CPU for video editing

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

dbr1

Member
Jan 23, 2011
53
18
81
I was thinking about building a budget video editing system around the FX8350 but apparently benchmarks using Adobe Premiere Pro, which is what I use, are poor. This despite that the FX8350 has 8 cores and Premiere Pro is multithreaded.

I found this from a review of the Vishera:
The multi-core nature of the architecture makes the FX-8350 more competitive in our handbrake H.264 video encoding test, but a score of 2,701 points at stock and 3,164 points when overclocked is still slower than Intel’s Core i5-2500K, let alone the Ivy Bridge i5-3570K which scores 3,1160 points at stock and 4,245 when overclocked.
Can't quite understand why this CPU is not better for this application, particularly Adobe Premiere. I would build a cheap video workstation around one if it were.
 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
12 minutes 14 seconds,23.5 FPS, i5-2500k, Linux, pure RAM.

11 minutes 08 seconds, 26 FPS, i5-2500k, Windows 7 (64 bit version of Avidemux), HD to HD.

12 minutes 27 sec, 23 FPS, x6 1055t, Linux, read from disk write to RAM. System does not have enough RAM for pure RAM operation.
 
Last edited:

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
I was thinking about building a budget video editing system around the FX8350 but apparently benchmarks using Adobe Premiere Pro, which is what I use, are poor. This despite that the FX8350 has 8 cores and Premiere Pro is multithreaded.
It depends on exactly which program. The FX8350 comes out on top in my tests using Avidemux and it came out on top when Anandtech tests video coding as well:



The 8350 isn't so bad with Premier especially if you factor in cost:



Note: I do not have a FX-8350 and I wouldn't suggest the chip for most but for some it is a good value...Seems to work good for video coding.
 

colonelciller

Senior member
Sep 29, 2012
915
0
0
I was thinking about building a budget video editing system around the FX8350 but apparently benchmarks using Adobe Premiere Pro, which is what I use, are poor. This despite that the FX8350 has 8 cores and Premiere Pro is multithreaded.

I found this from a review of the Vishera:
The multi-core nature of the architecture makes the FX-8350 more competitive in our handbrake H.264 video encoding test, but a score of 2,701 points at stock and 3,164 points when overclocked is still slower than Intel’s Core i5-2500K, let alone the Ivy Bridge i5-3570K which scores 3,1160 points at stock and 4,245 when overclocked.
Can't quite understand why this CPU is not better for this application, particularly Adobe Premiere. I would build a cheap video workstation around one if it were.
Premier Pro requires a beast of a workstation to shine.
dual 8-core CPUs / 32-64 GB Memory / GTX 670+ gpu

with such demanding software it is easy to see how non xeons yield less than stellar results. how much do you want to spend ideally? i've recently assembled the parts for a 2011 dual socket build for simarly demanding software on as tight a budget as i could muster and might have a few pointers. currently a dual socket mobo with just one socket to be populated... cpus to be switched out after the next xeon upgrades are released

my build cost was in the range of $3500 but it could easily have been closer to $3000 if i hadn't splurged in a few spots.
 

MrX8503

Diamond Member
Oct 23, 2005
4,529
0
0
For video, you want more cores/threads over clockspeed. A good rule of thumb is at least 4+ cores and 2GB ram per core. So if you have an 8 thread CPU, you want at least 16GB ram.
 

dbr1

Member
Jan 23, 2011
53
18
81
with such demanding software it is easy to see how non xeons yield less than stellar results. how much do you want to spend ideally? i've recently assembled the parts for a 2011 dual socket build for simarly demanding software on as tight a budget as i could muster and might have a few pointers. currently a dual socket mobo with just one socket to be populated... cpus to be switched out after the next xeon upgrades are released

my build cost was in the range of $3500 but it could easily have been closer to $3000 if i hadn't splurged in a few spots.

I was thinking of 600 or 700 dollars for CPU, MB and memory to use as a stopgap for maybe a year or two. If I can't get something worthwhile for that, I will continue to save and wait to see whether Haswell, Ivy Bridge E or Haswell E is the best bang for the buck for this application.

Interesting comment on the Adobe Premiere forums regarding AMD processors and SSE extensions, anyone can verify this:
In our PPBM5 Benchmark there is a FX-8350, overclocked to 4.6 GHz and it comes out at rank # 563, only preceded by 562 Intel systems, despite the overclocking. In comparison, my old i7-920 comes out at rank # 49 currently. I believe that says it all.

