FX6: The Sony ILME-FX6V Owners Club

There was a Facebook discussion about the M1 Max laptops yesterday and someone asked for a specific test so he could compare it to his desktop PC. He wanted to know how fast it would export a 1 minute 4K video. So I tried it on my M1 Max MacBook Pro with a 1 minute clip from my FX6. Using the XAVC-I codec so 4K UHD, 23.98fps, 10 bit 4.2.2. Exported the 1 minute of 4K to 1 minute of 4K H.264 in….35 seconds. If anyone (Doug?) is interested in any
other specific tests, just let me know what you’d like me
to test.
 
alaskacameradude said:
So I tried it on my M1 Max MacBook Pro with a 1 minute clip from my FX6. Using the XAVC-I codec so 4K UHD, 23.98fps, 10 bit 4.2.2. Exported the 1 minute of 4K to 1 minute of 4K H.264 in….35 seconds.
Click to expand...

That sounded slow to me so I ran my own testing just now. You didn't say what software you are using so I tried with both Resolve and Premiere.

2018 MacBook Pro
2.9 Ghz 6-core Intel i9
32GB memory
Radeon Pro Vega 20 4GB
Premiere Pro 22.2
Resolve 17


I think I matched your settings: 60 second 4K XAVC-I @ 23.98 clip recorded with FX6.
Rendered out to a 4K H.264 MP4 file with default settings.

Premiere: 26.7 seconds
Premiere with a simple text graphic superimposed over the video: 28.8 seconds

Resolve: 24.03 seconds
Resolve with a simple grade using curves: 24.06 seconds

That's about 45% faster than your M1 Max.
 
Last edited:
Doug Jensen said:
That sounded slow to me so I ran my own testing just now. You didn't say what software you are using so I tried with both Resolve and Premiere.

2018 MacBook Pro
2.9 Ghz 6-core Intel i9
32GB memory
Radeon Pro Vega 20 4GB
Premiere Pro 22.2
Resolve 17


I think I matched your settings: 60 second 4K XAVC-I @ 23.98 clip recorded with FX6.
Rendered out to a 4K H.264 MP4 file with default settings.

Premiere: 26.7 seconds
Premiere with a simple text graphic superimposed over the video: 28.8 seconds

Resolve: 24.03 seconds
Resolve with a simple grade using curves: 24.06 seconds

That's about 45% faster than your M1 Max.
Click to expand...

Wow that’s crazy. I work my gear for a long time between
upgrades so I was upgrading from a Sony FS700 and a
2014 MacBook Pro, I didn’t know the 2018 MacBook Pro
was that fast. For info purposes, I was using FCPX not
Premiere or Resolve. Also had the Sony
Slog 3/SGamet3Cine LUT applied to the clip without knowing it as FCPX did that automatically.
 
I'd be shocked if an Intel MBP were faster than an M1 Max MBP, based on all the benchmarks I've seen. But I think the big difference in H.265 handling just comes down to hardware accelerated graphics. And if you could somehow add the M1 Pro/Max media engine to an Intel, it would level the playing field when dealing with those compressed formats. Even my M1 Mac Mini plays back H.265 far better than my i7 MBP. (But once you start adding color correction and effects, my M1 Max MBP outshines the M1 Mini)
 
Just to add another data point:

setup:
2021 14" MacBook Pro
M1 Max, 10-core
32 GB ram


rendering out a 60s UHD XAVC-I 23.98 clip (from my FX6) to an H.264 MP4 in Resolve:

no corrections:
17.81 seconds

with LUT:
17.69 seconds (difference is probably due to user error with the stopwatch—no significant difference)

with LUT and curves:
18.16 seconds (still seems insignificant: Resolve clocks all of these at 17 seconds)

with LUT, curves, and Noise Reduction effect (default settings: just added the effect to a node):
1:05.26
 
drboffa said:
Just to add another data point:

setup:
2021 14" MacBook Pro
M1 Max, 10-core
32 GB ram


rendering out a 60s UHD XAVC-I 23.98 clip (from my FX6) to an H.264 MP4 in Resolve:

no corrections:
17.81 seconds

with LUT:
17.69 seconds (difference is probably due to user error with the stopwatch—no significant difference)

with LUT and curves:
18.16 seconds (still seems insignificant: Resolve clocks all of these at 17 seconds)

with LUT, curves, and Noise Reduction effect (default settings: just added the effect to a node):
1:05.26
Click to expand...

