v1.0b are also for TE4, and the 1st one is even more important in TE4 than in TEM2 : now, when starting on the Junior Tour, it'll be much easier to win your first matches, and then your 1st tournament. Before it could take years !

Also, the fictional players will have a better variety in their generated names as I added an additional player name base for that only purpose.

As expected, making the game handles the old Tour bases & new ones at the same time proved to be a nest of bugs, but apparently not as bad as I had expected, as after a couple of updates, the v1.0b seems stable enough. The most tricky part was to handle the Load/Save screen to show or not the "Invalid" state ; it now involves detecting & caching some specific details for each Tour base and thus requires to load them depending on the exact configuration : men/women, new/old bases, modded/unmodded, game in progress bases / saved game bases.

But it's a really good thing done, because now I won't have to hesitate so much if I feel the Tour bases should be updated, as the users won't lose their current saved games anymore.

It'll create a few additional headaches though : some users will be stuck with old bugs from the old bases, as long as they don't start a new career. For example, there were a lot of players not playing in the year 1992, making a few draws with missing players. So these users might report thing that are long fixed, but I'll have to check every time if they use the new bases or the old ones.



TE4 all the way !

This means now I can focus nearly 100% on TE4, and hopefully the pace will be pick up a bit. Nothing too crazy as I still have a young boy plus a baby at home, who eat up a lot of my time.

The next step is to finally finish the integration of the female customizations (mostly the hairs). It should be done by the end of the week (I feel I have already written this 2 months ago ).

The male production should be nearly done soon as well.

Then I'll focus on the gameplay, to iron out the little issues that keep being reported. Once done, the game may be ready for the Early Access tag, and gets a demo shortly afterward.

So cu soon enough to see how all this will go. (likely horribly, as usually )

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=267029#p267029

[manutoo]

Dev Update

So I didn't do much in the past 4 weeks...

I just finished a few days ago the customization of the female player. It's all done, but there are still a few glitches in the 3D model so I have to wait for the 3D artist to send me the corrected version. He's quite slow lately, so it's unnerving, especially when he sends me a new version with "everything fixed", but actually less than half is done.

Also, it still misses a few face modifications, but there again, on my side it's more or less alright already.

I had to do a little script to automatically add 2-sided polygons where it was needed, because for the shirt he did the polygons only on 1 side of the collar, which obviously looks bad in some direction. At 1st, I just put a double-sided shader on the whole thing, but in terms of optimization it was really bad.

I also did the shader to display the damaged grass, by combining the grass texture with a dirt texture. It works pretty well, and it's better than having only 1 texture, as it allows high details for close up cameras without having too gigantic textures nor detailed geometry.

And I had to fix a few bad bugs on TEM2 ; but don't worry, it was needed for TE4, as it was mostly about the new possibility to switch between Tour bases, which will allow me to update TEM2 & TE4 Tour bases if the need arises, without invalidating your current saved games, as they will just keep on using the previous Tour bases.

The 3D male player should be almost done by the 3D artist ; I got an intermediary version a couple of weeks ago and it looks like this :

2020-06 - Player Male - Tennis Mafioso.jpg

And this morning, I studied how to make a patch system. The game installer is already nearly 1GB big, and it'll likely grow to 2GB, if not more. I can't ask you to re-download the whole thing every time I'll change 2 lines of code to tune the gameplay, or fix a bug, especially when you'll be a lot of people playing the game, as it'd crush my server and you'll need forever to get the updated version. So I have to create a patch much smaller than that. It's not easy to do, especially on Mac, so I guess it'll be only for Windows, at least for a while. I think I'm just going to use a plugging for the NSis installer I'm using. I'll test it soon and hopefully it'll do the job.

So overall, as you can see, I took things slow. My son still can't go back to school, so he's eating my time & my energy way more than I'd have wished, even if his grandmother takes care of him often. This means I'm almost always tired, so I do what I can do.

Cu right away in the next post for some surprising fact !

A comment, cheer or question about this mini-blog ? => topic15-31540.php

[manutoo]

What is the most played paid tennis game on PC in the whole PC history ?

By "most played", I mean the cumulated hours ; eg : the sum of all played hours by all players of that tennis game. It's not the number of people who have played the game.

So think a bit about it...

You might have guessed, else I'd have not talked about it, but it's very likely TE2013 (or more exactly TE3 as TE2009/11/13)..!

LibraryLogo_1280x720.png

I always check what is the most played tennis game on Steam with this : https://steamcharts.com/cmp/346470,1072 ... ,758410#6m .

But it's only recently I realized this chart meant that TE2013 was the most played tennis game in the history of Steam, which gave it a chance at being the most played on PC. At 1st, I thought it was a silly idea, but then I did some math.

1st, the number of total players of TE2013 isn't what you see on this graph. You likely have to double these numbers to include the ones playing with the Mana Games version. I don't have exact numbers for the ones playing offline, as I only have the "checking for a new version" stat for these users, but for the online, it multiplies by 2 or more, without counting the demo players. Offline numbers are compatible with that (actually it's more like x3, but I don't know how much people check for a new version more often or less often than the automatic every 7 days). A recent new stats from Steam helped me to confirm that : they now show the average daily users on the past 7 days, which is a stat I can approximate as well from the new-version checking. Lastly, about these numbers, I consider that players on Mana Games play as much the game as Steam players, but actually they might play a bit more in average because more of them tried the demo before buying the game, so they are more likely to enjoy the game than Steam players.

All the games in the charts above are Steam exclusive (except for TE2013), this means they don't have more players than shown there.

So these charts give cumulated players for each month (it's proportional to the total played hours, not the total players), so it's easy to do a grand total :
- TE2013 : 2462, since March 2015
- AOT1 : 493, since May 2018
- TWT : 361, since June 2018
- AOT2 : 391, since January 2020 ; TE2013 beats it in the past 30 days, so if this tendency doesn't change, this game won't beat TE2013's total, even without invoking MG players

There's also Virtua Tennis 4, which hasn't been sold on PC for many years : 934, since July 2012
VT4 has been sold on Microsoft platform, but likely way less than on Steam, as its total sales on PC are estimated at 40k units by VgChartz : https://www.vgchartz.com/game/52501/vir ... region=All .

