HeartOfGermany

Unfortunately the cheapes one is 500€ without lenses. I now bought the Sony NEX 5N. I hope it will work great. It was VERY cheap. I payd 160€ with a nice 15-55mm F3.5/5.6 lense. It has both a smooth Focus and a smooth Zoom wheel. That is already very very great. It is not recording raw, it cannot shoot very high bitrates. But .h264 at 24mbit/s is quite okay. Yesterday I did some comparision renders from a real 8k down to 1080p (so I can be sure, that the source image is *at least* 4 times more detailed). I rendered it with: 2mbit/s - Like a bad 1080p youtube video 5mbit/s - better than almost any 1080p youtube video 10mbit/s - almost crisp image - like 1440p youtube video 24mbit/s - very crisp image. Like a 4k render, "lossless" (bitrate wise) rendered down to 1080p Note, that: - I compared with 200% zoom, to really see the differences (10mbit/s for Youtube is more than enough in most cases) - The video was with slow movement only. Of course it looks a bit cleaner than fast moving images - The camera might encode in a different way But: If the image comes close to my 24mbit render, I am absolutely sattisfied. It will likely look still great, when moving the camera very fast. I should be worried more about ISO>Noise etc., lens distortion (yes, the lens it includes is supposed to have some distortion - but honestly, I did not notice it at all in the videos. This 1% difference is waaaay out of my budget. Also, thanks to the EU I have a 1 year warranty and can return it. Yes, those 160€ where from commercial seller. So if the camera sucks or has defects I can return it. So that should be okay. Other things to note: - My Lense has F3.5-5.6 (Maybe upgrade to a low Aperture lense for some "workloads". I saw some for only 50€ with warranty - High apeprture lense maybe worth a thought very late (like next year or later) - Let camera arrive than there is more to say. Thanks for your time!

Well, the camera I mentioned at the end seems to be promising. However the Video quality on youtube sucks, so I cannot actually tell, if it realy is good or not. I am realy thinking of buying this one. Used for 160€. And I was hoping, by pumping the price a bit, there might be 1 or 2 cams that are a bit better and only cost a bit more. The Sony NEX 5N is quire promising. If no one knows a better choice, this will be the one. Maybe I can dial down contrast and saturation a bit, to get more dynamic range. That is, what a youtuber recommendet. But he dialed it down by 3, which is a lot - who knows, if stuff still looks good than... If PCs where 100 times faster, I would just use this cam and throw the videos trough an AI for enhancing dynamics and removing noise. (Yes, there are some solutions for this out already, but I would call them "Alpha development" (Gigapixel AI and JPEGtoRAW both from Topaz). Thos 2 programs infact can enhance a ton of images - though just a few look horrible after that. Main problem is time: Resizing something by 2 to 1080p on a 1060 takes between 15-30 sec. per picture. And JPEGtoRAW takes 2x to 3x longer per foto. Now imagine, I start with 1080p content. This will take absolutely forever. So anything better than Sony NEX 5N for less then... say 300€?

