Compressor VS Adobe Media Encoder CC speed

[orbitalpunk]

Has anyone compared these two on a 2013 Mac Pro? Other macs have different processor with different instruction sets and GPU's so they won't apply. I am seeing AME being twice as fast as compressor. Are others seeing the same thing?

 

[joema2]

Has anyone compared these two on a 2013 Mac Pro? Other macs have different processor with different instruction sets and GPU's so they won't apply. I am seeing AME being twice as fast as compressor. Are others seeing the same thing?

I don't have a Mac Pro but I've tested export speed of Premiere CC 2017 vs FCPX on a 2015 iMac 27 many times. In general FCPX is about 4x faster.

At 19:23 in this video, Max Yurev tested 4k H264 export time on a 2013 Mac Pro using both Premiere and FCPX. FCPX was 4.5x faster:

This was an interesting case because the Xeon CPU in the Mac Pro does not have Quick Sync, which is normally the stated reason FCPX is so much faster than Premiere. So even excluding Quick Sync, FCPX was much faster than Premiere CC. That was apparently not AME.

However Quick Sync is obviously a factor because the iMac was exported much faster than the Mac Pro (both running FCPX).

Here is a video comparing video export times of FCPX (inc'l Compressor) vs Premiere the 2016 MacBook Pro vs the Dell XPS 15:

In the above video, Compressor was much faster than Premiere, but he doesn't say whether it was AME or not. However exporting from Compressor was slower than FCPX. Compressor has supposedly been upgraded with performance improvements over the last year or so, and is supposedly as fast as FCPX but in my tests it's generally slower. However (like AME and Premiere itself) Compressor has so many export options that doing an equal test requires a lot of study and work. It is easy for some deeply-nested option to throw off a timed test.

 

[Ritsuka]

If you are compressing to H.264, avoid both and just use whatever app you want that includes the x264 encoder.

 

[orbitalpunk]

i made a screencast comparing FCP vs AME since doing it from FCP timeline is faster than Compressor alone. FCP did transcode the video as well.

Times:
FCP: 52 seconds
AME: 23 seconds

Compression Setting:
.h264 1280x720
2 VBR Multi-pass 16mbit
AAC 320kbps
Profile High 3.2
Entropy CABAC
I turned off Add Clean Aperture Information for FCP for faster speed
I turned on Maximum Render Quality and Maximum Render Depth for AME. Almost doubles the encode time.
Instances was turned Off in Compressor as it is slower when turned On

Software Version:
Compressor 4.3
FCP 10.3
AME CC2017.00

Source Video: http://4ksamples.com/4k-chimei-inn-60mbps/
I trimmed it down to 10 seconds in QuicktimeX

So why is AME more than twice as fast when everyone is saying FCP/Compressor is faster? If anyone can review the video and tell me what I'm doing wrong, I'd really appreciate it.

Thanks

 

[joema2]

i made a screencast comparing FCP vs AME since doing it from FCP timeline is faster than Compressor alone. FCP did transcode the video as well...
So why is AME more than twice as fast when everyone is saying FCP/Compressor is faster? If anyone can review the video and tell me what I'm doing wrong, I'd really appreciate it...

I watched your video and ran 13 different tests using both FCPX 10.3.1, Premiere CC 2017 and AME 2017 to try and figure this out. My source media was a 60 second 3840x2160@29.97 H264 from a Panasonic GH4. Encode output in most cases was H264 720p/29.97. Hardware was a top-spec 2015 iMac 27 with content on a Thunderbolt 2 RAID-5 array. My results below:

FCPX (File>Share>Master File>Settings: Computer, H264 Faster Encode, 3840x2160): 44.2 sec
FCPX (File>Share>Master File>Settings: Computer, H264 Better Quality, 3840x2160): 1 min 28 sec
FCPX (File>Share>Master File>Settings: Computer, H264 Faster Encode, 720p): 21.2 sec (3x real time)
FCPX (File>Share>Master File>Settings: Computer, H264, Better Quality, 720p): 41.4 sec
Compressor 720p/29.97 H264 2-pass, 16 mbps: 45.1 sec
Compressor 720p/29.97 H264 1-pass, 16 mbps: 26.0 sec

Premiere CC 2017 (Match source settings, ie 3840x2160): 5 min 11 sec
Premiere CC 2017 720p/29.97, VBR 2-pass, 16 mbps: 2 min 57 sec
Premiere CC 2017 720p/29.97, VBR 1-pass, 1 min 30 sec

AME CC 2017: H264 720p/29.97 VBR 2-pass, 16 mbps: 2 min 55 sec
AME CC 2017: H264 720p/29.97 VBR 1-pass, 16 mbps: 1 min 20 sec

Comments: FCPX and Compressor were faster in all cases, often 4x faster. I can't explain the performance difference you've seen. You are on a nMP which does not have Quick Sync but in the previously-posted video test on a nMP, FCPX was still much faster than Premiere.

In general I'd recommend exporting directly from FCPX using the Computer/H264/Faster Encode setting. This is single-pass but I cannot visually see any difference between this and multi-pass. I generally use single-pass on all proof copies and (if I remember) switch to multi-pass for the absolute final export, but sometimes not even then.

For H264 720p export using the above settings, FCPX uses a bitrate of about 12 mbps, which is not far from the 16 mbps you selected. 12 mbps is more than double the bitrate that Youtube recommends for 720p.

I did two final tests comparing FCPX export performance of 4k material using both a 4k project (ie "sequence"), and a 1080p project. Exporting 720p from a 4k project was about 2x slower than from a 1080p project. All of the above tests for FCPX and Compressor used a 1080p project since that is the common practice for editing and exporting 1080p or below from 4k material.

FCPX enables full resolution of the underlying 4k material when editing in a 1080p project -- you can crop/zoom into the frame and still get the full 4k benefit, yet many operations will go faster. I have never tested this before but that apparently includes export performance. If you ever need to do a full 4k export you can simply copy/paste the 1080p timeline into a 4k project and that will enable 4k export -- all the edits will still be there.