JVCs, D90, Sony HD1000 side by side tests at DVinfo.net

Ted Ramasola · January 14th, 2009, 10:02 AM

My brother came home for the holidays and showed me his D90, I was excited to see the footage that he DP'd for an indie feature using the video capable Nikon DSLR, for kicks we decided to do a shootout last new year's eve in our home studio.

Since the D90 shoots 720 24P we decided to see how it fared against whats probably the best affordable 720P cam out there, the JVC HDxx cams.

To liven it up, we decided to throw in 2 jvcs, HD100 with stock fujinon 16x, and HD200 with a 35mm adapter.
AND, since the D90 is a CMOS cam, we added a SONY single CMOS HD cam, the shoulder mountable HD1000 in the mix.

The reason we used two JVCs is simply to hasten our tests using stock and 35mm adapter so later on I'll just refer to the two as JVC cam and simply state what lens was used.

First we did a res chart test, then we did a setup to test it color and bokeh handling.

So here goes, lets start with the set up. Since our Christmas tree was still up it would be good to test its lights for bokeh.
We set up a lifesize bust 9ft away from the tree, our cam position 4.5 ft away from the bust.

And below are the results. We didnt use the sony in this test.
We used a very fast mid format 80mm f1.9 with a nikon adapter on the D90 and JVC with 35mm adapter.

A2 -D90 with mamiya at f1.9
A1 -JVC with adapter and mamiya at f1.9
A3 -JVC with fujinon at f1.4 (this is not ideal, note the lateral CA)

While we know the stock fujinon is not at its best at wide, and while we know CA disappears at f2.8 we decided to push it to see if the stocklens+1/3chips can "theoretically" produce a pleasant shallow DOF by themselves versus an adapter or the D90.

Next post, the charts.

William Hohauser · January 14th, 2009, 11:56 AM

What compression was the Nikon's video shot with? The image is very posterized compared to the JVC although that might look nice if it doesn't behave oddly with motion. It's a cruder image but sort of painterly.

What provisions does the Nikon have for sound recording?

Ted Ramasola · January 14th, 2009, 12:00 PM

Ok, heres the chart test.

We decided to use a 50mm f1.4 on the D90 and JVC with 35adapter.
We set up the 50mm on the D90 at f2.8.

On the JVC w 35adapter I used the same 50mm at f2.

On the JVC w Fujinon 16x stock we shot at f2.8

We had initial problems getting exposure on the D90 since we lit the charts with a flo softbox. The overhead flourescents that ran on ballasts seems to introduce lines at certain exposures. When we finally got a clear screen we locked exposure and took the shot.

C1- D90 50mmf1.4 @ f2.8
(notes: viewing the shot on a large CRT monitor it seems ok, but we noticed that the recorded image is significantly deteriorated when we played it back on the NLE.
Tell tale mosquito artifacts appear around the numbers on the chart.
The D90 seem to resolve around 460 to 500 on the horizontal lines. BUT the sensor seems to introduce a bluish/pinkish/yellowish mishmash on the moire areas especially on the thin black horizontal lines. This is the reason we pulled out the sony cam and included it in our test.)

C4- Sony HD1000
(notes: we included this 1080i single CMOS cam to see if the mishmash of colors on the D90 chart is a product of the CMOS chip or the Codec.
Looking closely at the trumpets and the horizontal lines at the 700-900 range of the Sony chart we saw SIMILAR rainbowish colors patterns in areas that are supposedy thin black lines.
The D90 color "mishmash" on the thin moire prone areas of the chart are more pronounced which rules out the D90 cam if color accuracy is needed.
Is this CMOS issue or the MJPEG codec that aggravated it? I Don't know. I've used the MJPEG codec during my Matrox Digisuite editing days and didn't notice this.)

C2- JVC with adapter 50mmf1.4 @ f2
(notes: I decided to open up a notch since at f2.8 the achilles heel of my 35 adapter, the GG pattern, starts to appear very slightly as a swirl. I noted no color was introduced, the thin black lines that the cam+adapter fail to resolve appear as gray which is normal, the extreme edges though show slight yellow & blue fringe due to the addition of an achromat in an adapter set up.)