So the 2 big missing games are Top Spin 1 & Top Spin 2 ; VgChartz doesn't have sale numbers for their PC version (actually they show "17" for Top Spin 2 on PC, but it seems a bit low ), but checking the sale estimates for the consoles and the ratio console/pc for Virtua Tennis, it doesn't seem possible that Topspin sold enough units to beat TE2013 Steam score, and even less for TE2013 Steam + MG, and even less for TE3 (you likely can multiply TE2013 Steam score by 3 to approximate it), or its players should have played the game like crazy.

So I very likely created the most played paid tennis game in the History of PC.

Fuck yeah !

So thanks for all your help, without all your feedback and without you playing the hell out of TE3 it'd not have been possible !

And let's see if you can help TE4 to take the title on consoles as well ! It might be harder, but while there's life there's hope !


Cu next time with hopefully a new version and some gameplay improvements !

A comment, cheer or question about this mini-blog ? => topic15-31540.php

[manutoo]

Advancing...

So the female player is finally done ! (except maybe for a few facial tuning)

And thus I could update the game after a 2-month break, adding all the stuff from TEM2 (eg: improved Junior tour).

As usual, the pace has been slow but relatively steady. My son got back to school for 2 weeks and now the summer holidays have started, so the situation will depend again on my mum's energy to take care of him.


Into the heart of tennis

Anyway I also finally worked on the gameplay, tuning a bit the biggest issues reported so far : too many aces, too weak topspin, overpowered sliced serve & aggressive slice bounce.

So this was quite refreshing for me to work on this kind of things after a couple of years porting the engine to Unity, doing production, world tour, stadiums, and other stuff not related to the core fun of what is a tennis simulation.

And by working on the slice bounce, I discovered something very important : my physics formula for the spin after the bounce didn't work in many cases, and it mattered a damn lot..!

Indeed that spin after bounce determines a lot how high and how far the ball will travel after the bounce. At least as much as the ball speed, and even significantly more in extreme cases.

So that formula has to be right and thus for all cases, else we get overpowered or underpowered serve, slice, topspin and so on..!

I quickly checked the concurrent tennis games, and none got it right, at least not for all the common cases. Usually, they just tune their bounce so the ball goes nicely around hip height so it matches their animations. (ok, it was not a surprise the other games weren't much realistic in that area, but I wanted to be sure before writing this post )

2020-07 - TE4 - Ball Bounce.jpg

It's relatively easy to get the ball-in-the-air physics right (although many tennis games fail on that as well), but getting the bounce right is a real challenge. Even TE2013 is mostly wrong about it, although it looks good enough if you don't look too closely, because almost every shot performs relatively close to real life. There's 1 big exception to that : the drop shot, which just travels way too much after its bounce, and coupled with the players' high running speed, it made it a noticeably underperforming shot.

Unfortunately, it's very hard to find compelling data showing how much spin is produced in real life after the bounce, so I did some tuning based on videos, in different cases, till it felt alright. I may tune it some more if I find some extreme cases that don't work correctly, but so far, I think this new version works pretty well.

The most satisfying part was to be able to prevent the CPU to do too many winners by using only the topspin strike, which was helped by the new charging rule to get the best acceleration possible : now you have to charge from the ideal positioning for 0.25s, else your acceleration is a bit less strong.

It felt like playing TE4 for the 1st time..!

The road is still long to get all the gameplay together, but it sure feels like it'll be worth it.


Next Tasks

So right now, I'm still waiting for an update to the update system I bought (no pun intended ). Once it'll be there, I'll use it so updating to a new version of TE4 will be easier & faster for you.

I started to look into dynamic IK to reposition the racket in real-time, so it should always have a correct contact. I'm not sure if it'll work well, but if it does, then it'll change a lot how we estimate if we strike the ball correctly or not : instead of checking if we hit it with the racket cords, we'll have to check if our player posture is correct or not.

And the final male player version is coming as well (still with possibly a few facial tuning to add later on). So I'll have to check it ; hopefully, the process will be faster than for the female...


Cu next time to see the next little bricks put into Tennis Elbow 4 !

A comment, cheer or question about this mini-blog ? => topic15-31540.php

[manutoo]

This is Revolution !

So after I wrote that previous blog post, I got a couple of weeks with barely any development done : as often, I got taken in everyday business work & debugging or other little stuff, plus I also have been taken by personal life once more, and as a bonus my mother couldn't take care of my son as much as before, so all in all, I almost didn't do anything...

But the last couple of weeks have gotten better and I have been able to work very seriously on the new dynamic animation system, and the result is a... revolution !

But have a quick look here :

2020-08 - TE4 - Dynamic Animation - Lower Hit Position.jpg

If you know TE4 (or TE2013 or TEM2) well, you know there's no strike animation where the player puts his knee almost on the floor. But here it is !

And all that is done in real-time, the whole animation, all animations, are now adapted to the ball position, so there's always a perfect contact between the racket and the ball on hit (when the reach autorizes it, of course). This is a huge change, both visually and in term of gameplay : now, if the ball is too high, too low or too far, you can see it very clearly as your player adapts his body and strike accordingly.

The result is very impressive ; have another look :

2020-08 - TE4 - Dynamic Animation - Stretching.jpg

Once again, a never-seen-before-in-TE animation..!

So not only this, but I also improved the hit prediction (else the IK system really failed, so I had no choice but to do it ), the strike on the run (the racket positioning was widely wrong on these ones), and hitting when walking back where the whole body rotation could be even worse than when running to the sides.

So all in all, TE4 got a huge, big revamping & improvement on the animation side (and it's hopefully not over ; it should likely have new animations within 1 year ).

One compartment that especially benefited from that boost is the volleying : now, the racket goes on the high and low balls, and it really helps to read what's going on and have a better feeling for playing :

2020-08 - TE4 - Dynamic Animation - High Volley.jpg

Now, for the bad, as everything is unfortunately not perfect :
- to make the contact perfect, the speed of the animation is slightly changed ; at normal speed, I can't really see it, but at 50% speed, it's clear for me, so if your eyes are better than mine, it might be shocking for you at normal speed ; if it does, let me know and I'll try to work on a solution (there are several ways to deal with that, all are complicated, so if no one complains about it, I won't do anything )
- the adaptation is sometimes a bit abrupt, or look a bit weird in some extreme cases ; once again, I'll need your feedback to know what is acceptable, and what isn't
- some animations are not adapted to the change ; for example, on the high volley above, it looks good enough on the screenshot, but in-game, the animation is very strange and there it'd really need a hand-made high volley animation
- same for the stretching; it will still need some specific hand-made animations there
- a mistiming produces an even worst animation glitch, so I'll have to ensure they never happen anymore (ie: correct the animation timing ASAP the mistiming is detected)
- sometimes the arms or the legs go into each other ; I did some stuff to prevent that, but it seems impossible to avoid that in all cases, especially for the arms

Also, it significantly raises your player reach (by ~40cm) and thus it'll require some tuning in the strike physics to take this into account, but this one was expectable.

So I'm really impatient for you to test all that to know where the game stands with these changes, but it'll have to wait a couple more weeks, because 1st I have to finish all this :
- new dynamic animation for the head to look at the ball ; it'll be more natural as the upper body will move a bit as well
- smooth the transition in the dynamic animations (eg: when the player stops to run while doing his strike, it can jerk a bit)
- get the off-hand on the racket for 2-handed strikes (it's going to be a long & painful task )
- fix the replay after a skipped point (the dynamic animation system is really not compatible with that feature for the moment)
- integrate the finished male model (it's almost done now), which should take me a couple of days
- integrate the new updating system so updating the game will be easier & faster
- tune & balance a bit more the topspin & acceleration efficiency, so defending is possible (eg: Nadal should win sometimes )


Charging ? Yes & no (or the opposite)

In the previous mini-blog post, I talked about the new charge system for the acceleration & shoulder-height acceleration bonus.

This word is a bit misleading and of course, it doesn't mean your player is putting his elbow in the back for a long time to charge his power (Sir, we're not in Dragon Ball Z here, in case someone hadn't noticed ).

This "charging" time actually represents the time you need to perfectly position yourself, go a bit down on the legs, then push up & forward to hit the hell out of the ball ; like here :


And you still can hit up to 150km/h missiles on the run if you manage to be perfectly positioned ; it's just for the 188km/h rockets that you'll need that extra preparation time...


Hopefully cu in a couple of weeks for a new version with a lot of stuff to test out !

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=270695#p270695

[manutoo]

When does it get too silly ?

So as usual, things went slower than hoped, but still, there's progress.

The bases of the new dynamic system are now done and the worst common cases are fixed. Now there are still a good bunch of less common cases to tune & fix (some being not that uncommon ).

Although some might not be possible to really fix ; eg: a very low ball a bit far will make your player bends strangely. Should I let it like that or should I make that kind of ball impossible to hit ? Personally, I'd rather be able to hit the ball, but in term of marketing it might make the game looks silly, so I'm not sure what direction to take.

So if you already pre-ordered Tennis Elbow 4 , don't hesitate to try the latest Alpha version and tell me what you think about all this here : topic15-31985.php .


Into the heart of gameplay & realism (and vice versa)

So for me, the more important part is now to get a good balance between the defense & the attack, with the topspin being a good defensive weapon. To check all that, one of my tools is to play a CPU vs CPU match on clay between Federer & Nadal. Federer was beating Nadal all the time in the previous version and now after many tuning they are close to 50/50 points won on each side, but still with an advantage for Fed. I'd like that advantage to switch to Nadal, but I've noticed that Nadal doesn't use much topspin and does dumb passing-shots way too often, so I guess there's a couple of IA issues there to fix 1st, and then maybe it'll be already alright..! (one can always hope, at least )

So some of this tuning was to make the Fed's backhand weaker against high balls. I didn't over-do it and the result is interesting : he still can hit accelerations sometimes, but they lack power & precision.

To achieve that result, I also introduced a new very important gameplay concept : now if you hit the ball during its descending phase, and even more during its ascending phase, your accelerations will get significantly slower.

This alone mostly fixed the issue with the acceleration feast we were having in the previous version. I'm not sure how realistic this concept is, but from my own experience, I'm relatively confident it doesn't betray the real tennis experience.

The concept of ascending/descending ball is already in the game, through ball difficulty, which then translates into different consequences, but the effect wasn't strong enough most of the time. So now the link is direct and the effect strength easily tunable.


Tennis is actually not that fast

While working on the acceleration speed, I came toward one big issue : the maximal backhand acceleration speed was very close to the maximal forehand acceleration speed (without counting the Shoulder Height bonus which is significantly less on the BH side). It worked like that in TE2013 and it wasn't that disturbing because the maximal speed wasn't that high : it was ~150km/h against ~160km/h now with a 0% topspin skill, and ~130km/h vs ~150km/h with a 100% topspin skill.

So till the previous version, we could see plenty of very fast & efficient backhand accelerations. Thinking about all that, I looked into the maximal recorded BH speed in real life : the record seems to be held by Nalbandian with 170km/h. Nadal stands not far with a ~166km/h one. Knowing the fastest FH was the Blake's one around ~200km/h, I thought I'll just proportionally lower the backhand fastest speed accordingly.

And then I remembered I saw a site showing statistics for BH at the AO grand slam, and Wawrinka got the fastest one ; which was in par with my idea that the 1H-BH has a better potential for speed in the common case.

This is that page : http://on-the-t.com/2016/10/22/aoleader ... and-speed/ .

ATP Backhand Speeds - AO 2014-2016.png

So I checked the average speed and thought I could maybe add that stat in TE4 to see how it compares. And then I noticed the actual Wawrinka maximum speed : it's only 134km/h..! So over 3 years, the fastest recorded BH was just a bit faster than a common good strike (110~120km/h), or the Wawrinka median (120km/h)..!

This made me realize that having half of the backhand accelerations going at 140+km/h instead of 150+km/h would still not be realistic enough. Which was good because gameplay-wise, it was still too efficient.