That is more like the speed improvement I'd expect to see and possibly good enough for justifying an upgrade.

@ alaskacameradude
Where is your media stored? Are you using a fast external SSD or what?
I'm going to bet you are using a slow HDD or the internal drive on your computer. If so, that is why your speeds are so much slower than ours. Drives make all the difference.
 
Doug Jensen said:
That is more like the speed improvement I'd expect to see and possibly good enough for justifying an upgrade.

@ alaskacameradude
Where is your media stored? Are you using a fast external SSD or what?
I'm going to bet you are using a slow HDD or the internal drive on your computer. If so, that is why your speeds are so much slower than ours. Drives make all the difference.
Click to expand...

Doug,
I was using the internal drive. However, the internal
drive is an SSD with speeds which make external SSDs
look slow by comparison. It is 2TB and I have about 130GB
worth of applications on it so almost 1.9TB free on it. It
’should’ be faster than any external SSD by a very significant
amount. But you have made me curious so I’ll test to
an external SSD….
 
This interests me. I always thought that media needs to be on a different drive to the operating system for maximum speed - so that's what I've done ever since I started in this business, usually putting media on external Thunderbolt RAID0 or RAID5 systems (all G-Tech GRAID kit - I've loads of it!).

After months of waiting on a back order I have just got a new 16 inch MBP M1 MAX with 64GB RAM - lovely bit of kit, gorgeous screen! I also bought a Caldigit Thunderbolt 4 Element Hub and a 4TB Sandisk Extreme Pro SSD USB 3.2 Version 2, this thing:

https://smile.amazon.co.uk/SanDisk-...xtreme+PRO+Portable+SSD+V2,aps,50&sr=8-3&th=1

Plus... I bought a Sandisk/Western Digital 2TB Extreme Pro Thunderbolt 3 SSD - link here:

https://www.westerndigital.com/en-g...rive-pro-thunderbolt-3-ssd#SDPS51F-002T-GBANB

I also bought a 24TB GRAID 2 Thunderbolt 3 system - more for more general storage and quick access for media files that are not archived on my dozens of Terabytes of (mostly spinning disc) drives elsewhere, this one:

https://cvp.com/product/sandisk-pro...szrcdEvwWT2EP_IerQGVcyp6O6W4xfxUaAgfLEALw_wcB

I have yet to use this system in anger (just got it!) but initial Write and Read speed tests using the Blackmagic utility with the drives attached suggest the following:

WD/Sandisk G-Drive 2TB Extreme Pro Thunderbolt 4 = Write around 1,984 MB/s, Read around 2,540 MB/s.

Sandisk Extreme Pro 4TB USB 3.2.2 = Write around 872 MB/s, Read around 833 MB/s.

GRAID 2 (set up as RAID0) Thunderbolt 3 = Write around 454 MB/s, Read around 467 MB/s.

If I do the Blackmagic speed test on the internal Mac's SSD (it won't let me do that directly as it says it's read only) but if I select say "Shared" I get over 6,000 MB/s Write and a Read speed of well over 5,300 MB/s...just wow!

If I'd have known that before I might have bough the otherwise maxed out spec MBP M1MAX with a much bigger internal drive!

Not sure how useful this might be - but thought I should share in view of the current discussion on this (otherwise FX6 camera related) thread!
 