(OT)Before writing the title, I have to say one thing: I was thinking of the topic above and was searching "fotographic forum". Than I thought, what would be best for tech? Propably LTT forum. I love this place. This can really help dudes out there, because the membership is just huge. Love it. Thank you linus, even just for that, you earned a big place in my heart, thought you'll never read that. Thanks dude! BACK on topic: I'll likely make it short, if you think about my requirements: I would love a camera below 200€/$. Required features: - A good dynamic range in video mode (no over and/or under exposure when filming in a garage, sun coming in on a hot and bright day, someone talking in the shadow, somewhere in the garage. My camera points not directly into the sun (hidden by roof). This is a very difficult scene for almost every camera <1000 bucks. I want to clearly see the details of his clothes (like a saturated, lightly dark blue) and yet see the or at least quite some details on the bright grey (freshly cleaned, smooth) concrete. If you understand the difficulty of this requirement, you are the man to possibly help me - 1080p60 is absolutely required. I want to shoot some nice stock footage to earn some extra money, when I go to school this year for IT education. If I want some 1/2 speed slow mition, it should look good. - RAW would be nice. This would help to (1) correct some slightly wrong exposed video and (2) get even greater dynamic range. *not required* - Big sensore would be great (because I want clean image if you think of the next requirement which is:) - Possibility to change lenses for some very high aperture lenses later - if not RAW at least a high bitrate recording like .264 compression >50mbit/s or .265 >20mbit/s. It of course it very likely will not be in my budget, to get a .265 encoding camera. - included lense has a wheel to focus and zoom (non of this button or software rubbish. would dave jones from the EEVblog say it like that? - Many cameras cannot catch the violet/purple tones when sun goes up or down. This absolutely is required. Without that many videos will look like garbage. Just imagine seeing a purple flower, which looks totally blue on camera! AAND: - CAN be a smidge more expensive, if it totally is worth it - you can consider me buying *used*, if you can recommend it (no recommendation for a 2 year old camera, that only lasts 2 years. ) What I thought of: Sony NEX 5N But at 10 years on the market (i think) it sounds like it will have to many drawbacks. I looked at some footage on video with low light and some dynamic scenes. I mean, I love the colors and the accuracy with a custom profile in a nature scene. And at least through the horrible compression of the video (Youtube as always...) it seems, that it is not that bad in noise. But I think in video mode, especially with the 24mbit/s old AVC/MP4 to keep in mind, it will lack to much dynamics. Complex scenes will likely loose detail in the bright and dark spots. But realy, sometimes the colors are absolutely astonishing! I saw a red purple sky when sun goes down. On mobile I never can catch thos violet/purple tones at all. This camera does that. It looks so good for its age and price! It could almost (!) be the one. Okay, this was kinda long as always Can you possibly recomment something with this specific requirements? Thank you. Edit: I added some lines for social distancing as required on april first. Nice one!

You are kidding, right? A reasonable GPU will no problem run 24/7 at 70 for 2 years or more. About the proof, that it is supported: First of all: What are you waiting for? Why do you not update your Nvidia driver? This can in many circumstances lower some games performance by 50%. (Yeah, some older games over years doubled performance with updates) Also support could improofe. And with a KVM virtual machine the GPU will likely not run without modding KVM. And as far as I can tell, KVM is the only good solution for virtualization on near 100% performance. Nontheless: Please update your driver to the latest version and report, if this fixed your issue. Edit: Overread the part with the missing gl file. This gl looks a bit like Open GL. And did you mean "libgl.so.1" ? If so, please remember, that the precise name can help the helpers in finding a solution. Also other peoples might want to find this thread later and that likely will not happen, if the names do not match. https://ubuntuforums.org/showthread.php?t=2233005 This could be your solution. It sounds like that. Please report back.

Virtualbox simply is not suitable for accelerated workloads. Try to get it running through KVM. It is propably much harder to do, but it realy passes through the GPU. If you cannot archive it, you should go to just a windows install as 2nd boot. KVM with PCIe passthrough is very difficult and very specific for each combination of hardware and software. But it is a very very powerful tool. I know of absolutely nothing even close to that! Of course you have to find a way, to hide the VM for the drivers. The drivers will not (REALY NOT) work at all, if you skip this step. It does not even matter what version of driver inside of windows you are using. As far as I know every version is locked down in that way. Sadly noone else is posting anything about it. Maybe people are affraid of law/NVIDIA? Whatever. I am also confused of your driver version. Was the Linux driver outdated or the Windows driver? Whatever it is, try to update your drivers in Linux. Why are you talking about P106 anyway? Did you buy a mining card for that (the cards with no video out)? This will likely not work well. As far as I can tell, on windows you actually needed a modified driver for it to install as a GTX series video card. This gets even more difficult when inside of linux. If you cannot find a tutorial about that, you are likely lost. Here is one very important lesson: If you save 10-20% but have less features and no support, do not even bother. It is a waste of time and time just runs away. I make this mistake till today. But it still is a mistake.