So now, I might have gone overboard as with the new ascending/descending phase I hit plenty of 115km/h BH accelerations (with 100% topspin), but like this we are very likely closer to the reality, especially as it's still possible to hit a 140km/h BH when everything is nearly perfectly in order.

Seeing the result on the FH side confirmed this. And more than that, a well-placed 120km/h strike is still very efficient, and if it doesn't guarantee a winner, it often gives a great opportunity to conclude on the next racket hit. So gameplay-wise, I think we are arriving in good lands. Now, there's still a lot of tuning to do, but the roots seem better positioned.

Funnily, almost whatever changes I threw on the acceleration speed, it didn't change much the results of my Nadal-Federer matches. Lower speed made the average rally length raise a bit, but nothing too crazy (from 5.1 to 5.8 at maximum). I guess it just shows that top tennis players can adapt to the conditions and always play their best tennis...

All these tests lead me to tune the use of the acceleration by the AI, and apparently, it became much more effective (when testing the new smarter AI vs the old one), so you may feel it's harder to play it than before.

At last but not least, thanks to the extending reach brought by the dynamic animation system, the ace numbers are much more realistic, at least on clay, as I didn't test on other surfaces so far.


So all in all, if things go slowly, it's also because there's a lot stuff to do...

The male model is now more or less done (except some of his facial configuration), so I'll try to integrate him into the game this week.

Cu next time for more dev news !

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=271447#p271447

[manutoo]

Basic Dev Update

Once more, this post is firstly for me to keep track of what's going on before I forget everything...

So in these last 3 weeks, I did mainly 3 things.

The 1st one has been to convert the outfit texture creation to the new Burst mode made by Unity. This made the outfit texture processing more than 10 times faster ! This allowed me to raise the texture resolution of the male player (I did only the skin & hairs for the female, and I'll do her clothes later on), which is mostly visible in the menus, especially if you play in 1440p or above and with the 3rd person camera mode where our player takes a big part of the screen.

So it's useless 95% of the time, but it'll help to give a better 1st impression of the game, and it'll also be useful if I ever do some cutscenes with close-ups on the players.

The 2nd thing has been to integrate all the new male outfit stuff. It has been a bit longer than planned, partly because there were a lot of small & bigger glitches ; some were really surprising. So I spent a long time explaining what's wrong with the player. Here's a typical screenshot I sent to the artists ; on this screenshot, you can see the new player version has his shoulder that doesn't deform as smoothly as the version 1 from last December (in the middle) ; on the right, it's the model in the current version of the game, and on the left it's the corrected version that is still not 100% correct :

2020-09 - Male Player - Shoulder Rigging Little Issue #1.jpg

Hopefully, it'll be all fixed within a few days.

C# sucks

And the 3rd thing was to convert the Bicubic downscaling to Burst, which is used for the mipmapping.

Note : a mipmap = a lower resolution version of a texture, shown when the texture is small on the screen.

Unity, the dev engine I'm using, can use only 2 algorithms to create the mipmaps : a classic average one, which makes everything a bit blurry, and the Kaiser one, which is both too aliased and still a bit blurry.

Bicubic downscaling is overall slightly superior ; it is especially good when there's a text to read in the texture (eg: the ads on the court). In some cases, it can be seen as vastly superior (depending on your taste) as it'll be able to keep the contrast generated from high-frequency patterns, that would just be blurred with the other algorithms and thus look flat in the mipmaps. This is especially true when using a sharpening version of the Bicubic algo like I do.

The huge downside is that it was really very slow because I could implement only in C#, and thus I couldn't use it in-game ; and even in preprocessing the textures it was just too damn slow, so I had turned it on only for a few selected textures.

So here now with Burst conversion, it got more than 50 times faster..!

I think it mainly shows that C# is a very slow language when it comes to manipulating a lot of data, especially with stuff that should use pointers in C++. All the memory checking to prevent nasty bug is very nice to work with on an everyday basis, but when it comes to code some stuff requiring real optimization, C# is really a performance killer.

Although, some of that speed gain is from splitting the processing over all the CPU cores instead of using only one.

Anyway, now the outfit mipmaps are created using the bicubic algo and they look a bit sharper. Some look surprisingly sharper, so I'll wait & see for your feedback in case you think it's just too much...

Note : it's not in the current version, it'll be there with the next update.

Planning

So I did almost everything from the last planning (on August 15th) and some more, but I still didn't do :
- the handling of the dynamic animation system in the replay after a skipped point
- integrating the updating system so updating the game will be faster & easier

And here are the other things I'll do soon :
- check a bit more Federer vs Nadal on grass in case there are some obvious little fixes to do (more in-depth stuff will come later on)
- integrate the dynamic animation system in TEM2, so I could update the game and make it enjoy the changes from TE4 in physics & gameplay
- finish translating the TEM2 documentation to french (it should be just a couple of hours, but I do it little piece by little piece when I still have the energy to work but not too much time available, which actually almost never happens )
- do a little basic system for the commentaries so Darren will be able to test his comments in-game, get a feeling of the final result and possibly tune a few things if needed

Then I'll finally switch to the online mode, while likely working a bit on the gameplay tuning now and then, with the goal to have the online mode ready for the end of this year...


That's all for today, cu next time !

[manutoo]

The return of the son's vengeance

*** skip to the next section if you just want development news ; this is the Calimero section ***

Calimero.jpg

So so so, more than 1 month since the last post in this mini-blog... I'd want to say that time flies, but actually it feels like many months since the last post for me, as I always have a ton of things to do...

Let's have a look at my agenda for these past 2 months, with on the right, for comparison, the one from September last year :

2020-11 - Agenda.jpg

To be fair, now I try to also mark down when I go to the supermarket so I get a better idea of how much time I'm losing with all these appointments, errands, and this kind of stuff...
Each blue line makes me lose on average more than 1 hour, so as you can see my development work time has been seriously hit in these pasts 2 months. I have less stuff planned for the upcoming months though, partially thanks to the new lockdown in place since a few days in France...

