00:00:00
when the Ronin 4D first came out for a
00:00:02
very good reason everyone was so focused
00:00:04
on the Revolutionary gimbal camera combo
00:00:06
DJI introduced for the first time but
00:00:08
there was barely any useful information
00:00:09
about the most important part of the
00:00:11
camera the sensor so I thought maybe
00:00:13
there's not much to say about it fast
00:00:15
forward one year DJI asked me if I would
00:00:17
test out its sensor and as I compared it
00:00:19
against the fx3 and the red V Raptor I
00:00:21
was blown away by what I saw and was
00:00:23
shocked that no one thought of giving
00:00:24
more attention to the bigger elephant in
00:00:26
the room the zenmuse X9 sensor we all
00:00:30
know Sony was the queen of low light
00:00:31
performance for the past couple of years
00:00:32
but I was really surprised by the way
00:00:34
the Ronin 14 not only outperformed the
00:00:36
fx3 in low light but it also played in
00:00:38
the same field as the big players like
00:00:40
the V Raptor in this episode I'll show
00:00:42
you the tests and comparisons that led
00:00:43
me to these very objective conclusions
00:00:45
along with the two major flaws I found
00:00:47
at the 4D you know there's no such thing
00:00:49
as a perfect camera after all and
00:00:50
finally I'll share some very unexpected
00:00:52
discoveries about the native ISO and
00:00:54
dynamic range that were very impressive
00:00:56
and disappointing at the same time quick
00:00:57
heads up this episode is extremely
00:00:59
technical I tried to simplify my nerdy
00:01:02
two months of testing but it might still
00:01:03
be intense for many filmmakers so clear
00:01:06
your minds raise yourselves it's gonna
00:01:07
be a fun ride
00:01:09
[Music]
00:01:18
let's start with a quick rundown of the
00:01:20
sensor highlights this is dji's own
00:01:23
zenmuse X9 they have a 6K and an 8K
00:01:25
version they just released I'll only be
00:01:27
talking about the 6K here which I feel
00:01:29
is a sweet spot between 4K and 8K it's a
00:01:32
full-frame sensor with dual native ISO
00:01:34
of 800 and 5000 with 14 plus stops of
00:01:37
dynamic range 9 stops of built-in ND it
00:01:40
can shoot 24 frames for Cinema
00:01:41
Productions and Records pro-res raw
00:01:44
internally along with a bunch of other
00:01:45
prores compressions and h.264 you
00:01:48
probably noticed already how these
00:01:50
features feel like a step up from the
00:01:51
typical mirrorless camera you find
00:01:52
everywhere they're more of a cinema
00:01:54
camera kind of specs looking at the full
00:01:56
frame sensor I was surprised that it
00:01:58
doesn't have any open gate or 16x9 modes
00:02:01
all we have was the DCI 17x9 image area
00:02:04
also a cinema standard so it seems
00:02:06
they're positioning it more as a
00:02:08
dedicated cinema camera or at least a
00:02:10
mirrorless with Cinema quality and
00:02:11
features so this 17x9 image area gives
00:02:14
us a 6K clip maxing out at 48 frames for
00:02:17
higher speeds it crops to a 239 to 1
00:02:20
ratio for up to 60 frames you can still
00:02:22
go even higher if you crop the super 35
00:02:24
and drop to 4K for up to 96 frames and
00:02:27
for the final speed boost you guessed it
00:02:29
your crafted 239 to 1 for up to 120
00:02:32
frames
00:02:33
so I went for a casual test drive with
00:02:35
the camera to get some First Impressions
00:02:37
before I dissected I mainly shot in raw
00:02:39
to get the most honest image the camera
00:02:40
can capture before any internal
00:02:42
Corrections or denoising are applied I
00:02:45
shot scenes with high dynamic range and
00:02:46
different lighting conditions to try to
00:02:48
find the breaking point of the camera
00:02:50
here is a strongly backlit shot I just
00:02:52
applied the official 709 conversion lot
00:02:54
and I remember on the monitor the sky
00:02:56
looked like it was blown out but in post
00:02:58
the waveform showed the Highlight
00:03:00
retained all the details this guy didn't
00:03:02
have much color that day anyway
00:03:04
the log image shows you the latitude it
00:03:06
has in those highlights you can play
00:03:07
with so these were early promising signs
00:03:10
on how the d-lock curve is clinging
00:03:11
pretty well onto the highlights which
00:03:13
I'll confirm with you in another control
00:03:15
test I did but what about the Shadows so
00:03:17
