Hey,
1) For video it's really complicated, so I'm just going to give you a BEST CASE SCENARIO using Handbrake.
Handbrake is almost completely proportional to the overclock. Thus, if we overclock it by 10% we get the video done 10% faster. The default max Turbo for the i7-4790K is 4.4GHz, and for the i7-4790 it is 4GHz.
So 10%, though you can go as high as 4.8GHz if you win the "silicon lottery) and your motherboard, settings, and cooler are all adequate and setup. I don't personally push things to the limit, and as you said that needs a much better CPU cooler.
So, up to 20% is possible. 10 to 15% more practical.
Other editing programs are often more limited by the speed of one of the cores not total processing power for much of the editing process (final render may be more like the Handbrake example.)
2) You can also TWEAK the CPU even if it's non-overclockable. I have an i7-3770K but this probably still applies->
*You can adjust the TURBO values. For example, on an i7-4790 the max Turbo of 4GHz might only be for one main core usage. So you may see FOUR different multipliers such as 40, 39, 38, 37. Thus running at roughly 3.7GHz when rendering video (you can look at TASK MANAGER).
On my i7-3770K I did a quick overclock (no voltage), adjusted my multipliers and I run at 4.2GHz (full load) instead of the default 3.6GHz in the same situation.
3) GAMING?
Not really.
In general, with your CPU it's not going to make much difference at all. If the CPU is fast enough (it's usually the main thread of code that's the issue) then running the CPU even faster doesn't make a difference.
In the 2013 Tomb Raider, look at the big graph to see what CPU SCALING looks like:
http://www.techspot.com/review/645-tomb-raider-performance/page5.html
Now look at Fallout 4 which is very poorly optimized. Really not acceptable:
http://www.gamersnexus.net/game-bench/2182-fallout-4-cpu-benchmark-huge-performance-difference
*You can see a big difference between an i7-4790K and i5-4690K. A big part of the difference is the 500MHz default frequency difference but it's relevant to your question (The "4GHz" for the i7 is the non-Turbo speed so it's probably running about 4.1GHz in reality as per my discussion near the top about turbo). It's really NOT common to see a gaming need for more than an i5-4690 or similar though.
It completely varies by the game. If you'd had a weaker CPU like the FX-4300 and a good GPU the difference between games can be well over 40% vs a modern Intel (like minimal bottleneck in Tomb Raider, but big bottlenecks in some MMO's, Skyrim, Starcraft 2 and others).
4) i5 vs i7->
A faster i5 can make sense as well, especially if it saves a $100 or so.
The i7 mainly benefits really well threaded workloads, however being able to overclock a similar i5 can give similar or even better performance in some circumstances.
$100 for a GPU on the other hand goes a long way.
5) CPU cooler->
You didn't ask, however my top two recommendations for budget coolers (for a modern i5/i7) are:
a) Cooler Master Hyper 212 EVO
b) Cryorig H7
The H7 is my top choice. The "EVO" is okay but the fan gets noisier than I'd like. Far better than stock though. You also need to adjust the fan profile (usually the motherboard fan software).
6) System Memory must be fast enough to prevent CPU bottleneck. Do not use one stick only. I generally recommend->
i5/i7 4xxx-> DDR3, 1866 or 2133 CAS9 (2x4GB, or 2x8GB kit).
i5/i7 6xxx (Skylake)-> DDR4, 2666MHz CL16 (or close; there was a pretty in-depth article, but pricing doesn't affect total PC cost much. Maybe $15 or so for a better 16GB kit.)
7) What about SKYLAKE CPU's?
Maybe an i5-6600K?
Depends on total budget, and difference in cost to a similar Haswell setup. There's a few advantages you may or may not care about but that's too long as I made a book already with my post.
*You MIGHT want to consider building without an addon GPU first and get an AMD Polaris or NVidia Pascal card instead. My MAIN reasons for this are:
a) AMD-> current GPU's put out a lot of heat (my room would get too warm), but
b) NVidia-> performance loss due to lack of ASync Compute hardware for future DX12 titles is uncertain. We won't have RELIABLE information for a while, but we have some indications with two beta titles (Hitman, AotS). It's NOT clear how much NVidia can tweak their drivers.
Both of the new architectures are not only really power efficient, but we'll also see future games use their hardware much better (put another way, current GPU's won't do as well comparatively in future titles).
*It's not a BAD thing to buy now. I'm still rocking a GTX680 and love it (similar to GTX770), but for me it's not nearly good enough for some upcoming titles including STAR CITIZEN.