The idea of separating the system/applications and scratch/other was taught for probably two decades by now, but I also started using my internal SSD about 7-8 years ago as there were no differences in my case and the speeds were already incredible (which is why every MBA or MBP I get always has a pretty hefty SSD).

But back in the day, you'd get some pretty major improvements separating them or of course using a RAID (as mentioned).

With improved technology and much better software/hardware, using the internal SSD for everything is okay for many.
 
NorBro said:
The idea of separating the system/applications and scratch/other was taught for probably two decades by now, but I also started using my internal SSD about 7-8 years ago as there were no differences in my case and the speeds were already incredible (which is why every MBA or MBP I get always has a pretty hefty SSD).

But back in the day, you'd get some pretty major improvements separating them or of course using a RAID (as mentioned).

With improved technology and much better software/hardware, using the internal SSD for everything is okay for many.
Click to expand...

How would explain that my 2018 MacBookPro appears to be 45% faster than alaska's new M1 Max when rendering the same type of footage?
 
Usually it's user error or people lie and/or exaggerate to prove their point.

I don't know though; I don't make numbers conclusions based on what's said on forums if I have to accurately gauge results.
 
If you want, post a 60-second clip, something beefy with moving trees or something else (actually almost anything from your nature footage would be perfect), and I'll tell you - or record it - how long a M1 took just exporting the S-Log3 as is - in its native 4K timeline - from FCP X.

At least a simple base test like this could be interesting, even though there is no heavy effects, color grading, etc. to make it interesting.
 
Yeah I had the same idea. I don't have time this week (just wrapped a 14 hour day and still have email to do) but probably next week I could upload something.
Another possible explanation is the bit rate of the target render. He said 4K H.264, but didn't say VBR or CBR and at what rate? For example, 50Mbps CBR might render at a different speed than 25Mbps VBR. This aspect of the comparison was never discussed.
 
Doug Jensen said:
Yeah I had the same idea. I don't have time this week (just wrapped a 14 hour day and still have email to do) but probably next week I could upload something.
Another possible explanation is the bit rate of the target render. He said 4K H.264, but didn't say VBR or CBR and at what rate? For example, 50Mbps CBR might render at a different speed than 25Mbps VBR. This aspect of the comparison was never discussed.
Click to expand...

Doug,
So I went back and looked at it in more detail. The default 'share' setting in FCPX was a 'multi pass' H.264 encode. I switched it to 'single pass'
and will reveal those times shortly. But as for the external drive thing....I had always used external drives and still have a couple thunderbolt
3 external SSD's which have read and write speeds around 1000 MB/s. But the internal SSD has read and write speeds around 6000 MB/s.
It's the fastest way to edit, encode, render or whatever. It's why I took the hit and bought one with a 2TB drive (and why they sell them with internal
drives up to 8TB.) The internal drives are spendy but 1000% the way to go if you want pure speed and can spring for it.

Test
1 minute render of FX6 footage (4K/UHD 23.98fps XAVC-I 10 bit 4.2.2) rendered to (4K/UHD 23.98 H.264 40 Mbps single pass)
Computer: M1 Max 14inch MacBook Pro. 10 Core CPU 32 Core GPU 64GB unified memory. 2TB internal SSD

Speed using external thunderbolt 3 SSD: 17.84 seconds
Speed using internal SSD: 16.2 seconds

So my takeaway is......it's important to check every little thing to make sure you are testing 'like for like'.
Multi Pass encodes will take longer by their nature and until I dug into the encode to check the bitrate and
other particulars, I didn't even realize I was encoding multi pass on this test.
 
Last edited:
Where do you see 'multi-pass' in FCP? (It is indeed the default setting for H.264 in Compressor which FCP uses its engine for everything so I don't doubt it, but just wondering where it says multi-pass in the NLE, unless you just analyzed the final file?)
 
NorBro said:
Where do you see 'multi-pass' in FCP? (It is indeed the default setting for H.264 in Compressor which FCP uses its engine for everything so I don't doubt it, but just wondering where it says multi-pass in the NLE, unless you just analyzed the final file?)
Click to expand...