in this clip you can see a lot of good
00:03:19
details in the shadow area this might
00:03:21
seem crushed but it actually has so much
00:03:23
details you can recover if you lift it
00:03:24
up a bit and here's the lock to see
00:03:26
everything the camera captured also
00:03:28
notice how the Shadows look pretty clean
00:03:30
this was the native ISO of 800 but still
00:03:32
the visible noise looked more like film
00:03:34
grain to me so I needed to push the
00:03:36
camera a lot more than that so I went
00:03:38
for an eye test to see if these clean
00:03:39
shadows will translate well in high isos
00:03:41
I was pretty much punishing the camera
00:03:43
in New York's typical bad quality street
00:03:45
lights a dp's nightmare in other words
00:03:48
to me the image looked pretty good
00:03:50
no major color issues even when I
00:03:53
underexposed the skin I was able to
00:03:55
bring it back to life with a minor
00:03:56
mid-cone lift
00:03:57
then up close you can see a very
00:04:00
reasonable amount of noise keep in mind
00:04:02
this is wrong the noisiest format in the
00:04:04
camera since there's no internal
00:04:06
denoising applied here here's how it
00:04:08
would look if I add the denoising pass
00:04:10
so overall I felt the camera performed
00:04:12
super well in low light but that didn't
00:04:14
break the camera still I needed to go
00:04:16
for a more controlled stress test and do
00:04:18
some comparisons with other cameras I'm
00:04:20
more familiar with which leads us to a
00:04:22
very interesting Discovery in our next
00:04:23
topic ISO and low light performance we
00:04:26
know the camera has two ISO ranges first
00:04:28
one starts from 200 all the way up to
00:04:31
4000 where 800 is the first Native ISO
00:04:33
then the second native ISO starts at
00:04:35
5000 then reaching all the way up to 12
00:04:38
800. we all know the logic says the
00:04:40
higher you go away from the Native the
00:04:42
noisier the image gets then it would
00:04:44
reset the noise again at the second
00:04:45
native ISO to reach the highest nose
00:04:47
levels at the maximum of 12 800 with
00:04:50
that logic a 4000 is the furthest
00:04:52
setting from the cleanest 800 native
00:04:54
right before switching to the second
00:04:55
native of 5000 then it should be the
00:04:57
noisiest in that range but that doesn't
00:05:00
seem to be the case
00:05:01
here we have two clips shot at ISO 4000
00:05:04
and 5000 but zooming into the Shadows
00:05:06
the 4000 actually looks a bit cleaner
00:05:08
than 5000. for a final quantifiable
00:05:11
confirmation I measured the noise levels
00:05:13
on my denoiser app picked the same spot
00:05:15
on both and got my confirmation the 4000
00:05:18
gave me 6.6 noise levels and the 5000
00:05:21
gave me 7.8 so I took those results to
00:05:23
my studio to replicate it in a more
00:05:25
controlled environment I measured the
00:05:27
noise level in h264 prores Quad 4 and
00:05:29
prores draw across all ISO settings and
00:05:31
put them in this graph I'll show the
00:05:33
actual Clips in a second but the graph
00:05:35
again confirmed what I just saw ISO 4000
00:05:38
has lower noise than ISO 5000 with a
00:05:40
very dramatic dip in the noisy and
00:05:42
prores 444 and a more subtle one in h264
00:05:45
which might mean that native ISO is 4000
00:05:48
not 5000 but if we isolate the Raw
00:05:51
results there's a subtle dip at 4000 but
00:05:53
probably not enough to qualify as a
00:05:55
second native ISO the whole graph looks
00:05:57
almost linear meaning noise just gets
00:05:59
worse the higher you go on the iso scale
00:06:01
at no point it resets in the
00:06:03
conventional sense we know which could
00:06:05
mean the camera doesn't really have a
00:06:07
second native ISO the other good news
00:06:09
the difference between the lowest and
00:06:11
highest noise levels is pretty small and
00:06:13
this is true in all the codecs I
00:06:14
measured this reminds me of the red V
00:06:16
Raptor I was never scared of pushing it
00:06:18
Beyond 10 000 ISO
00:06:20
if we compare both red and the 4D raw
00:06:23
formats keep in mind red has a huge
00:06:25
Advantage being 8K meaning noise is a
00:06:27
lot smaller in size making it look a lot
00:06:29
cleaner with that in mind red has a
00:06:31
lower noise level of 14 at ISO 800 but
00:06:34
both have the same maximum noise level
00:06:36
of 65.