C3- JVC Fujinon 16x @f2.8
(notes: slight red/green fringe CA at the extreme sides, which is the nature of the 16x. Otherwise the Chart showed true circles and diagonal lines.)

Ted Ramasola · January 14th, 2009, 12:09 PM

Quote:

Originally Posted by William Hohauser

What compression was the Nikon's video shot with? The image is very posterized compared to the JVC although that might look nice if it doesn't behave oddly with motion. It's a cruder image but sort of painterly.

What provisions does the Nikon have for sound recording?

William,

The Nikon uses m-jpeg with bitrate under 16mb/s.
Audio is a small onboard mic which records 16bit MONO audio.
There is no provision for external mic.

It behaves beautifully under natural light, very difficult under flos. We tested using MH lights and the horizontal running lines on screen worsened.

Ethan Cooper · January 14th, 2009, 04:42 PM

Poor D90 looks bad in those chart tests. I've often thought it was the codec and bitrate holding it back, but maybe it's even more than that.

William Hohauser · January 14th, 2009, 04:48 PM

Take photo of the chart with the D90. That will lead to some answers. Are there different video compression rates available?

Ethan Cooper · January 14th, 2009, 04:57 PM

Quote:

Originally Posted by William Hohauser

Take photo of the chart with the D90. That will lead to some answers. Are there different video compression rates available?

Nope. It's a $1000 still photo camera that happens to shoot video, they don't give you many options.

You can shoot standard def video on it.... hmmm.. is that the same bitrate too? Might look way better if it is.

Ted Ramasola · January 20th, 2009, 12:53 AM

Quote:

Originally Posted by William Hohauser

Take photo of the chart with the D90. That will lead to some answers. Are there different video compression rates available?

William,

I know the D90 can shoot also in SD, as to this will give better bitrate, I don't know.

What we know is shooting at HD, editing in HD and delivering in SD gives acceptable results after we burned the material onto DVD.

Ted

Alex Humphrey · January 20th, 2009, 09:50 AM

which 35mm adapter were you using? Also was sharpening on Normal on the JVC's? How about -7 or Min and what about the Sony?

Jerry Porter · January 20th, 2009, 10:40 AM

I have a D90 also and as soon as you put motion in front of it or you make it move the wheels sort of come off. As soon as I got mine (not for video use) I took it out and tried it. As long as nothing is moving it's Ok

Ted Ramasola · January 20th, 2009, 12:47 PM

Quote:

Originally Posted by Alex Humphrey

which 35mm adapter were you using? Also was sharpening on Normal on the JVC's? How about -7 or Min and what about the Sony?

Alex,

I used an spinning GG adapter I made myself. It uses an SGPRO achromat. the GG is just a CD case

The detail setting on the JVC is at normal. Blackstretch at 5.

Sony sharpness is at 6 on a scale of 1 to 8.

The D90 had sharpening at 0 and contrast at -3 for better latitude.

Let me just add as a side note, since you asked about 35adapters, that the JVC or a videocam with a capable 35 adapter will still give a wider FOV than the D90 using the same lens due to its crop factor.

On our tests,

to acquire the approximately the same framing on the bust and christmas lights, the
JVC with 35mm using a 50mm f1.4 lens was measured at 119.5 cm from the eyes of the bust to the focus plane marker (GG plane) of the adapter.

The D90 with the same 50mm had to be moved back to 149cm to get the same framing.(but note that the image posted is still cropped closer as compared to the JVC framing, this means that we still would have needed to move back further to achieve the exact framing composition)

Ted

William Hohauser · January 20th, 2009, 02:14 PM

It would be quite interesting if a D90 owner could post a few moving clips to Vimeo. I am interested in the camera for shooting abstract stills with limited motion in the frame (leaves moving, water, etc.).

Thanks for the information!

Robert Altman · January 20th, 2009, 10:12 PM

Actually using custom curves, proper exposure control, and a tripod/monopod/steadicam the d90 can produce stunningly beautiful and filmic images. It needs proper care--the soccer mom crowd will be very disappointed!

Just check out vimeo--do a search for d90--to see the best (and worst) of the d90. It can handle movement well--but it does not tolerate being hand held at all!

Using custom curves one can improve the jpeg compression and create an incredible dynamic range that blows the garden variety 1/3" and 2/3" video cameras away.

Ted Ramasola · January 20th, 2009, 11:33 PM

Quote:

Originally Posted by Robert Altman

Actually using custom curves, proper exposure control, and a tripod/monopod/steadicam the d90 can produce stunningly beautiful and filmic images. It needs proper care--the soccer mom crowd will be very disappointed!

Just check out vimeo--do a search for d90--to see the best (and worst) of the d90. It can handle movement well--but it does not tolerate being hand held at all!

Using custom curves one can improve the jpeg compression and create an incredible dynamic range that blows the garden variety 1/3" and 2/3" video cameras away.

Robert,

There is no argument that that it can, In fact as I said, after the tests we did, my brother continued to use it in an indie film and BTS shots, both still and video.

The fact is we were impressed by its low light performance compared to a lot of existing cams,even with the cams we tested it with.

I wish I can show it here, the indie film they are doing is even a horror/thriller and not just drama with controlled movement.

BUT, in a business like what we do, its just one of the tools out there and right now its not something I would take out of the toolbox for most paid work. It has specialized uses.

In our tests we mounted the D90 Not just on a tripod but even dressed it up with a matteBOX and rails and even tested it on a mini skater for dolly shots.

We tested various curves and used a low contrast setting to avoid crushing the blacks.

It can create great images. BUT contrary to what you posted elsewhere that people dont shoot charts but shoot people, hey we do shoot a lot of things, we just tested our tools and want to know how sharp our knives are so we'll know which knife to use when cutting our meat.

I take offense to what you said there since it seems to patronize some honest testing that we are doing in the hope of finding out the performance of these tools to better show people how they behave. NOT to prevent them from buying them but to help them avoid these issues when doing their work.

Take for instance the JVC. It came out and I read here everything about CA and SSEs. Then people came out with solutions; Avoiding long shots and lowering opening at or below 2.8. Avoiding high gain setting etc.
Knowing these issues and how to avoid them gave me the confidence to purchase 2 units.

As I said these cameras are just tools, we just need to know which is right for the job.
The story/subject/content/finished project is how you get judged at what you did and not what camera you use.

Respectfully,

Ted

Robert Altman · January 21st, 2009, 07:56 AM

I am sorry for any offense with my posts--I was simply trying to keep people's minds open about the d90 as a useful tool. Your testing is very valuable. My point was that a camera's performance cannot be completely defined by these tests--much as a student's performance cannot be completely captured by an SAT score.

I fully realize the weaknesses in the d90--your tests highlight some of these. However, upon viewing video done with the camera under proper control the values of the d90 image are also apparent. I have always looked at the d90 as a great dSLR, and a stop-gap device to capture video with a more 35mm filmic look than my 3-chip HDV camera. I am sure that in a year or two I will be buying a d90 replacement that will have addressed many of the problems--and my complement of lenses and accessories will come with me. I would not invest in the Canon 5D Mk II because I believe that at its current price I would have a harder time viewing it in the same stop-gap role--and it too has many limitations.

A few years from now the merging of dSLR and video will be much more mature. Nikon I think is in the best position to lead the way, as it has no current video business to worry about cannibalizing.

In the mean time--please continue your testing, and keeping the rest of us from having to reinvent the wheel!

January 14th, 2009, 10:02 AM	#1
Ted Ramasola Trustee Join Date: May 2005 Location: Allen, Texas U.S.A Posts: 1,117	JVCs, D90, Sony HD1000 side by side tests My brother came home for the holidays and showed me his D90, I was excited to see the footage that he DP'd for an indie feature using the video capable Nikon DSLR, for kicks we decided to do a shootout last new year's eve in our home studio. Since the D90 shoots 720 24P we decided to see how it fared against whats probably the best affordable 720P cam out there, the JVC HDxx cams. To liven it up, we decided to throw in 2 jvcs, HD100 with stock fujinon 16x, and HD200 with a 35mm adapter. AND, since the D90 is a CMOS cam, we added a SONY single CMOS HD cam, the shoulder mountable HD1000 in the mix. The reason we used two JVCs is simply to hasten our tests using stock and 35mm adapter so later on I'll just refer to the two as JVC cam and simply state what lens was used. First we did a res chart test, then we did a setup to test it color and bokeh handling. So here goes, lets start with the set up. Since our Christmas tree was still up it would be good to test its lights for bokeh. We set up a lifesize bust 9ft away from the tree, our cam position 4.5 ft away from the bust. And below are the results. We didnt use the sony in this test. We used a very fast mid format 80mm f1.9 with a nikon adapter on the D90 and JVC with 35mm adapter. A2 -D90 with mamiya at f1.9 A1 -JVC with adapter and mamiya at f1.9 A3 -JVC with fujinon at f1.4 (this is not ideal, note the lateral CA) While we know the stock fujinon is not at its best at wide, and while we know CA disappears at f2.8 we decided to push it to see if the stocklens+1/3chips can "theoretically" produce a pleasant shallow DOF by themselves versus an adapter or the D90. Next post, the charts. Attached Thumbnails Last edited by Ted Ramasola; January 14th, 2009 at 10:28 AM. Reason: added text

January 14th, 2009, 11:56 AM	#2
William Hohauser Inner Circle Join Date: Dec 2005 Location: New York City Posts: 2,650	What compression was the Nikon's video shot with? The image is very posterized compared to the JVC although that might look nice if it doesn't behave oddly with motion. It's a cruder image but sort of painterly. What provisions does the Nikon have for sound recording? __________________ William Hohauser - New York City Producer/Edit/Camera/Animation

January 14th, 2009, 12:00 PM	#3
Ted Ramasola Trustee Join Date: May 2005 Location: Allen, Texas U.S.A Posts: 1,117	Charts Ok, heres the chart test. We decided to use a 50mm f1.4 on the D90 and JVC with 35adapter. We set up the 50mm on the D90 at f2.8. On the JVC w 35adapter I used the same 50mm at f2. On the JVC w Fujinon 16x stock we shot at f2.8 We had initial problems getting exposure on the D90 since we lit the charts with a flo softbox. The overhead flourescents that ran on ballasts seems to introduce lines at certain exposures. When we finally got a clear screen we locked exposure and took the shot. C1- D90 50mmf1.4 @ f2.8 (notes: viewing the shot on a large CRT monitor it seems ok, but we noticed that the recorded image is significantly deteriorated when we played it back on the NLE. Tell tale mosquito artifacts appear around the numbers on the chart. The D90 seem to resolve around 460 to 500 on the horizontal lines. BUT the sensor seems to introduce a bluish/pinkish/yellowish mishmash on the moire areas especially on the thin black horizontal lines. This is the reason we pulled out the sony cam and included it in our test.) C4- Sony HD1000 (notes: we included this 1080i single CMOS cam to see if the mishmash of colors on the D90 chart is a product of the CMOS chip or the Codec. Looking closely at the trumpets and the horizontal lines at the 700-900 range of the Sony chart we saw SIMILAR rainbowish colors patterns in areas that are supposedy thin black lines. The D90 color "mishmash" on the thin moire prone areas of the chart are more pronounced which rules out the D90 cam if color accuracy is needed. Is this CMOS issue or the MJPEG codec that aggravated it? I Don't know. I've used the MJPEG codec during my Matrox Digisuite editing days and didn't notice this.) C2- JVC with adapter 50mmf1.4 @ f2 (notes: I decided to open up a notch since at f2.8 the achilles heel of my 35 adapter, the GG pattern, starts to appear very slightly as a swirl. I noted no color was introduced, the thin black lines that the cam+adapter fail to resolve appear as gray which is normal, the extreme edges though show slight yellow & blue fringe due to the addition of an achromat in an adapter set up.) C3- JVC Fujinon 16x @f2.8 (notes: slight red/green fringe CA at the extreme sides, which is the nature of the 16x. Otherwise the Chart showed true circles and diagonal lines.) Attached Thumbnails

January 14th, 2009, 04:42 PM	#5
Ethan Cooper Trustee Join Date: Dec 2003 Location: Houma, La. Posts: 1,400 Images: 5	Poor D90 looks bad in those chart tests. I've often thought it was the codec and bitrate holding it back, but maybe it's even more than that. __________________ ∅ -Ethan Cooper

January 14th, 2009, 04:48 PM	#6
William Hohauser Inner Circle Join Date: Dec 2005 Location: New York City Posts: 2,650	Take photo of the chart with the D90. That will lead to some answers. Are there different video compression rates available? __________________ William Hohauser - New York City Producer/Edit/Camera/Animation

January 20th, 2009, 09:50 AM	#9
Alex Humphrey Major Player Join Date: Mar 2007 Location: california North and South Posts: 642	which 35mm adapter were you using? Also was sharpening on Normal on the JVC's? How about -7 or Min and what about the Sony?

January 20th, 2009, 10:40 AM	#10
Jerry Porter Major Player Join Date: Aug 2005 Location: Satellite Beach, Fl Posts: 784	I have a D90 also and as soon as you put motion in front of it or you make it move the wheels sort of come off. As soon as I got mine (not for video use) I took it out and tried it. As long as nothing is moving it's Ok

January 20th, 2009, 02:14 PM	#12
William Hohauser Inner Circle Join Date: Dec 2005 Location: New York City Posts: 2,650	It would be quite interesting if a D90 owner could post a few moving clips to Vimeo. I am interested in the camera for shooting abstract stills with limited motion in the frame (leaves moving, water, etc.). Thanks for the information! __________________ William Hohauser - New York City Producer/Edit/Camera/Animation

January 20th, 2009, 10:12 PM	#13
Robert Altman Regular Crew Join Date: Jun 2007 Location: New York Posts: 28	Actually using custom curves, proper exposure control, and a tripod/monopod/steadicam the d90 can produce stunningly beautiful and filmic images. It needs proper care--the soccer mom crowd will be very disappointed! Just check out vimeo--do a search for d90--to see the best (and worst) of the d90. It can handle movement well--but it does not tolerate being hand held at all! Using custom curves one can improve the jpeg compression and create an incredible dynamic range that blows the garden variety 1/3" and 2/3" video cameras away.

January 21st, 2009, 07:56 AM	#15
Robert Altman Regular Crew Join Date: Jun 2007 Location: New York Posts: 28	I am sorry for any offense with my posts--I was simply trying to keep people's minds open about the d90 as a useful tool. Your testing is very valuable. My point was that a camera's performance cannot be completely defined by these tests--much as a student's performance cannot be completely captured by an SAT score. I fully realize the weaknesses in the d90--your tests highlight some of these. However, upon viewing video done with the camera under proper control the values of the d90 image are also apparent. I have always looked at the d90 as a great dSLR, and a stop-gap device to capture video with a more 35mm filmic look than my 3-chip HDV camera. I am sure that in a year or two I will be buying a d90 replacement that will have addressed many of the problems--and my complement of lenses and accessories will come with me. I would not invest in the Canon 5D Mk II because I believe that at its current price I would have a harder time viewing it in the same stop-gap role--and it too has many limitations. A few years from now the merging of dSLR and video will be much more mature. Nikon I think is in the best position to lead the way, as it has no current video business to worry about cannibalizing. In the mean time--please continue your testing, and keeping the rest of us from having to reinvent the wheel!