The reason is that the AMD's all have very limited support for SSE extensions, that are used by PR all the time. AMD either does not support all required extensions or they are implemented in a very slow way. For around the same amount of money you could get an i5-3570K that is also unlocked, can be overclocked and is more than two times faster. Even an i5-3470 is way faster. See rank # 319.

 

Shephard

Senior member
Nov 3, 2012
765
0
0
Ok here is my comparison. I should have looked at lakedude's screenshot I just guessed the settings. My first time using avidemux.

pc:

i5-3570k @ stock.
8gb samsung golden ram
recorded to 7200rpm sata 3 hd.

video :

raw video - 17gb
game - starcraft 2
recorded with fraps @ 30 fps fullscreen
windows 7 64 bit

avidemux:

video output - mpeg4 avc (x264)
audio - copy
output format - avi muxer
32bit avidemux
priority - high

total time - 8 minutes

final file size = 735mb, 13 minute video no editing.

 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
My friend with the fx8350 has Sony Vegas but Vegas will not work with MKV files so we didn't test it.

Shephard based on your 85% CPU load I'm guessing you are IO limited. On all my tests the CPU load was 97 to 100% most of the time...

If you are IO limited a faster CPU will not help one bit.
 

Shephard

Senior member
Nov 3, 2012
765
0
0
My friend with the fx8350 has Sony Vegas but Vegas will not work with MKV files so we didn't test it.

Shephard based on your 85% CPU load I'm guessing you are IO limited. On all my tests the CPU load was 97 to 100% most of the time...

If you are IO limited a faster CPU will not help one bit.
what do you mean IO limited? I also did high priority and I also had Starcraft 2 running in the background as you can see from my screenshot. I don't know if that would affect rendering speed.

Difference between MKV and AVI muxer? Any quality difference?

By the way I uploaded to youtube took like 2 hours.

http://www.youtube.com/watch?v=iPxvzWEnQYo

looks pretty good even after the youtube compression. Although I have seen better Starcraft 2 uploads. So I am not sure what I did wrong.
 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
what do you mean IO limited?
IO = Input Output = The hard drive.

I mean that it is possible that the speed at which the RAW file can be read from the hard drive is the limiting factor, not the CPU.

Doing any job like this is a team effort and the job will only go as fast as the slowest link.

OK it looks like a typical HD can read at about 75MB/s which is about 4.5 GB/min so it looks like it would take about 4 minutes just to read the 17GB RAW file from the HD. This means even if your CPU was instantly fast and your write back was instantly fast, the process would still take about 4 minutes minimum because of the read speed of your HD.

BTW that SATA 3 speed thing is BS. You are limited by the mechanical speed of your HD. You will never hit the speed quoted for SATA 3.

See this chart for HD average read speeds:

http://www.tomshardware.com/charts/...ad-Throughput-Average-h2benchw-3.16,2901.html

I'll try to do a similar RAW file test like you did, once using an SSD (fast) and once using a normal HD. If there is a big difference it means that there is an IO limitation.
 
Last edited:

Shephard

Senior member
Nov 3, 2012
765
0
0
well I only have 120gb SSD with 70gb free so don't know if it's really worth doing it on an SSD.

Looking forward to your test.

My 7200 rpm is a Western Digital Black. The only better mechanical than that is a 10000rpm and I don't know anyone who renders who uses those. Waste of money too with SSD's out.

Did you see the quality of my video? What do you think?
 

Idontcare

Elite Member
Oct 10, 1999
21,110
59
91
Did you see the quality of my video? What do you think?

Quality looked good, a tad too much compression though IMO as one can see the compression artifacts around the text in the middle of the screen. Other than those artifacts though the rest of the image was beautiful.

(I played back at 720p from youtube with full-screen at my end)
 

Shephard

Senior member
Nov 3, 2012
765
0
0
Quality looked good, a tad too much compression though IMO as one can see the compression artifacts around the text in the middle of the screen. Other than those artifacts though the rest of the image was beautiful.

(I played back at 720p from youtube with full-screen at my end)
do you mean the text is not so crisp? I notice that. The colour is off a bit. It could have been fraps or the compression that done that. Nothing major though.

I didn't do 1080p because I don't got that monitor size. I could have probably done it though if I ran it in a window.

Maybe there is more compression because I chose AVI muxer? I wasn't sure what the difference was there. That was default option.
 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
Difference between MKV and AVI muxer? Any quality difference?
MKV and AVI are "container" formats. AVI is a much older Microsoft standard, while MKV is newer, open source and more flexible. Either may contain a H.264 video stream, the extension (AVI vs MKV) tell you nothing about the contents of the file.