00:06:38
I need you to remember this number for a
00:06:39
second because I thought if I bring the
00:06:41
Sony fx3 to this comparison being the
00:06:44
most famous low-light monster I was
00:06:45
expecting it to be a tough Contender and
00:06:47
maybe even destroy them both but what I
00:06:50
saw was quite the opposite Sony's noise
00:06:52
performance is by far worse than both
00:06:54
cameras it reached all the way up to 187
00:06:57
at 10 000 ISO then it resets at the
00:06:59
second native ISO of 12800 at 66 noise
00:07:02
score but remember both 4D and the
00:07:05
Raptor max out at 65 at 12 800 ISO
00:07:09
here's the graph of the three cameras
00:07:10
together you can see how the fx3 is way
00:07:12
off the charts that at 10 000 ISO it has
00:07:15
three times the amount of noise of the
00:07:17
4D it even had a marginal defeat in the
00:07:19
Battle of native ISO of 800. the most
00:07:22
interesting is how both red and the
00:07:24
Ronin 4D max out their noise levels at
00:07:26
12 800 which matches Sony's ISO 3200
00:07:29
score
00:07:30
okay let's see how that looks like
00:07:32
shooting side by side first set ISO 3200
00:07:35
you can already see a bit of noise in
00:07:37
the Sony as we pick set peep you can
00:07:39
clearly see the massive difference
00:07:40
between both this test couldn't be more
00:07:43
fair for both because this is six ISO
00:07:46
clicks from both their first Native ISO
00:07:48
of 800. the nosing can clear all this
00:07:50
out of course and they look equally good
00:07:53
now let's max out the 4D at 12 800 and
00:07:56
keep the fx3 at 3200 based on my control
00:07:59
tests they should have similar noise
00:08:01
levels and as expected they do look
00:08:03
similar that's pretty impressive
00:08:05
honestly about the 4D once again if we
00:08:07
apply the nosing on both they look
00:08:09
equally good
00:08:10
let's go back to my control test super
00:08:12
quick this is the scene I shot to test
00:08:14
noise color and highlight roll-off so at
00:08:17
800 both are on their native ISO the fx3
00:08:20
looks a bit softer simply because it's
00:08:21
4K compared to 6K and the 4D jumping to
00:08:24
4000 the last stop before 4D flips into
00:08:27
the second native ISO and fx3 is halfway
00:08:29
through its second ISO honestly the 4D
00:08:32
looks pretty good the fx3 is starting to
00:08:34
struggle already
00:08:35
at 8 000 4D is holding on pretty good
00:08:38
fx3 is falling apart 10 000 fx3 is
00:08:42
pretty much destroyed 12 800 finally fx3
00:08:46
resets at its second native but noise
00:08:48
levels look pretty close between them
00:08:50
for reference here's the red at 12 800
00:08:53
the noise levels look pretty similar
00:08:55
only Red's noise look more monochromatic
00:08:57
which looks better for sure
00:08:59
also here's the 4d's best and worst
00:09:01
noise settings at Native 800 and at the
00:09:04
maximum ISO of 12800 that's what I meant
00:09:06
when I said the difference is not that
00:09:08
bad between the first Native and its
00:09:10
maximum ISO setting
00:09:11
so let's check those high isos in
00:09:14
another real-life comparison to check
00:09:15
once again how things look on skin tones
00:09:17
I shot in Pro is wrong with both cameras
00:09:20
side by side matching their white
00:09:22
balance at 4000 Kelvin to visualize my
00:09:24
test on the iso scale these are the
00:09:26
native isos for the 4D even though I
00:09:28
believe it doesn't have a second native
00:09:30
but let's play along the fx3 has a
00:09:32
confirmed native of 800 and 12 800.