And so on top of that it still doesn't get much better with my sons : the big one is still a rebel without a cause nor a pause, although he calmed down a bit, mostly because he can play for a longer period of time by himself. During the holidays, it's still far from enough, so he took a lot of my time, and tired me like hell, especially as his grandmother couldn't take care of him too much, coz he starts to make her too tired too. (so I direly hope they won't close the schools during the lockdown)

The smaller one isn't that small anymore and he really very wants to discover the world, which is good for his development but not for mine (third league pun intended ), so I try to take care of him a bit more than before to make him walk around & stuff (as his mum doesn't like to do that much ).

And to top all that, my work SDD died last week. It was a Crucial 2TB. I had bought 3 of them 1 year and half ago (so I could do a ton of backup in silence ). The 1st one died in last April. This is the 1st time I get such a high rate failure, and now I expect the 3rd one to die within a few months (likely after the 2-year warranty will have ended, though ). But that last one is only a backup of the 1st one, so its loss will not impact me like the 2nd one loss.

For that one, fortunately, I backup most of my work in several different places, so I could restore most of it very quickly. After I had to reinstall several softwares, with some of them being really not happy to not find the files that were supposed to be still installed, so I had to sweat to find a way to reinstall them : I ended installing some of them on my 2nd PC and then copying over the files to my main PC. The software engineers who have designed install software that can fail as soon as the installed files are lost should be tarred and feathered..!

And finally, I found out I didn't really back up everything correctly, so I had to dig into my old work PCs & drives to restore a few special files & setup...

So I lost likely a bit less than 2 days on all that.

And all in all, with all the sh*t coming at me, I'm getting f*cking exhausted. Also probably because I took like 1 or 2 half-days of rest since I got sick on last January, which is clearly not enough...

So I rested another 2 half-days... I can't afford to rest a full day as long as I have kids at home...

"Hello, I'm Darren Kilfara ! Next !"

So anyway, since the last post, I did everything planned (except for the updating system), but it took me more time than hoped, like very often.

The main part was of course to put the basic system in place for the commentaries.

It has met an unexpected problem, though (of course !) : despite recording everything in relatively similar conditions, there were some big discrepencies in the sound volume of the different comments. There are more than 10'000 of them, so it'd have been a titanic stack, if not impossible, to level them manually.

Fortunately, after have scratched my head for a couple of days, I found out tools to do loudness normalization. If you're an sound engineer you might be well aware of that, but for me it was a relatively foreign domain. I knew only MP3gain and it was not perfect. Nor are the results here, but it's damn pretty close : now, I will only have to boost a bit the biggest exclamation comments and it should be ok.

So now if you launch the game and go to the Audio menu, you'll see this :

2020-11 - Menu - Commentary Download.jpg

The system is very rudimentary, as I organized the comments only by their winner / error / service context, which means if you try them you'll hear greatly inappropriate things ; eg: "this was a very well touched drop shot" right after a 150-km/h forehand acceleration winner.

But it's enough to get a sense of the final result, and it should be pretty nice as Darren put great care in crafting all his comments.

However, to reach that final result, I likely have 1 big month of hard work to add the many, many cases to handle all the different specific comments, so I'll do that next year, because right now, the priority is on the Online mode.

Network Limbo

But of course, here again, it's sh*tty land.

Unity, the engine I'm using, has been quite poor on the network side for a decade, and now it's probably in its worst state ever.

The last solution they implemented is not supported since more 2 years, and the new one isn't done yet. At 1st, I thought I'd just pick up a 3rd-party solution : they are plenty of them, and one seems to do pretty much what I need (something very low level but still handling correctly all the basics).

But there was a catch : the network library shipped in the game has to be certified by Sony & Microsoft to allow the release of the game on their console..!

It seems I could get that certification by myself, but I have 0 idea of how long it'd take ; maybe it's 2 weeks of work, maybe it's many months. So I'm stuck.

The only reasonable solution in sight thus is to use the previous Unity network library, as it's still accepted by Sony & MS. However, it may change in the near future if the new library is finally done by Unity. And there's a problem with this one : it's pushing a server-client architecture and completely dropped the P2P support of their previous solution, which is the architecture I'm using for TE. It wouldn't be a total show-stopper as it might be possible to stick the TE protocol over the server-client structure without performance cost (if their stuff is well designed, which is not a taken), but that would still put uncertainity on the whole thing on one side, and on the other one, it'd be the doom of the online doubles with 4 separated PCs/consoles ; not that I intended to do it, but still, I'd like that possibility to be still open if I ever want to do it one day...

Anyway, that's where I stand today.

I'm going to try to get a bit more info to see where I'm going to head, but there's 1 sure thing right now : I have to design my code so it'll support as easily as possible a change of network library...

And right now, I'm finally going to try to integrate that patch system so it'll be easier & faster to update the game, which is indispensable for the online mode to survive more than 2 days of gameplay update...

Cu next time for more headache news..!

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=271790#p271790

[manutoo]

Development Hell

Pfffff, this has been another difficult month...

So I thought I'd need a couple of big days to integrate the new patching system, or 1 week at worst.

How wrong I was ! It took me 1 month of freaking headaches, with 1 issue leading to another till my head felt ready to explode...

It also took more time than hoped because I did the things correctly : I prepared everything in a good way so the demo bundling is already almost done ; I also put in place bundle handling for the stadiums : this means that each stadium lies into its own separate file, which makes releasing a new version a bit faster as Unity doesn't have to repack everything every time. This will also be useful if a Modder wants to create a 3D stadium one day.

For the demo, I wish I could have kept it under 300MB, but it'll be closer to 550MB. I hope it won't be a deterrent for new users to try the game. It's 11 times bigger than TE2013's demo, which is also the full game...
And to achieve that, I'll have to do a trick : for the demo, I'll repack the player textures in .jpg instead of using Unity format. That format is faster to load, but it's much bigger than .jpg files. Quality-wise, the loss shouldn't be too visible, especially as the textures are relatively big now, even for the female as I did the switch when preparing all that.

On the little things, I also checked 2 tools to decimate polygons : they allow to remove polygons from the player models to create the low-detail models ; unfortunately, both have some serious shortcomings, so I guess I'll have to ask my artist to do it by hand, so it's going to take a long time and a lot of back & forth to get everything in a decent way...

For the Mac version, I took this occasion to notarize TE4 as well as TEM2 : like this, the Mac users won't have to bypass a security warning and manually authorize the game installation. Hopefully, it'll help to raise a bit the managames.com sale of TEM2 on Mac, as they are quite asthmatic lately...

So for the good news here the new updater/launcher :

2020-12 - TE4 - Launcher.jpg

It works decently well, although I had to do a lot of little things to make it work as I want : I added the current local version & the new version on the download bar ; I enhanced the logging so you'll have a better idea at what's going on, and other little stuff like that.

It's been on the website for 10 days now for the Windows version, and since yesterday for the Mac version, and no one reported any issue, so I guess it's globally working well enough...

So now, I won't have to worry to destroy your Internet boxes when doing a new update : a small change of gameplay leads to an update of less than 1MB, which is much much faster to download than the full 1.5GB installer (which will grow to 3GB once everything will be there !), especially as it's automatic ! Launch the patcher and it takes care of everything.

I also tried to get TE4 into the next Steam Game Festival (in February) to build a bit of awareness around the game, so I spent 1 day preparing the TE4 Steam store page only to find out it had to be validated by Valve before I could enter TE4 into the Festival (*), which takes around 1 week, while the deadline was the next day... Bummer...
(*) : the explanation said it wasn't required, but the validation button label stated the opposite. I reported that glitch, but they didn't fix it. Valve's style.

Anyway, it gave me the occasion to do a new banner for TE4, I hope you like it :




Online, here we come !

So I also contacted Sony about what network library to use ; their answer has been cheerleading : basically, it just has to work. So there's nothing especially tricky or hard there. Also, for the certification, it's just about implementing features in a standard way so it's not really problematic. I guess it'll have a lot of headaches on the way, but theoretically, I can take any network library and if I have access to the code, I shall be able to make it work for PS4 if it works on Windows (assuming the PS4 network sockets work like on PC, which I wasn't able to confirm in my 15 minutes navigating through the PS4 documentation).

I really wanted to get the online playable (if not 100% done) for January 1st 2021, but it seems unlikely to happen now, except if everything goes incredibly smoothly (and if you read that mini-blog often, you should know now they never do ), especially as I think I'll take a couple of weeks of holidays for Xmas as I'm always f*cking exhausted, plus right now I'm sick (bad sore throat & light cough ; no, it's unlikely to be the Covid19 because I don't have fever ), and my personal life is still quite busy...

Anyway, let's all cross fingers very hard and maybe we'll have a little something to play within a couple of weeks, who knows..!

That's all for today ! Have a nice X-mas and/or end of year celebrations & holidays.

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=273402#p273402

[manutoo]

Sick Month

Pfff, I've been sick the whole d*mn month of December..!



I chained rhinitis, pharyngitis, bronchitis, back to pharyngitis, and then a persistent light throat-ache for the last 10 days...

It was unlikely to be the Covid19, though...

In bonus, my 2nd son is now 15-month old and he's screaming non-stop when no one takes care of him. So I can hear him when my wife's cooking and stuff...
He has to be out of his park to roam free to be alright, and thus it can be quite perilous for him as our home is not 100% baby-proof.

Networking in progress

So I was quite tired the whole month and couldn't work much. At 1st, I thought I'd work hard in the 1st half of December and then rest and enjoy some family time during the holidays, but as I started to port the network code when I got a bit better, I had to finish ASAP to not forget parts here & there, so finally I worked a bit during the holidays to wrap 80% of the netcode port. Now, there's only the prediction engine left to port.

This is actually a double-port : from C++ to C#, and from the DirectPlay network library (from DirectX) to LiteLibNet which seems very neat for my needs.

I still didn't test anything, though. Right now, I'm upgrading the menus so I could launch the connection between the host & the client ; here the main network menu is done :

2021-01 - Network Menu is Back !.jpg

Note : this is the TEM2 menu theme because I work with TEM2 now, as it compiles twice faster than TE4, because it uses Unity 5.6 which is much faster than Unity 2019.4 used with TE4.

So now, if lucky (2% chance to happen ), there won't be too many issues & mistakes and the online will work soon (ie: end of this week), but TE2009 style : without prediction, and thus with perceptible lag.
The other scenario is that there's a sh*tload of ugly things to fix and it'll burn my brain once more.

Anyway, once it'll work like this, I'll release a new version ; it'll be easier to see core issues without the prediction engine on top of everything, so if anyone is brave enough to give it a try, it could be useful to find some bugs.

After that, it should take 2 weeks to port the prediction stuff, and then add the push-through server, so I'm relatively confident to see the online mode done by the end of January if there's no issue arising in my personal life.


Business Consideration

With the end of the year, I did a bit of accounting for my company and I checked how much each game contributes to my turnover.

So a little warning 1st : the income from Steam is from November 2019 to November 2020 because I get paid with a 1-month delay, while the income from managames.com is only on the 2020 year (but it's significantly lower than Steam sales), so it mixes slightly different periods and thus it's not very accurate but it gives a good idea of the big picture.

So here's the rankings :
- TEM2 = 43%
- TE2013 = 30%
- TE4 = 13%
- Dungeon Guardians = 13%
- TEM1 = 1%

Actually, TEM1 is 2% but I consider 1% goes to TEM2 instead because most people get TEM1 in the Steam TE Bundle to get the discount on TEM2...

So now TEM2 + TE4 were approximately 56% of my income in 2020. This means without these 2 games, my situation moneywise would be really problematic.

TE2013 still sells decently but the sales from managmes.com are now very low and most income on Steam is done during special sales, so despite quite low sales, TE4 brings almost half of the TE2013 income. So that's a good occasion to say thanks to everybody who pre-order the game and help me to not worry about money... (although the game is very playable & played as analytics show ; I'll have to write about this in the next post)

Dungeon Guardians still sells a bit as well ; it's not very impressive, but as it requires nearly 0 work, it's pretty nice...

This is all for today, see you next time for a nearly finished online mode !

Meanwhile, my best wishes to all of you for 2021.

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=273704#p273704

[manutoo]

Online (almost) fully up & running !

After having released the online mode without the Prediction engine 2 weeks ago, I got a bunch of Bug Reports so I spent a while fixing many little & bigger bugs, till about 1 week ago when things finally settled down.