Well, the drivers inside of Linux are not VM blocked. But to use the GPU for a VM you actually need to run the Linux host on another GPU (even a cheap one). To pass through the GPU to the VM, the GPU cannot be taken by the Linux OS. Try to update the drivers. If this does not work, you can try the open source drivers. (last resort, since I don't think the acceleration will be as optimized as the original drivers) Unfortunately there are tons of reasons, why it might not work. Even the version of the same game can matter a lot. Sometimes even the old version runs better. It is just not really predictable. I mean, GPU stuff is really complex. And also the Windows<>Linux abstraction has thousands of problems. If you cannot fix it, you could try to use Win10 as 2nd OS (dual boot). The VM will work, but it might be far to hard to actually get it working. If it works, it just is amazing. It is basically like just using Windows on bare metal. VMs have come a long way! Do not upgrade your GPU just for this reason. Rather wait, till you need a faster one and than choose AMD, which is preferred by many Linux users, since AMD is not such a big d!ck like Nvidia (and also Intel unfortunately). Please report, if Wine only works for you. Wine up to date also? Good luck.

Photon as you are talking is the Steam API for running unsupported games on Linux, do I get that right? The problem might be simple and very unsattisfactory: (my information is not the latest, so don't take it for granted!) Photon is in my Opinion still in the Alpha/Beta state. It is not stable. It will not work with every game and every hardware combination. That might be the main problem. I know, it is very complicated, but if you realy want to game on linux, a look into KVM with iommu (GPU passthrough mode) might be the solution. And yes, you obviously are required to run Windows for that. You also need to hide the VM from the NVIDIA drivers (I never found a soultion for my problem and gave up prematurely). And this, you have to absoultely understand: This is not allowed by NVIDIA. So not only do it at your own risk, you also might break the law by doing so. If you see no problem with that you can try it out. If you get it working, you can expect performance up to maybe 97-98% of the bare metal performance. Yes, this is f****** impressive. It realy is amazing, if you get it to work. Tip: Maybe AMD is a consideration for the next buy - since AMD is not locking their drivers in VM and I think they do not forbid it. One thing to say: If I would have 100 million dollars, I would absolutely spend 2 of them for the development of a REALY good compatibility layer that supports DX12 gaming even. But in this world... Unlikely to happen by anyone soon. And when that happened, Microsoft will have a new DX or just update DX Ver.12, so that latest games are also not really working in linux. What I cannot read out of all of this: What kind of driver did you install for your GPU and did you try an alternative? This can have a lot of effects. Also did you update them, if even an update is available? Also it migth sound stupid, but did you try just using plain Wine? Wine should work with Vulkan! So might be worth a try if nothing else works.

Okay clarification: I do not like the physics engine. I want to basically build my own from scratch. This is the main goal. (For both learning and improvement reasons) Also, the built in physics engine is not accurate in my usecase. Also, I want some special stuff to happen on impact (sounds much simpler) Both objects have a Collider and a RigidBody2D. Both colliders are triggers. And yes, that is absolutely intentional. If no purely trigger relatet way is possible, I will try the following: multiple raycasts (360° [in ~5° to 15° steps] from the start, mid and end of the bullet), to get the closest point. But I would love, if there is a simpler, more integrated way to detect trigger collision position (just use an already existing variable). I do not understand, why this is not possible in an obvious way. I read multiple threads where the solution was to invert the "isTrigger=true". It is not intended, to use unity physics for that. This is no solution for me.