NorBro,
Sorry, you are correct that it doesn't show up in FCPX necessarily (at least I didn't see it there.) I opened up Compressor to check the bitrate of
the default setting used in my version of FCP X for 4k export. Mainly because Doug had mentioned how a different bitrate (as well as CBR vs VBR)
could make a difference in the export times. And as I scrolled through the settings, I saw that 'multi pass' was checked. So basically, I duplicated
the default, and unchecked 'multi pass' saved it as a preset, and then that new preset shows up in FCPX to use as an export choice. I chose that
new export choice instead of the default one which uses 'multi pass' and got much faster times.
 
ahalpert said:
I know I'm not supposed to admit this, but I rather prefer using mirrorless cameras over FX/FS line. I'm rarely running sound to camera and I can get exposure tools through external monitors. Give me internal ND and I would be set.

Touch tracking through an external monitor -- for this I had been using Imaging Edge Mobile via wifi connection on my smartphone.

BUT I had a game-changing discovery last week. The app Monitor+ can do everything Imagine Edge Mobile can do (except store a list of cameras for quick reconnection - it only stores one). Monitor+ can give you touch tracking. But it can also give you false color, which Imaging Edge Mobile can't do. Which means that I can kick my 5" monitors off my gimbals and solely rely on my phone from now on.

But it gets better. A couple days ago I found out Monitor+ will connect to the camera via USB-C with the camera in "PC Remote" mode. So now you have a wired connection (with the attendant benefits of lower latency and higher reliability) that can do touch tracking, etc. And the connection is very robust in that if you disconnect or power the camera off and then reconnect to the phone, your picture comes back as quickly as it does with an HDMI connection. Obviously, when shooting on gimbal there are benefits to using a wireless connection rather than having a tether tugging on the side of the camera, but with the right cables rigged the right way I think it's going to be a non-issue. I'm debuting the wired solution on my shoot tomorrow
Click to expand...

That's really good to know. I tried Monitor+ few months ago and it looked promising, it just didn't work - image freezing. I have a 2016 iphone SE so that could be the problem. I'll definitely give it another go. I'd even consider buying an Android phone just to use as a monitor, as a wired solution is obviously preferable. How did the shoot go?

I used the FX6 for a news feature job this week and I wondered afterwards if I should have taken the a7Siii instead, mainly for the IBIS. I've definitely moved towards wider FOVs in recent years. On this last job I frequently felt that 24mm (Sony 24-105 f/4) was not wide enough and there don't seem to be any Sony options at the wide end with OSS, it's frustrating. The 16-35 2.8 GM is a great lens but of course, unstabilised.

Having no stabilisation assistance at all is not ideal for this kind of work. There is a limited budget of time and energy and getting useable handheld with an unstabilised lens (particularly for walking shots) is not a wise investment of either. The best I can hope for is that it isn't a problem for the producer. No one is ever going to say "Hey, we really liked the way the image was a bit shaky, can you do that again next time?" I'm actually pretty good at stable, walking handheld but it's a very distinctive look and not one that's necessarily welcome.

I've got an old Canon 16-35 f/4 IS with a Metabones adapter. I'm going to test it tomorrow with the FX6 and compare it to the a7Siii with the Sony 16-35 GM (and IBIS) as on this cannot go. I might end up going all in with the a7Siii for some jobs, especially if the monitoring situation is made more acceptable by Monitor+.
 
OK, I take some of the above back (in case anyone is following this [laugh with tears emoji]) after testing.

Walking shots: IBIS (standard) is not as useful in this situation as I thought. I think the side handle on the FX6 contributes more to stability than IBIS standard does to the a7Siii. IBIS (active - with crop) is dramatically smoother but then it becomes more of a psuedo gimbal shot which is not usually what I'm aiming for.

