Go to youtube and check.

In my opinion UDK is much better. First of all udk have much better graphics and it is easy to use.

The main difference is about 1.5GB in download size. UDK is 2gb and Unity3D is about 400mb.

The engines target different things.

UDK is a big toolkit that lets you make (basically)standalone Unreal mods, and targets beefy hardware.

Unity is smaller, user friendly, more flexible game engine that targets a wider variety of platforms and hardware specs.

Regardless of which engine you use, you are going to have to adjust your content and design to fit into the engine's way of doing things. UDK is overkill for most indy projects, and can be a pain in the ass to work with (been using it on and off since 1998, and I have never liked their way of doing things). Unity IMO is much easier to work with, and more user friendly. It's also easy to just throw your work up on a dropbox-like site, and let people try it in their web browser with the plugin.

Feature-wise, there isn't as much in it as you might expect. The UDK has the edge when it comes to rendering, but I'd like to think that the gap is shrinking with recent Unity releases. Unity has a significantly easier learning curve, but the toolset isn't quite as mature. Unity has more publishing options...

it is worth considering that the licensing terms/costs are drastically different:

Unity has a free indie license, but the feature set is significantly crippled - my impression is that most serious developers will be looking at a commercial license instead. A commercial unity license costs $1,500 per developer (plus $400 for each of iOS and Android publishing, if you need those), but you never have to pay any royalties.

The UDK requires a $100 publication license when you are preparing to ship your product, but also charges 25% royalties on any revenue after the first $50,000.

This presents you with an interesting cost equation. Unity's per-seat licenses are expensive upfront, but after that you never need to pay a dime. The UDK costs pretty much nothing upfront, but if your game produces a high revenue, you will be paying a fair bit of money out in royalties. My back-of-the-envelope calculation suggests that the UDK might be more cost effective for a game that grosses under ~$120,000, and Unity is significantly more cost effective for a game that exceeds that figure - but you should run your own projections to validate the choice for your particular situation.

People severely overstate the missing features in the free Unity. It mostly comes from the crowd who can't actually make a game and just likes to screw with advanced rendering options. "Hey look, I can give my procedurally generated crap real time shadows, or do deferred rendering on it". None of that has anything to do with making an actual working game.

Unity Free lacks real time shadows, render to texture, and a few small things. There is nothing stopping you from making a completely functional game, and then just turning on those features when you upgrade. For someone like me who bakes their lighting, and needs circle shadow blobs under characters anyways, there is no perceivable difference between the free and paid version other than the 2 second Unity watermark in the bottom corner when I run my game.

Once you've actually finished a game to the point of being able to do a proper release, the upgrade price is negligible, and in line with all your other required expenses. Especially since most people using Unity3D will never get to the level of completion anyways.

UDK has a better renderer at the moment, but your graphics quality is ultimately up to the quality of your content. A nice, properly colored, well composed scene will look great in any engine, and crappy, poorly colored, badly composed programmer art will look bad in any engine.

If you are making a simple 2D game, it's ill advised to use either of these engines.

[quote name='Daaark' timestamp='1358878192' post='5024375']
People severely overstate the missing features in the free Unity.[/quote]

Aye, that's a fair point. I'm building the entire game world via GPGPU procedural generation, so the lack of render-to-texture is a non-starter for me. YMMV.

Lack of RTT hurts me a bit too. I want to render something onto the screen of an arcade machine, but I will have to do without for now. It's just a background detail anyways, and doesn't effect my game one way or the other. I'll use a UV coords hack to atlas some frames and get around it.

it is worth considering that the licensing terms/costs are drastically different:

Unity has a free indie license, but the feature set is significantly crippled - my impression is that most serious developers will be looking at a commercial license instead. A commercial unity license costs $1,500 per developer (plus $400 for each of iOS and Android publishing, if you need those), but you never have to pay any royalties.

The UDK requires a $100 publication license when you are preparing to ship your product, but also charges 25% royalties on any revenue after the first $50,000.

This presents you with an interesting cost equation. Unity's per-seat licenses are expensive upfront, but after that you never need to pay a dime. The UDK costs pretty much nothing upfront, but if your game produces a high revenue, you will be paying a fair bit of money out in royalties. My back-of-the-envelope calculation suggests that the UDK might be more cost effective for a game that grosses under ~$120,000, and Unity is significantly more cost effective for a game that exceeds that figure - but you should run your own projections to validate the choice for your particular situation.

One other hard fact is that you will have to shed a huge amount of cash out if you need to change something in the source codes as well. I guess options like Ogre and Torque could be worth considering as well. In fact this link could help out also: free engines

A lot of those engines on that list aren't that good, or severely limited.

Ogre is just a rendering engine. Nothing wrong with that, it does it's job very well. But that's an apples to oranges comparison. It's not a game engine.

Irrilict was a mess. Nothing worked very well. Basically made by a group of guys who just copied everything Quake2 did. Anytime you see an engine, and all the supported functionality revolves around what quake did, and quake file formats AVOID AT ALL COSTS! Forum is full of 'tard know-it-all, OSS zealots, who can't help you with anything when it comes to making games.

GTKRadiant (which is on the list) is not a good editor in 2013. I say this as someone who used to live in various flavors of Radiant for years and years. Unless you are modifying old ID games, it won't do anything for you. It's not 1996, we can model with real polygons now, and not have to be limited to using an outdated brush system, and a unneeded power of 2 coordinate system. Brushes are abstractions over polygons, and you can't even drag a vertex without the editor modifying the rest of the brush to maintain a legal shape. Unless you are making a corridor shooter made out of very coarse and simple shapes, it will fail you miserably.

Panda3D is awesome if you are using Python. It advertises C++ compatibility, but when I tried it years ago, there wasn't any C++ documentation, and they didn't make it very easy to use it.

Blender's built in game engine is good for Python too. Just model stuff, add some logic blocks, and start the game engine to watch it all in action. Very good for prototyping ideas.

The rest of that stuff is not worth anyone's time. It's all just someone's hobby project. They don't have good support, or knowledgeable user bases. Unity and UDK have tons of KNOWLEDGEABLE people making cool stuff to get help and inspiration from. With Unity at least, you can just google "Unity3D <Whatever problem you are having>" and get a ton of results from their forums and answers site, and wiki.