Why I need that precision? Well, it looks very dumb, if I want to spawn animations at that point. Also, I want some physics to be applied in a precise way. That absolutely requires the correct position. I want to let the asteroid spin and/or move away from the impact (0 Gravity environment). Depending on the hit location and the resulting angle AngleA between line(asteroid.midpoint, hitpoint) and line(bulletspawnpoint, hitpoint) the asteroid should start to spin in some way. The bigger the angle AngleA is, the more of the kinetic energy should get transformed into rotation depending on the mass and distanance(hitpoint, asteroid.midpoint). If the Angle is 180°, the energy should only transform into linear motion. Also the impact is supposed to create preasure and depending on the shape of the entity the energy reflects in different ways - which either results in spin, linear motion or a combination of both. Why be so crazy about it? - It is not a REAL game. I am trying to work on my Unity and C sharp skills. Precisely understanding the relations between all this is important for later projects, that actually should work properly. Long story short (as always): So is there a way, to get a collision position, that actually is correct without using any rayCast method? With rayCast it will be much, much worse in precision, which I'd love to avoid. And using a collider with "isTrigger=false;" will not work, since it would react to other non-trigger colliders, which is also not wanted. Why? The problem is, realism. Standard collider with rigidbody2D will not correctly react to torsion forces. Just imagine a ball hitting another ball on the side with high speed. Some of the friction will induce a spin on the other ball. This is not working with basic unity physics. P.S.: yeah, I know I have to code every interaction with other objects from scratch by avoiding the otherwhise usable physics of Unity2D. But that is intended. First: Old console games (SNES, GBA...) do it all by scratch and rather use the same aproach I want to use (only mine in a refined way). Second: After getting all the physics working properly I think physics training in Unity is complete to me. Nothing to learn in that direction. (Only thousands of other topics to master left. )

Well, it does not install any audio device - which on a GTX 1060 for example will be installed, even if unused. So we now know: - DVI CAN support Audio out - HDMI DOES - Some iGPUs output, some don't (Likely more a Bios/Motherboard issue!***) *** Why? Well, I heard even of some Motherboards with the ability, to do realtine Dolby encoding, but it being disabled in the Bios firmware. Some tweakers actually modified the binaries of said board and got an output. So it seems, that there is a multiple and absolutely unpredictable chaos out there when it comes to Onboard Video+Sound. My solution was to buy a passively cooled GPU with HDMI for 2.04€ from work. (~2$) It is capable of 1080p60 and gives sound AND is also still supported by windows 10 - no idea why in 2020, but it works and that is nice. Btw: 2m Cable length. Also, if I remember correctly, the Audio is transmitted over the same data lines as the video. So an unstable connection will likely not just disable audio and suddenly work for video only. (Maybe a 1 in 1M) Got it to also work on 10m cable. thanks for reply!

You just wrote VM? So are you actually running regular Geforce drivers? In a Virtual Machine actually Code 43 should happen, since Nvidia does not allow drivers of consumer cards to run in VM (there is a workaround to hide the VM, do not ask me how - never got it working. Amd works out of the box)

Hello to everyone. I am having trouble in finding a way, to detect the position, where 2 Objects collided. What I have: Bullet with Rigidbody2D and CapsulleCollider as trigger Target with Rigidbody2D and PolygonCollider as trigger It seems, this would work nice without using the Colliders as triggers, but I can not allow the objects to react to physics outside of my code. private void OnTriggerEnter2D(Collider2D collisionData) { //Debug.Log(collisionData.collision.position); Debug.Log("Here it is: " + collisionData.gameObject.GetComponent<Collider>().ClosestPointOnBounds(transform.position)); } The problem is: I get the error: "MissingComponenteException: There is no 'Collider' attached to the "Asteroid2" game object, but a script is trying to acces it." Ah thanks Unity. I know I want to acces it. But it is not working. Why not, it has a Collider? What is not a good idea: - Raycast > Since I realy want the exact position respecting the shape of projectile and target (the Colliders I chose in the first place) - set "Is Trigger" to false on the collider of the target (disabling also gives the same error anyway) What I realy need: The hit point or hit points of the collision. How do I archive that? Thanks for help.

Thank you a lot. The answer is perfect now! I gues, now I almost fully understand C# basics. That was very helpful, so realy thank you for your time.

So far so good. This helps actually. It is pointing me in the right direction. Now I also understand something. private int something_integer_alike; This variable is only for THIS gameobject, right? If not, this variable could only be used once, because it would be a public integer, right? So is it possible, to be called outside of the script? That is the reason, why I thought of something like objectname.variablename in the first place. I mean, what, if I have 1000 asteroids. How do I call the health value of the asteroid 534? Actually I am not using visual studio. I used PSPad. It formats it different. I actually used notepad.exe for all of my CoD4 scripting. It worked perfectly. The formatting part however makes it interesting to usue visual studio. It autoformats something like: if (...whatever) { something; } to if (...whatever) { something; } ? If so, I realy love it. I hate seeing if(...) { This bracket belongs on the next line! Sorry for the formatting anyway. Will not continue to use PSPad for anything complex, since the formatting is not my flavour. Thaaaaaaanks a loooot, guys!