Static shots (with cinesaddle): IBIS does make a difference here and it's quite subtle. It still looks like handheld but it looks like handheld with a heavier camera. On the other hand, it's not static shots I was so concerned about.

So, I'll carry on moaning about the lack of stabilisation on the FX6 / Sony lenses but I won't be switching to the Canon 16-35 f/4 on the FX6.
 
Andy9 said:
That's really good to know. I tried Monitor+ few months ago and it looked promising, it just didn't work - image freezing. I have a 2016 iphone SE so that could be the problem. I'll definitely give it another go. I'd even consider buying an Android phone just to use as a monitor, as a wired solution is obviously preferable. How did the shoot go?

I used the FX6 for a news feature job this week and I wondered afterwards if I should have taken the a7Siii instead, mainly for the IBIS. I've definitely moved towards wider FOVs in recent years. On this last job I frequently felt that 24mm (Sony 24-105 f/4) was not wide enough and there don't seem to be any Sony options at the wide end with OSS, it's frustrating. The 16-35 2.8 GM is a great lens but of course, unstabilised.

Having no stabilisation assistance at all is not ideal for this kind of work. There is a limited budget of time and energy and getting useable handheld with an unstabilised lens (particularly for walking shots) is not a wise investment of either. The best I can hope for is that it isn't a problem for the producer. No one is ever going to say "Hey, we really liked the way the image was a bit shaky, can you do that again next time?" I'm actually pretty good at stable, walking handheld but it's a very distinctive look and not one that's necessarily welcome.

I've got an old Canon 16-35 f/4 IS with a Metabones adapter. I'm going to test it tomorrow with the FX6 and compare it to the a7Siii with the Sony 16-35 GM (and IBIS) as on this cannot go. I might end up going all in with the a7Siii for some jobs, especially if the monitoring situation is made more acceptable by Monitor+.
Click to expand...

Monitor+ has been rock-solid for me. And that was on the Galaxy A20, which is a low-end phone a few years old. My 2-year old smashed that phone a week ago and I've replaced it with a Galaxy S22+ which is more expensive and claims to have a 1750-nit screen. If you want to buy a cheap android for such purposes, I might recommend the Galaxy A22 or A23, but I'm not very knowledgable about phones.

I did have one issue with Monitor+ which turned out to be user error. On my first shoot with it, after a half hour of smooth monitoring/control, the app began to crash every time I hit the record button. I had to stop using it, but later I realized that the default setting is to record the liveview feed to the phone. My phone's hard drive had filled up after a half hour, and that was why the app was crashing when I hit record. I cleared those files and turned the record setting off, and since then it's been a dream to use.

It's kind of what I've always wished were available: remote monitoring and control, especially to a display of your own choosing. I think Portkeys makes some solutions that can control a Komodo or Z Cam or something. But the reviews I read suggested it was a sluggish, imperfect interface. Monitor+ is smooth, responsive, well laid-out and actually improves on native camera controls in a couple ways. For example, when you open the ISO menu, it offers several presets including low and high base ISO (which are labeled NATIVE) alongside a full range selector.

re: IBIS, i feel like it tends to work less well with wide-angles than telephotos. But Sony also has the worst IBIS in the business

What about the Sony vario-tessar 16-35 f/4 OSS? Or maybe the addition of an extra grip on the a7siii could get the handling to where you have with the FX6, and the IBIS could take it over the top. or maybe the addition of a roll-axis stabilizer like the letus helix. I find roll shake to be more bothersome as a viewer than pan or tilt shake. I will say that I have a client who handed me an indefinite recurring assignment because I'm an all-day gimbal operator and they weren't happy with the handheld look their previous hire was providing, but obviously gimbal isn't for everything

EDIT: just saw on the Komodo thread that you can control a Komodo with a smartphone via wired connection now, and the user seems happy with it. Using RED's control app. Tangential, but i stand corrected
 
Last edited:
You must log in or register to reply here.
Back
Top