I7 5820K vs I7 4790K Streaming PERFORMANCE DROP

MrJagsterS

Reputable
Dec 26, 2014
56
0
4,630
Before we start I'd like to say I don't expect this to happen for a while and anyone who takes part in my proposition I am most grateful :)

I'm looking to build a gaming/streaming PC however I'd like to know which CPU would cause a greater FPS drop when streaming. I know that the 5820K is gonna be the winner but my question is by how much.

My proposition is to anyone with either a 4790K or a 5820K, is to record an average FPS for a popular AAA game e.g. BF4, CS:GO, Star Wars Battle Front etc, then record an average FPS when streaming using the following settings:
CBR and ENABLE CBR PADDING
Bitrate 2700
Audio Bitrate 128
Base RES 1080p
downscale 720p
Filter Lanczos
FPS 48
264 CPU preset veryfast (I may do a different thread comparing veryfast and faster if this goes well)

If you could clearly post the before and after FPS with your CPU I would be most greatful :)
I hope that this not only is educational for me but also for the rest of the community.
Thanks xD

Side note - I know not everyone has the exact same systems but as we are doing FPS change, it should be accurate enough for me.
 
Solution
If you have NVENC (on nvidia gpu) or AMDs equivalent (AVE i think), youre better off using that, it gives virtually no effect on gaming, cpu encoding is a little old time now.

i did try nvenc vs cpu (both at max quality for recording, then about 2500kbits for streaming) and both times nvenc made no effect on fps, while cpu encoding dropped me between 10-40 fps depending on cpu load (higher drop when cpu got closer to 100%).

My machine is i7 4790k with titan x, tested at 1080p ultra settings in cod black ops 3.

MrJagsterS

Reputable
Dec 26, 2014
56
0
4,630


Everything you say I already know. That was not the point of this thread. Please only reply with information relevant to the test that I have proposed :)
 

Uberragen21

Distinguished
Sep 3, 2009
285
1
18,810
The limiting factor will never be your CPU when comparing two high end CPUs for game streaming. Your Internet speed and quality (both upstream at home and downstream wherever you are gaming remotely) will have a much greater influence on your streaming quality than those two high end CPUs.

thor220 is correct in that going for the 5820k will cost you a lot more in the short run and you won't realize any perceptible gain over the 4790k.
 

MrJagsterS

Reputable
Dec 26, 2014
56
0
4,630


I take on board what you are saying. don't worry though, my internet will be more than fine. However I am not asking about making the best streaming PC in this thread. I am simply curious about the different FPS drops due to streaming with the CPU's. I understand the results may vary with different people having different internet speeds but hopefully most people who have high end PC's will also have "Good enough internet".
 
You're asking people to do an awful lot and you're asking them to do it in an uncontrolled environment. In this manner, even if someone did want to go through all the trouble, their results would likely have no statistical significance because of the way the tests were conducted.

You'd have much better luck looking for professional reviews that can do this in a controlled environment.
 

gotovato

Reputable
Nov 15, 2014
438
0
4,860
I've done both. My i7 4790k was at 4.8ghz. Handled bf4 at 1080p 60fps with barely any issues at all. This was at 5000kbps. My in game fps did take a hit for sure, but maintained above 60fps full ultra settings. This is with overclocked 780's in sli. My CPU usage would steady hit 75+% and at times hit 100% and my frames would dip but never below 60(if I remember correctly). My 5930k@4.3ghz? Well...I've streamed a bit of battlefront/bf4 with it at 1080p 60fps 5000kbps bitrate(which is too high for twitch) but my CPU usage hangs out around 60% hitting up to 80%. Frame rates are well above 60, closer to 100 average. The 5930k does handle the streaming better, on the obs website they actually recommend not streaming battlefield 4 because its so CPU heavy. Basically, if you can afford it, and want to stream 1080p 60fps with a bitrate of 3000-4500kbps(that's the sweet spot) the 5820k/5930k will handle that better. And I really wouldn't worry about the slower clock speed and losing a few fps at stock. Get a good cooler, good motherboard, and overclock the x99 chip. Of course this adds to cost but that's just how it goes.
 

ValkyrieStar

Reputable
Oct 26, 2015
23
0
4,520
If you have NVENC (on nvidia gpu) or AMDs equivalent (AVE i think), youre better off using that, it gives virtually no effect on gaming, cpu encoding is a little old time now.

i did try nvenc vs cpu (both at max quality for recording, then about 2500kbits for streaming) and both times nvenc made no effect on fps, while cpu encoding dropped me between 10-40 fps depending on cpu load (higher drop when cpu got closer to 100%).

My machine is i7 4790k with titan x, tested at 1080p ultra settings in cod black ops 3.
 
Solution

ValkyrieStar

Reputable
Oct 26, 2015
23
0
4,520


Id still recommend using nvenc or ave directly on the gpu, takes away the load from the cpu and doesnt effect fps from what ive found. Either CPU will be fine in that case but id recommend a 5820K for longevity, 6 cores is good, and you could potentially upgrade to 8 or even 10 cores (broadwell e is expected to have 10 cores).
 

MrJagsterS

Reputable
Dec 26, 2014
56
0
4,630
Used to use nvenc but quality was bad. I'm investing in good hardware for good quality stream. I will probably go with 5820k as I would like to not have to upgrade for a long time.
 

ValkyrieStar

Reputable
Oct 26, 2015
23
0
4,520


5820K will be your best bet, currently i have 4790K and a Titan X, will be getting the 5960X or 6950X in the near future.

If you wish, im streaming CoD:BO3 currently at 720p using nvenc, i would use higher if only i had better upload speed (50 down 3 up); http://twitch.tv/spartanc001

With nvenc i find its very good for recording, and produces very nice quality video, but for streaming the low bitrate doesnt fare well for it.