00:09:34
first comparison was at ISO 5000 giving
00:09:37
the advantage to the 4D then I repeated
00:09:39
this at ISO 10000 where it's not native
00:09:41
for either of them but one step before
00:09:43
the 4D Max is out and also one step
00:09:45
before the fx3 flips to the second
00:09:47
native then finally at 12 800 the second
00:09:50
native for the fx3 and the absolute
00:09:52
maximum for the 4D giving the advantage
00:09:55
to the fx3 this time
00:09:57
starting with 5000 ISO you can clearly
00:10:00
see how Sony has a warm yellowish tint
00:10:02
something I've seen a lot in Sony's
00:10:03
color science in general while the 4D
00:10:05
looks more neutral so I'll just set Sony
00:10:07
to 3500 for a more neutral look remember
00:10:10
this is raw so what I'm doing is not
00:10:12
affecting the noise in any way also Sony
00:10:14
looks a bit lifted in the shadows and
00:10:16
has a nicer highlight roll-off on her
00:10:18
face so I'll just get them closer to
00:10:19
each other with a simple curve for both
00:10:21
as for noise from a distance the 4D
00:10:24
looks clean and I can already see some
00:10:26
noise in the fx3 zooming in it's pretty
00:10:29
clear Sony's not doing well
00:10:31
then at 10 000 it's much clearer now how
00:10:34
Sony has a Sandy colored noise texture
00:10:36
which doesn't look good I tried to
00:10:38
denoise it and it looks a lot better but
00:10:40
I struggled to completely remove the
00:10:42
noise honestly you can still see it
00:10:43
lurking in the background in front of
00:10:45
her eyes here while in the 4D it's
00:10:46
pretty clean also a lot of details were
00:10:49
already lost in the fx3 look at the hair
00:10:51
roots for example on both
00:10:53
you literally see the individual hairs
00:10:55
in the 4D but the fx3 pretty much lost
00:10:57
all those details to be fair the 6K
00:11:00
resolution in the 4D is surely helping
00:11:02
but to also be fair the noise in the fx3
00:11:05
was too much to handle I guess that
00:11:07
debunks the myth about bigger pixels
00:11:09
means better light sensitivity
00:11:11
now at 12 800 finally Sony is back to
00:11:13
the comfort of its second native ISO
00:11:15
looking a lot cleaner while the 4D as we
00:11:18
already established still looks good and
00:11:20
holding on even up close they look
00:11:22
pretty similar when denoised I guess
00:11:24
it's clear that the 4D held onto details
00:11:26
a lot better
00:11:28
so bottom line we saw how Sony quickly
00:11:30
reaches the red zone as early as 3200
00:11:33
ISO while the 4D never really reached
00:11:35
any red zones noise just gets worse as
00:11:38
you go up then I discovered it has a
00:11:40
subtle reset at 4000 its unofficial
00:11:42
second native but noise always stayed
00:11:44
within a very reasonable and manageable
00:11:46
level all the way up to 12 800 and it's
00:11:48
funny how at its worst it still looked
00:11:50
virtually identical to Sony's second
00:11:52
native ISO of 12800 of course Sony has
00:11:55
the awesome advantage of going Beyond
00:11:56
12800 and all the way up to over a
00:11:59
hundred thousand ISO and in the extended
00:12:01
mode all the way up to 400 000. now I'm
00:12:04
actually curious to see how the 4D would
00:12:06
perform in those crazy isos DJI should
00:12:08
explore that in future updates so
00:12:11
shooting with a 4D you can feel pretty
00:12:12
confident maxing out your ISO while
00:12:15
shooting with Sony I'd start getting
00:12:16
worried between 4000 and 10 000 which is
00:12:19
a pretty common range we use in night
00:12:20
scenes
00:12:21
so yeah when it comes to low light
00:12:23
performance I think it's fair to
00:12:25
objectively say Ronin 4D destroyed the
00:12:27
fx3 and I can safely say it's my current
00:12:29
Queen of low light right after V Raptor
00:12:31
of course
00:12:32
moving on to Rolling shutter that's
00:12:34
where the DJI lost against the fx3 here
00:12:37
I shot the same vertical line with both
00:12:38
cameras and superimposed them in this
00:12:40
shot fx3 is in blue shooting full frame
00:12:43
at 4K and the 4D is in yellow shooting
00:12:45
full frame as well but in 6K as I whip
00:12:48
pen you can see how the DJI has a much
00:12:51
more skewed angle than the fx3 doing the