Fortunately, I got a good idea to speed up the port of the Prediction engine and thus I could finish it yesterday, despite being taken by many other things.

I think it's working nicely now, but I already got a couple of reports about having a bit too many desyncs so I'll have to look into these and try to fix them. On my PC, I don't get any desync, but my setup is pretty simple and thus it doesn't cover all the possible cases when playing online.

Anyway, now, we just need the Push-Through server and a new way to choose the Talents, and the online will be all good ! (more or less )

Hopefully, it'll be done by the end of the week, but as it's the holidays for my son and my mum won't take care of him every day, I might spend some time with him...


(small) Victory !

Ok, I already hinted to that several times, so finally here the analytics of TE4, showing the game is already well-received, despite its Alpha state (I blanked out the user scale numbers) :

2021-02 - TE4 - Analytics.jpg

It's not seen on this graph, but the "monthly users / total purchases" ratio is nearly 30%, which is pretty good. The Sticky Factor, which is an indicator of how much the users are engaged with the game is 0.20, which corresponds to a relatively successful finished & released game.

Now, let's compare these numbers with TEM2 :

2021-02 - TEM2 - Analytics.jpg

TEM2 is meant to be less played than TE4 (ie: 1 career in TEM2 lasts 20~30 hours, while you won't even finish your 1st season in TE4 in that time), but it gives some ideas to situate the relative success of TE4 so far.

So the "monthly users / total purchases" ratio was at best at ~30% and now it fell to ~15% and the Sticky Factor is around 0.13 ; these are still reasonable numbers, but they are significantly less good than TE4 ones, so that's why I allowed myself to think that TE4 is already fun and played regularly by a bunch of people, although I won't allow myself to rest (too much) before the game will be complete..!

Note : these stats don't cover people who play without an Internet connection or who have opted-out Analyctis ; if you want to do so, go to Settings -> Misc Options and click on "Unity Analytics" ; although it's anonymous and I have access only to aggregated stats and nothing personal.


Near Future

So a little word about what's next.

I have these things in mind, although I'm not sure in what order they'll come :

• Demo : so it'd be easier to find people to play online (and possibly switch the dying TE2013 Online Tour to TE4)
• Anim modding support : as it has been one of the most popular requests for a very long time, and there's no very important feature missing from TE4 now (ie: feature-wise, it's already better than TE2013, except for the Modding support)
• Integrating the new NPCs : the new ball boys & the new Chair Umpire are ready to be put in the game (at least I hope so !)
• I'd like to do a few small & bigger quality of life improvements (eg: automatic video quality selection on 1st game launch, so people with IGP won't have too much hassle to setup the game at 1st)
• Adding a couple of new stadiums (although it's not very important and I might end working on the gameplay before getting to that)

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=274719#p274719

[manutoo]

Trouble is a friend of mine

Once again, problems stacked on my front door.

So it all started with my 1st son getting sick (like often since he goes to school). Then his little brother got sick. Then his mum. Then me.



Little variation this time : I got only mildly sick and had to slow down work only a couple of days. It was on a weekend so for once it felt I had a kind of normal life.

But, big but, my wife started to cough blood a few days afterward.

Spoiler alert : coughing blood is bad, and it can get pretty ugly fast ! Think "does drowning in my own blood sound fun ?" ; if not, then phone your doctor. So we did that, and my wife ended to do an emergency scanner. Fortunately, it was bad but not too bad, so she ended with strong antibiotics for 10 days and had to sleep & rest more.

This means that for a short week, I took care of our baby for a big part of my days (in addition to my 1st son) and thus barely worked. (it was nice nonetheless to spend time with him, he's an extraordinary little dude )

My wife is better now, but she still needs a bit of time to fully recover.

And in bonus on the last few days my baby got sick again and it gave us some worries and another trip to the doctor...

He's ok now and hopefully stays that way.

Dead end

After that, I had some problems getting back to work, especially as I had to fix many little bugs here & there, deal with issues in the online mode, and so on.

Eventually, I was able to work at full speed end of last week. I focused on a polygon decimation tool. I'd like to have a version of the 3D players with a lower polycount so the game will run fast enough on low-end GPUs & decent IGPs (like in TE2013).

At 1st, I had decided to ask my artist to do that LOD version, but as he's nearly 1 year late on his schedule, and that he always sends me stuff full of issues that I need a long time to find & detail to him, I'd really prefer to not depend on him for this.

Last November, I bought 2 polygon decimators but their results were not completely satisfying.

So I spent nearly 1 week creating mine. It was kinda interesting, especially as it turned out it wasn't that hard to do. The ones in 3D Max are kinda sucky so I thought it'd be complicated to create only a half-decent one. But it takes time to be well polished, and I'd need a least a couple more weeks to get to a decent result, and I can't afford to lose so much time on this.

So I went back using one of the decimators I bought, especially as it has a very good topology option (the polygons are well spread & connected along with the 3D shape). I tuned its code, but the face is still a bit too much destroyed and an artist would likely get a better result, but I'm going to make a try like this and I'll see from your feedback if it's ok enough or not. If not, I guess I'll have to take the dreadful path of making the artist do it and get one thousand back & forth till it'll be alright.

Next ?

So I'm going to create the LOD versions for the male & female players. I hope they'll be done by end of next week. Then I'll try to incorporate the new ballboy & chair umpire.

Once done, I'll likely see for the animation modding. It feels far now and my mood is kinda low. I'm really tired of all of that constant sh*t coming my way. I hope it'll get better soon. I had 1 month & a half of good productive time and really would have appreciated being able to keep that rhythm at least till June to release the Early Version of TE4 on Steam in good conditions...

Cu next time to see what the Gods will have thrown on my path, but don't worry, they only can slow me down, because in the end, I shall prevail.

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=274879#p274879

[manutoo]

Slow Motion

Pffffffffff, I got even more annoying things happening to me than usual in this past month..!

It's getting so ridiculous, that I won't write everything here, but this includes :
- my neighbor threatening to break my legs because I told him to make his dog stop barking, so I had to go to the police station to report the incident (his dog stopped to bark though, so hopefully it'll be ok now )
- thanks to the TEM2 sales, I could pay to extend my house and stop to sleep in my office ; of course, the construction got delays and whatnot (many whatnot ) and we have been staying in our house without insulation for weeks while getting one of the coldest months of April I can remember of !
- electricity got cut 6 hours and a half yesterday (for "grid improvement") ; no insulation in house => got quite cold
- lockdown in France, including closing of school this time, so got my son at home 1 week earlier (and now it's the holidays)
- having workers in the house is actually quite distracting as they regularly need to ask questions (plus it's noisy of course ; it feels like I'm back to Thailand where my building seemed to be almost constantly in construction )
- and a good bunch of other sh*t on the personal side (but at least this time no one got too badly sick )

So once again, just getting to work has proven to be a daily challenge, especially as all that got me quite tired and I need to rest.


Animation Modding

Anyway, I managed to finish the animation modding support for TE4..!

2021-04 - TE4 - Anim Modding SDK.jpg

This also includes most of the things needed to add the modding support for the outfits (which thus should come soon).

You'll find more info about how to add your anim in the game here : topic29-33190.php .

But except bug fixing, that's pretty all I did in that past month, so that doesn't feel like a lot...

I tried to create the LOD version by myself, but I quickly ran into a big problem : when reducing the polygon-count of the apparels, they were not matching the player body anymore..! So it looked quite bad and it wasn't usable. So I asked my artist to do the LODs by himself. Hopefully, he'll deliver them soon enough.

I also tried to incorporate the new ballboy & chair umpire, but both had issues, so the artist had to fix them 1st. Now, they seem to be ok, so I should be able to put them in-game next week.

There's still a decent hope to get TE4 ready for its Steam Early Access release in June, so let's keep our fingers crossed, and see you next time to keep a list of all the horrible things happening to me !

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=275526#p275526

[manutoo]

Early Version

Ouch, already more than 2 months since the last update..!

But I have a good excuse : I both set TEM2 out of Early Access on Steam, and released TE4 on Early Access on ManaGames.com & Steam. In addition, I also released the demo/trial version : https://www.managames.com/tennis .

So if you follow a bit the activity of the site, you might have noticed several topics about all that, including the Roadmap for TE4 here : topic15-33253.php .

To reach the Early Version state for TE4, I got the artist to create the LOD versions for the male & female players, and fix almost all the issues of the NPCS. But not all : at the end I was fed up, so I took the matter in my hands and fixed the crowd by myself.

The artist wasn't able to correctly map & optimize a low poly-asset. With minimal effort, I removed more than 15% of the polygons (a good bunch was just totally hidden by other body parts ), as well as removing little holes (that were creating jittering in the distance), and in bonus I enhanced the rigging. I guess I should offer my services as a low-poly artist because I can outpass some 3D artists.

Coz of all that, I wasn't able to fix the weird animations before the release, nor include the new controller configuration system.

Since the yesterday's version, at least 80% of the weirdest animations should have been fixed, so now I hope to get the new controller setup done within a couple of weeks : it actually requires an awful amount of stuff to be done, but it's necessary for the console ports, and to ease newcomers into the game (else they get a headache setting up their controller).

I already spent a couple of days on that in May, but as I couldn't finish in time for the Early Version release, I reverted my change in the code. It's the 1st time ever I did that, and I hope it won't be too much of a hassle to get back the WIP from my source control base and merge it with the subsequent changes done afterward...

I'm also eager to put the customizable AI support in the player base (from TEM2 real-time coaching), and allow the Modding of the outfit so I could release the TE4 Modding SDK.

TE4_ss06.jpg

What's going with my life

I'm just going to rant about my personal life here, so you can ignore this part if you were there only for dev news...

So my house extension is still not done. It should have been finished a couple of weeks ago, but due to several issues, including worldwide shortages & price inflation on many construction parts, as well as closed factories due to Covid, it got severely delayed.

In bonus, some workers did a bad job, leading to extra work for themselves and other workers... So all in all, it might need another month, and it's really annoying to work with noise & extra activity in the house.

On another side, my baby is getting bigger, nearly a boy now, and he's really a handful so he drives my wife crazy and I have to jump in often and take care of him regularly. Luckily, my mum takes care of my 1st son often, so I still have a bit of time to work...


That's all foks for this time,
Cu soon enough for more dev soons about TE4,
Meanwhile, may you have a nice summer !

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=275526#p275526

[manutoo]

New Controller Setup 70% done !

Ouch, another month flew by...

I'm still stuck with the work in my house on one side, and on another side, with the summer holidays, my big boy is at home, needing help & attention, while the little one is growing up every day and I don't really want to miss much of that either...

I still manage to work, but the pace is awfully slow.

Anyway, the 1st main step of the new controller setup is done : now you assign the buttons (or combination of buttons !) to each strike directly, which is much simpler to apprehend than the old system :

2021-07 - TE4 - New Controller Setup.jpg

I still have to convert "Button 1" to a green-colored A for the XBox controller, and so on for the different buttons, as well as for the PS4/5 controllers, but hopefully it'll take significantly less time to reach that final state.

It felt a bit strange to change all the basics of the control scheme, even though the final default settings look the same for the keyboard (but not the controller) : I had put all that in place in 1994 when I wrote the TE prototype over a few weeks during the summer before entering university...

At that time, my goal was to make it possible to play with the Gravis Gamepad :

640px-Gravis_pc_gamepad.jpg

As you can see in the picture above, it had 4 buttons, not 2 like in the legacy TE setup, but there's a trick : only 2 buttons per gamepad were working when plugging 2 of these gamepads into the same PC, and I wanted to support local gamepad-based multiplier, thus the 2-button limit the TE series had for years, till ~2010 when I added the support for the strike remapping.

The current goals are the same as in the previous post of the mini-blog. Hopefully, my working pace will pick up a bit : the work in my house might not be done before September, but there's very little bit left to do, so now it's mostly a matter of waiting a few parts to arrive, and the last few cleaning & wrapping stuff...

So see (soon ?) for the next update !

A comment, cheer or question about this mini-blog ? => viewtopic.php?p=275526#p275526