Hello folks. I feel kinda stupid for asking. It actually is the simplest thing in the world, but I do not understand some basic logics this time. I watched a Video on YouTube called "How to Start Making Games - Unity Playground (Beginner Friendly)". And I tried to make a demo project like asteriods with "Unity Playgrounds". I got my ship shooting lasers and the asteroid gets destroid, once hit. The problem is, I want to add health, so you need to hit it 3 times. I need a variable like gameObject.health. From my CoD4 scripting experience it was like this: self.health = 0; predefined the child variale health to self which is the gameobject player. Why can I not do this with C sharp (C#), when using private/publich int gameObject.healtPoints = 15;? This does not work, but why? Because destroy gameObject; infact destroys the asteriod. So the parent "gameObject" seems to be correct as parent of the variable gameObject.health. How do I define this variable, to later be used ingame? What am I missing? Please do not tried to explain the basic C# rules - this problem is not just about C#. I need an example, to understand the rules behind that. So what did I do wrong in this case? It is important to me, I actually got quite some plans. But this basic barrier actually stops me from doing anything really useful. I would not have asked here, if answers.unity.com was of any help. zero answers after over a week! Nothing. I am frustrated that even the main page does not help. And videos are often also not helpful. They only show you how to copy code, but do not explain, what the code actually does. I am absolutely stuck here trying to understand this variable/object problem. Thought, this will be really simple, since I actually scriped a lot with CoD4 and was actually quite very good at the end. This is the problem code. All I have added is stuff named "..... healthPoints.....". Anything else was there before and was working perfectly. [AddComponentMenu("Playground/Attributes/Destroy For Points")] public class DestroyForPointsAttribute : MonoBehaviour { public int pointsWorth = 5; private UIScript userInterface; private int this.gameObject.healthPoints = 15; private void Start() { // Find the UI in the scene and store a reference for later use userInterface = GameObject.FindObjectOfType<UIScript>(); } //This will create a dialog window asking for which dialog to add private void Reset() { Utils.Collider2DDialogWindow(this.gameObject, false); } //duplication of the following function to accomodate both trigger and non-trigger Colliders private void OnCollisionEnter2D(Collision2D collisionData) { OnTriggerEnter2D(collisionData.collider); } // This function gets called everytime this object collides with another trigger private void OnTriggerEnter2D(Collider2D collisionData) { // is the other object a Bullet? if(collisionData.gameObject.CompareTag("Bullet")) { if(this.gameObject.healthPoints == null) { this.gameObject.healthPoints = 15; } this.gameObject.healthPoints -= 5; if (this.gameObject.healthPoints < 1) { if(userInterface != null) { // add one point BulletAttribute b = collisionData.gameObject.GetComponent<BulletAttribute>(); if(b != null) { userInterface.AddPoints(b.playerId, pointsWorth); } else { Debug.Log("Use a BulletAttribute on one of the objects involved in the collision if you want one of the players to receive points for destroying the target."); } } else { Debug.Log("There is no UI in the scene, hence points can't be displayed."); } // then destroy this object Destroy(gameObject); // Oh hello guys!!!! This is what I mean. It destroys the asteriod. So shouldnt int gameObject.health just be working? But it is, like gameObject is just locked down to what already is there. // Also I do not understand, how to add a script and get a prompt in the user interface for the prefab, // to predefine the health of an asteroid. Like: // int gameObject.health = createConfig("Asteroid health", "This variable (int) will set the health of the asteroid.", int 1); // -> explained: createConfig("Variable name for interface", "Help text that shows to explain the input", int default_value_if_undefined); // I am asking, because I keep seeing this, but never know, where those custom inputs in the user interface from scripts actually come from! // For example there is a script for projectile firing, that I usue for my laser. It asks for the speed for example in the UI on the right side of the screen. How? } } } } So what is wrong here? At last: A facepalm on me for not getting this, thanking you for the answer!