00:12:53
math the fx3 is 2.3 times faster than
00:12:56
the 4D pretty impressive honestly this
00:12:58
is on par with Cinema cameras I also
00:13:01
confirmed the score from cined's more
00:13:02
scientific test that led to the same
00:13:04
results so I thought maybe the 6K is was
00:13:06
slowing down the 4D so I dropped the
00:13:08
resolution to 4K and even used the super
00:13:11
35 mode to avoid any internal resizing
00:13:13
that might affect its performance I left
00:13:15
everything the same in the fx3 apart
00:13:17
from cropping it in post to match the
00:13:19
framing and funny enough the 4D was even
00:13:22
slower making the fx3 more than two and
00:13:24
a half times faster than the Ronin 4D
00:13:26
even if I consider a margin of error in
00:13:28
my test it won't be that much off so
00:13:31
this was pretty disappointing thing on
00:13:33
the 4D front I assume DJI designed this
00:13:35
camera with the intention to be used
00:13:37
with the gimbal which would slow down
00:13:38
your panning motions and hide the
00:13:40
effects of the Slow Rolling shutter but
00:13:42
if we take the gimbal out of the
00:13:43
equation then we're left with a pretty
00:13:45
slow sensor readout I guess it's fair to
00:13:47
say Sony destroyed DJI this time and the
00:13:50
rolling shutter battle
00:13:51
next comes dynamic range we have two
00:13:54
important informations you need to look
00:13:55
for first is the usable dynamic range
00:13:58
the sensor can register second is the
00:14:00
dynamic range versus ISO and that's the
00:14:02
most important one no one seems to
00:14:04
really pay attention to
00:14:06
let's start with the usable dynamic
00:14:07
range DJI claims the 4D can achieve over
00:14:10
14 stops on its website but always take
00:14:13
those marketing claim numbers with a
00:14:15
grain of salt simply because this number
00:14:16
reflects all Dynamic ranges including
00:14:19
the unusable ones buried deep under the
00:14:21
noise floor and that's usually one to
00:14:23
three stops above the usable range that
00:14:24
we care about some other brands probably
00:14:27
include some of the partially clipped
00:14:28
range in the highlights as well I'll
00:14:30
explain it and show you an example in a
00:14:32
second this is the synedes MS chart for
00:14:34
the 4D in its native ISO 800 and you can
00:14:37
see how we have 12 usable stops maybe 13
00:14:39
if we denoise it the test readings for a
00:14:42
usable image says 12.4 pretty good by
00:14:45
today's standards anyway
00:14:46
let's compare this to some of the
00:14:48
popular cameras out there here I'm
00:14:50
showing the camera's official claims in
00:14:51
red versus the lab tested usable dynamic
00:14:54
range in blue you can see how Sony and
00:14:56
red both famous for over hyping their
00:14:58
camera specs they exaggerated their
00:15:00
dynamic range claims by 2 and 3 stops
00:15:02
above the usable score so you always
00:15:04
need to take their claims with a bigger
00:15:06
grain of salt than others Blackmagic and
00:15:08
djis were more conservative in their
00:15:10
claims with less than two stops above
00:15:11
the usable then the most honest of them
00:15:13
all is none other than Ari their claim
00:15:16
is only half a stop off I guess you
00:15:18
don't really need to overhype your
00:15:19
camera when you've set the highest
00:15:21
quality bar for everyone else
00:15:23
now I found another small Discovery
00:15:25
between the fx3 and the 4D even though
00:15:27
they share the same usable dynamic range
00:15:29
score the 4D seems to perform better in
00:15:31
my test I shot the same scene and placed
00:15:34
this white card with a light shining at
00:15:36
its bottom to see how both cameras
00:15:37
render light roll off and check the
00:15:39
clipping point on each
00:15:41
side by side both look good but probably
00:15:43
you can't see the fx3 has some partial
00:15:45
clipping here you can see it better in
00:15:47
the waveform
00:15:48
the blue Channel seems okay but if you
00:15:51
look closely at the green and red
00:15:52
channels you can see some clipping
00:15:54
happening and that's what I meant
00:15:55
earlier when I said partially clipped
00:15:58
back to the Ronin 4D all channels have a