If I understand correctly Matroska (MKV) allows for a higher quality stream. AVI has some limitations that I don't really understand but quality is not as good (something about "B" frames).

http://en.wikipedia.org/wiki/Container_format_(digital)

http://en.wikipedia.org/wiki/Comparison_of_container_formats
 

Shephard

Senior member
Nov 3, 2012
765
0
0
maybe I should render the video again as MKV and then it might be better?

I will also do the test a q8200 when I get a spare hard drive.

Do you think that Starcraft running in the background slowed down the rendering?

Why didn't you choose high priority?
 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
Go for x3970 X79 or AMD x6 phenom II or AMD FX 8350 and min of 16GB 1866 CL8 1T
Based on my tests the AMD x6 would be roughly the same as what the OP already has so the Phenom is not a feasible option.

OTOH the FX 8350 did very well in my video coding test (it was the best actually).

A stock 3770k would be something like 40% faster than a stock 2500k. Is that worth it?

The socket 2011 chips are just so expensive...
 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
maybe I should render the video again as MKV and then it might be better?

I will also do the test a q8200 when I get a spare hard drive.

Do you think that Starcraft running in the background slowed down the rendering?

Why didn't you choose high priority?
I just left everything stock. I was not doing anything else with the computer so the priority shouldn't matter much.

There are some (a ton) of quality settings you can play with. The container is not the only issue...Higher quality tends to equal higher finished file size...

You can play with settings and balance quality with file size. The juggling is a bit of an art form.
 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
Ok some new results

W7, 64 bit Avidemux, HD back to the same HD, 7 minutes 20 seconds (slightly faster than while testing with the SSD which was 7 minutes 28 seconds). The only reason I can think of for this anomaly is that I was surfing during the SSD test but the computer was left alone during the hd test.

Using core affinity on the 3770s to simulate a 4 core i5-3570, W7, 64 bit Avidemux, hd to hd, 8 minutes 55 seconds.
 
Last edited:

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
Summery (All times using FatDog Linux from live CD because all the software is consistent that way).

FX 8350 7 minutes 22 seconds.
i7-3770 8 minutes 7 seconds.
i5-2500k 12 minutes 14 seconds.
x6 1055t 12 minutes 27 seconds.

Windows 7 times with the 64 bit version of Avidemux were slightly faster...

The very best time was the FX8350 at 6 minutes 53 seconds on Windows 8 but no other system has Windows 8 so there is nothing to directly compare to around here...
 

lakedude

Platinum Member
Mar 14, 2009
2,778
528
126
Last test for today. Took a 812 MB 30 second clip of Far Cry 2 and converted to a 40 MB 1920 by 1080 MKV in 22 seconds using the SSD on the i7-3770s system.

Free unregistered FRAPS is limiting me to a 30 second clip with a watermark.
 

Shephard

Senior member
Nov 3, 2012
765
0
0
Ok this time I rendered the video as MKV Muxer instead of AVI Muxer. I just copied what I saw in lakedude's screenshot.

I didn't go into the 'configure' settings and change any of that. I am not sure if you can make it better quality in there.

Same clip, except I shortened it to 30 seconds so it didn't take hours to upload.

Can you notice a quality difference?

http://www.youtube.com/watch?v=bEn77ESWM-Q

Also file size difference.

MKV Muxer - 22.8mb (30 second clip)
AVI Muxer - 23.1mb (30 second clip)
 

Piroko

Senior member
Jan 10, 2013
905
79
91
The container will not change the quality or size of your video (well, mkv is some 100kb more efficient, but that's it). What you're looking for is resembling this:

probably in that 'configure' settings tab. Encoding mode "Target quality" or "constant Quality" is the easiest to understand, the number next to it can vary between ~15 (rather bad quality) to ~25 (great quality, but massive size). I think this encoding mode with a setting of 18 is standard setting, if that is the case then perhaps check out 20.
But keep a look on the file size, you might want to keep it at its current quality setting or else uploading will take even more time. Testing won't hurt though.
 

Shephard

Senior member
Nov 3, 2012
765
0
0
Ok Video Output > Configure.

I didn't touch these settings before as I said.

Encoding Mode - Constant Rate Factor (Single Pass)
Default Quality - 20

I think you have it backwards because on the slider it says '0' is at the High Quality end. '51' is at the Low Quality end.

Fast First Press - Checked.
Macroblock Tree Rate Control - Checked.
Framerate lookahead - 40
IDC Level - Auto

-

There are more tabs with more stuff to change.

I am not familiar with this program compared to Vegas. That said I am no Vegas expert either.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |