CPU for video editing

[paul3281]

I'm going to build a system got video editing and general internet and office use. Not really into gaming. I've a budget of about £120-£150 for the CPU.

Any recommendations?

Thanks.

[CAT-THE-FIFTH]

What is the total budget for the build and what parts and software do you already have??

[paul3281]

Hi, I’ve got most of the components, eg case, psu, SSD hard drive, blu ray drive. Will just need CPU, motherboard and ram. Budget for these about £250 max. Will be using win 8. I’ve got a good collection of software and my budget is just for hardware.

[CAT-THE-FIFTH]

What PSU do you have and what case?? What video editing software do you intend to use??

What is your current build??

I would get the following parts:

http://www.amazon.co.uk/AMD-FX8320-E...eywords=fx8320
http://www.amazon.co.uk/M5A97-R2-0-M...ywords=amd+970
http://www.amazon.co.uk/Crucial-BLT2...=8gb+ddr3+1866
http://www.cclonline.com/product/123...-Card/VGA1896/

That is around £20 above budget,but the graphics card should give you a useful boost if the software uses OpenCL.

[paul3281]

I've an Antec smartpower PSU. Will be using rebox.net for video editing.
Have had a look at http://www.amazon.co.uk/AMD-FX-6350-...productDetails

Any good?

[CAT-THE-FIFTH]

Thats an old PSU,I would probably get a new one TBH! Have you got a PCI-E graphics card?? What case do you have??

The FX6300 and FX6350 should be fine too,as they beat the similarly priced Core i3 CPUs,but the FX8320 is a great budget CPU for video encoding IMHO.

The FX6350 should be around Core i5 3470 to Core i5 3570K level for video encoding AFAIK.

I would get this PSU:

http://www.cclonline.com/product/554...tion-/PSU0202/

Edit!!

Out of interest what are you upgrading from??

[ik9000]

had a quick browse through the anandtech bench results (see here: http://anandtech.com/bench/CPU/26)

Looks like it will depend which kind of encoding you'll be doing as to quite which Intel chips the FX rubs shoulders with... Unsurprisingly the FX8000 series is significantly faster than the FX6300 if you can afford the extra.

it's amazing to see though that even in the worst cases the FX chips are akin to the old preimum Intel chips of 3 years ago eg the FX8350 is faster than i7-965 on encoding/decoding. And considering the pricing that is quite something.

[Bonebreaker777]

had a quick browse through the anandtech bench results (see here: http://anandtech.com/bench/CPU/26)

Looks like it will depend which kind of encoding you'll be doing as to quite which Intel chips the FX rubs shoulders with... Unsurprisingly the FX8000 series is significantly faster than the FX6300 if you can afford the extra.

it's amazing to see though that even in the worst cases the FX chips are akin to the old preimum Intel chips of 3 years ago eg the FX8350 is faster than i7-965 on encoding/decoding. And considering the pricing that is quite something.

Well, I personally wouldn't call it significant difference, but there is a measurable. The rather interesting thing is that the difference between the FX-8320 and the FX-6300 is about the same as the difference between the FX-6300 and A10-5800K.

On the other hand, if this rebox.NET supports QuickSync, I would go without hesitation for a Intel CPU. Unfortunately I can't find any further information about this.
Just a comparison about the power of QuickSync in supported programs > AMD A10-5800K GPU vs Intel QuickSync.

[AETAaAS]

rebox.net is primarily a remuxer and so does not recode video and so Quicksync would not come into play.

I tried Quicksync for recoding and in my experience, it is a pretty poor system, like most fixed hardware encoders are when compared to x264. While fast, they typically deliver encoding which produces large files and strange atrifacting in certain scenes. If it's just to throw something onto your mobile for transient use, it's ok. But if you wanted to compress video while maintaining some measure of fidelity, I wouldn't use it.

[Bonebreaker777]

Haven't read about the artefacts anywhere yet. And haven't even seen any, even if I do 20x movies a week :-?
The larger files are not that significant, see the Hexus Handbrake results.

[AETAaAS]

Perhaps it depends but I tend to see artifacting on edges and on 720 and 1080 videos, high contrast edges like vibrant colours tend to have it worse.

The filesizes one should be qualified, the disparity grows as you drift from the normal profiles to more complicated encoding. Increasing reference and b frames, trellis, psyrd and some other tweaks can help drop file sizes with no appreciable quality loss. Fixed function encoders like QS don't benefit from tweaks which can be done to help specific inputs.

I play around with encoding animation because there is so much of it on the web that you can use without getting into trouble, and I was fascinated by encoders like ONS, Hi10 Anime and sites like MiniTheatre that they could encode 720p episodes to anywhere between 60-120mb with no appreciable difference from originals 3-7 times the filesize.

Animation is also the place where artifacts can be readily spotted as sharp lines and bright flashes/sparks make easy comparison points.

edit; I should also say that I encoded for redistribution sometimes, so there was a point where I really nitpicked and it was a case of me alt-tabbing between two players looking at specific frames on a 25' monitor about a foot away from my face, wondering if the benefit was appreciable enough to justify a 90 minute encode for each episode when there was a big backlog.

Also, the filesizes were quite a bit bigger now that I went back and had a look. On your own testing, one of the videos came out on average about 50% larger with QS enabled.

[watercooled]

rebox.net is primarily a remuxer and so does not recode video and so Quicksync would not come into play.

I tried Quicksync for recoding and in my experience, it is a pretty poor system, like most fixed hardware encoders are when compared to x264. While fast, they typically deliver encoding which produces large files and strange atrifacting in certain scenes. If it's just to throw something onto your mobile for transient use, it's ok. But if you wanted to compress video while maintaining some measure of fidelity, I wouldn't use it.

I second this entirely TBH - QS is generally OK to quickly get a video onto your tablet/phone to watch during a journey or something, but for anything I intended to keep, I'd still be using software any day of the week. The difference isn't always massive, but as you say there are times where QS does make a pig's ear of it; that's just not acceptable for archival media IMHO.

[Bonebreaker777]

Also, the filesizes were quite a bit bigger now that I went back and had a look. On your own testing, one of the videos came out on average about 50% larger with QS enabled.

Ah, touché I beg your pardon. Well, I use it anyway for "re-distribution" only. Or do quick Apple conversions.