00:16:01
nice roll-off curve in the highlights I
00:16:03
thought maybe the lot caused this so I
00:16:04
went back to the lock Clips right out of
00:16:06
the camera but it's way more obvious now
00:16:08
how the fx3 clearly clipped with a solid
00:16:11
Flat Line in both red and green channels
00:16:13
while the 4D has a super smooth slope
00:16:16
across all channels so this dynamic
00:16:18
range probably picked up on the
00:16:20
information from the blue Channel and
00:16:21
considered it as part of the fx3's 12.4
00:16:23
dynamic range score when it really
00:16:25
shouldn't just because of the missing
00:16:27
information from both red and green
00:16:28
channels I could be wrong I wish I have
00:16:31
the proper testing equipment to conduct
00:16:32
my own test and conclude this
00:16:34
here's another interesting Discovery
00:16:36
this is the red V Raptor under the same
00:16:38
conditions lock footage without any Luts
00:16:40
of course there's no clipping whatsoever
00:16:42
but check the waveforms side by side
00:16:44
remember how the 4ds Highlight roll-off
00:16:47
curves look super round and smooth
00:16:48
basically the more it looks like a
00:16:50
quarter round or oval the better and
00:16:52
smoother the Highlight roll-off will be
00:16:54
have a look at the ones from Red you can
00:16:56
see how they're a bit inconsistent not
00:16:58
as smooth and look a bit bumpy in some
00:17:00
points I might be splitting hairs here
00:17:02
but these are some small things that
00:17:04
might make a huge difference in some
00:17:06
cases with specular highlights
00:17:08
now we already established the 4d's
00:17:10
usable dynamic range is not 14 plus
00:17:12
stops but 12.4 instead but the camera
00:17:15
cannot sustain this dynamic range across
00:17:17
all ISO settings that's why if you
00:17:19
scroll further down on the website
00:17:21
you'll find the sensor's dynamic range
00:17:23
performance chart across ISO settings
00:17:24
that's what I referred to earlier as
00:17:26
dynamic range versus ISO this graph will
00:17:29
show you how you might be doing it all
00:17:31
wrong it will prove how increasing your
00:17:33
ISO when it's dark might not be the best
00:17:35
practice to get the best results out of
00:17:37
this sensor let me show you what I mean
00:17:39
so the chart gives you two very useful
00:17:41
informations first is how much dynamic
00:17:44
range you get from each corresponding
00:17:45
ISO setting so ISO 3200 and 12800 have
00:17:48
the lowest dynamic range so you should
00:17:50
avoid them as much as you can
00:17:52
while 400 and 800 give you the maximum
00:17:54
dynamic range possible which we know now
00:17:56
is actually 12.4 but the second crucial
00:17:59
information you get is the distribution
00:18:01
of that dynamic range across highlights
00:18:03
and shadows it's very important to know
00:18:05
because sometimes it's counterintuitive
00:18:07
and might not make any sense to what
00:18:09
we're used to here's a funny example ISO
00:18:12
200 is the worst for daytime or sunny
00:18:14
outdoor because of having only 5 stops
00:18:16
in the highlights but it's best for
00:18:18
shadow details because of its 9 stops
00:18:20
below so if you're shooting a daytime
00:18:22
scene you better not set it to 200 but
00:18:24
instead use a strong ND and shift to one
00:18:26
of these that have the highest highlight
00:18:28
range
00:18:29
same goes to 12800 it will only give you
00:18:31
6 stops in the shadows three full stops
00:18:34
lower than ISO 200 so it's by far the
00:18:36
worst choice for dark scenes I guess you
00:18:39
get my point on why this strut will help
00:18:40
you with some very counter-intuitive
00:18:42
decisions
00:18:43
moving on to the internal ND quality the
00:18:45
Ronin 4D gave us an impressive nine
00:18:48
consecutive stops of full spectrum ND
00:18:50
filters that's the maximum internal ND
00:18:52
system range I've seen so far so these 9
00:18:55
stops are 6 small stops than the
00:18:57
Blackmagic 6K Pro and even the brand new
00:18:59
state-of-the-art Alexa 35 both having
00:19:02
only two four and six stops of ND and
00:19:04
that's also 3 more stops than the Sony
00:19:06
fx6 and the brand new red V Raptor XL
00:19:09
but these have the advantage of using
00:19:11
the more advanced electronic ND system
00:19:13
unlike the Rune in 4d's fully mechanical
00:19:15
one I measured half a second to cycle
00:19:17
between each of these nine stops as they
00:19:19
shift up and down or stack on top of
00:19:21
each other and that's when I had some
00:19:23
concerns since the more you stack
00:19:25
filters and the higher you go with your
00:19:26
ND stops density the higher the risk of
00:19:28
introducing undesired color shifts and
00:19:31
all we care about in any neutral density
00:19:33
filter is to be as color neutral as
00:19:35
possible I learned this lesson when I
00:19:37
discovered in my episode about the
00:19:39
pocket 6K Pro how it shifted colors to
00:19:41
Blue in all its ND stops with an extra
00:19:43
blue shift in nd6
00:19:45
so to test the ronin's and the color
00:19:47
accuracy and consistency I use the
00:19:49
highest quality light I have the solar
00:19:52
30C it has an impressive SSI score of 90
00:19:55
at 3200 Kelvin looking virtually
00:19:57
indistinguishable from real function
00:19:58
light so I shot a color chart first with
00:20:01
a clear ND took a snapshot of its Vector
00:20:04
scope then did the same thing with all
00:20:06
nine ND filters and finally compare them
00:20:08
all to the first one with the clear ND
00:20:10
once again the results were impressive
00:20:12
with a couple of subtle but expected
00:20:14
shifts at higher stops so from stops one
00:20:16
to six they're pretty clean then 7 to 9
00:20:19
had a subtle green shift where the
00:20:21
non-stop has some extra green to show
00:20:23
you what I mean here's how the clear
00:20:25
reference looks like I want you to focus
00:20:27
on the vectorscope center dot here
00:20:28
showing a neutral color when it's dead
00:20:30
center in the graticule now when I shift
00:20:32
to the 7th stop did you notice this
00:20:34
subtle shift towards the green vector
00:20:36
let me toggle between them again
00:20:38
it's pretty subtle but it's there you
00:20:41
can also see it in the chart of course
00:20:43
then if I bring the 9th stop once again
00:20:45
let me toggle between them
00:20:48
you can see a stronger shift this time
00:20:51
so maybe avoid using those ND ranges if
00:20:53
you're shooting a green screen
00:20:54
but again these are very manageable
00:20:56
shifts all within an acceptable range
00:20:58
I'm personally happy and mostly
00:21:00
impressed how they managed to squeeze 9
00:21:02
stops of filtration with such a minimal
00:21:04
amount of color shifts considering how
00:21:06
hard it is to restrict it in such high
00:21:08
and defectors I have many other tests
00:21:11
and results that cover more aspects of
00:21:12
the sensor but that can take another
00:21:14
hour
00:21:15
so to summarize I believe the Zen music
00:21:17
X9 is a high-end sensor even with its
00:21:20
drawbacks it has great potential to
00:21:22
overtake the mirrorless Market I showed
00:21:24
you how it was on par with Hollywood
00:21:25
level Cinema cameras such as the V
00:21:26
Raptor but the sensor's glory moment was
00:21:29
overshadowed by the Revolutionary gimbal
00:21:31
camera hybrid everyone was talking about
00:21:33
which by all means is yet another
00:21:35
awesome once in a decade Innovation by
00:21:37
itself but I feel the sensor still needs
00:21:39
its 15 minutes of fame so here's my two
00:21:42
cents to DJI for their r d of the new
00:21:44
Ronin 4D Mark II create a modular system
00:21:47
that splits the camera into three
00:21:48
interchangeable parts the sensor the
00:21:51
Gimbal and the Brain where the gimbal
00:21:52
part is detachable to let you attach the
00:21:55
sensor directly to the body in a super
00:21:57
sturdy way which will solve the biggest
00:21:59
issue with this camera and that's
00:22:01
finally being able to mount Cinema
00:22:02
lenses or any other oversized heavy
00:22:04
lenses without the need for any sort of
00:22:06
convoluted Contraptions or cages so
00:22:09
until that happens I invite you and all
00:22:11
YouTubers to forget about the gimbal for
00:22:12
a moment and pay more attention to the
00:22:14
equally massive thing about the Ronin 4D
00:22:17
the sensor
00:22:20
[Music]
