View Full Version : Resolution Comparison


Petr Marusek
January 6th, 2006, 01:01 PM
Canon XL H1 from Shannon's chart http://www.cinemahill.com/hidef/xlh1&f900/images/IMG_0115.jpg
800Hx650V progressive
800Hx800V interlaced from German magazine, not from Shanon

Panasonic HVX-200 from Kaku's test http://home.arcor.de/martin.doppelbauer/ResCharts/0088YP.png
625Hx600V progressive

JVC HD-100 from Martin's chart http://home.arcor.de/martin.doppelbauer/ResCharts/JVC%20HD%20100%20Mittlere%20Blende.jpg
700Hx525V progressive

Sony FX-1 from Martin's chart http://home.arcor.de/martin.doppelbauer/ResCharts/Sony%20HDR-FX1%2025mm%20PP_Off%20f4.png
650Hx775V interlaced

I compensated for some factors, like for instance a chart not filling the screen properly.

The Canon is an absolute winner; combined with Wafan HDD or PC based recorder, 35 mm adapter and good quality relay nad 35 mm lenses, it should be an excellent camera for film production with theatrical release.

The Panasonic appears to have CCD's that employ both vertical and horizontal pixel shift, without full pixel count for the format, in either direction.

Martin Costa
January 6th, 2006, 01:45 PM
Petr, thanks for taking the trouble to do the comparisons, now if I could only make a film which was as amazing as the specs of these new hd cams!

Shannon Rawls
January 6th, 2006, 02:01 PM
What mode are these cameras in? 1080 or 720? You should put that in there so we know. I don't want people thinking the HVX-200 was in 1080 when it was in 720.

- ShannonRawls.com

John Mercer
January 6th, 2006, 02:04 PM
I'm sorry but am I missing something - all the rez charts seem to be different. How exactly have you "...compensated for some factors, like for instance a chart not filling the screen properly." ? The JVC chart particularly is way off filling the frame.

The .png images all seem to be 977 x 550 pixels too - what does this mean and how exactly can we relate that to each cameras native rez?

Petr Marusek
January 6th, 2006, 02:27 PM
http://home.arcor.de/martin.doppelbauer/ResCharts/

http://dvinfo.net/conf/showthread.php?t=57572&page=3

This is what Martin determined. The Panasonic was in 1080p mode.

Canon XLH1 about 800 lph (unconfirmed !) (1080i)
Sony HDR-FX1/Z1 about 700 lph (1080i)
Panasonic HVX200 about 600 lph (1080p)
JVC HD 100 about 500 lph (720p)

John Mercier, I don't know what brouser settings are you using, but mine .png images were close to 1910x1040. Note also that I was the only one who interpreted the HVX 1080p resolution over 600 lines. You need to see a few of these charts and go over them with an engineer to to learn how to interpret them properly.

The main difference between my interpretation and Martin's is that I went in 25 line increments. Another one is that if you study the Sony carefully, you see that the vertical resolution goes to about to 775. The JVC was the hardest to interpret. I also remember that the JVC loses resolution on iris closing more than the Sony does, per previous Italian test.

P.S.

I just looked all 4 of the HVX 1080p resolution charts. The camera will resolve 600 lines horizontally and vertically, but the resolution pattern is different than with other cameras. The individual strips with lines in 100 line increments are not as clean as with other cameras once we get close to the resolution limit, indicating IMHO both vertical and horizontal pixel shift. The diagonal resolution is higher, which again IMHO testifies for both V & H shift.

The resolution performance is similar to JVC HD-100.

The Canon is the only camerera in this group with notably high resolution.

Here is a resolotion shart shot with hand-held JVC HD10: http://home.earthlink.net/~lesd/hd/jvc-HD10-chart.jpg It indicates 600Hx450V.

John Mercer
January 7th, 2006, 03:08 AM
Hi Petr,

"You need to see a few of these charts and go over them with an engineer to to learn how to interpret them properly."

I'm sorry but this is quite an arrogant statement. I have been working in TV and film for over 20 years with both film and video - I know how to interpret a resolution chart. These are camera alignment charts - one a very old 4:3 NTSC tube one and the other a more modern Hi-def one. Mostly other types of charts are used for lens collimation. I have never heard of industry professionals testing cameras in this way - only obsessive prosummers.

Ok so I downloaded the .jpgs and see that they are native resolution - my mistake. But my point remains about framing - the JVC is very inaccurately framed - how can you determine accurately the chart here? How did you compensate for this framing? You would also need to look at the JVC image upscaled to 1920 x 1080 pixels to level all comparisons, preferably with the same chart and framing as the H1. The HVX200 OTOH is some wierd 1882 x 1040 - however slight some scaling is going on.

"The Canon is an absolute winner; combined with Wafan HDD or PC based recorder, 35 mm adapter and good quality relay nad 35 mm lenses, it should be an excellent camera for film production with theatrical release."

Even from these very crude comparisons it is far from clear to me that the Canon is "an absolute winner". As I have mentioned elsewhere the differences between these cameras are minimal and resolution charts are only 1 part of the story.

There is absolutely no evidence that the Canon is THE camera choice for theatrical release. a) it is highly unlikely and will be very rare that any of the footage from these cameras will end up in a theatrically released 35mm print, and b) IF that does happen both the HVX200 and the JVC will be equally suitable choices. Consider that 24f 1080p is about 20% less resolution than 1080i, which is about the same resolution as 720p for one thing.

Petr Marusek
January 7th, 2006, 06:31 AM
John,

Sorry if I sounded arrogant. I did not mean "you" particularly, but "someone in general". The charts are far from perfect. I don't want to start looking at them again. I mentioned that the JVC chart was the hardest to interpret. I don't rememeber which one was it, but one chart was a 4x3; that was OK, since H and V resolution lines and their spacing were identical. The chart was shot so it fills the frame vertically, which was proper.

You can consider the JVC interpretation as not too accurate. There were also some charts done somewhere in Italy that showed that the JVC is losing resolution with iris closing a lot more than the Sony; I already mentioned it. So for determining the JVC resolution accurately you would need more charts for sure.

As to cameras losing resolution when converted to progressive, that is for sure, but if you look at the Canon chart, you can see that Canon is doing something differently. I can only guess that although the chips have a native format resolution, a pixel shift allows higher resolution than 1440x1080 and this higher resolution is probably the starting point for Canon to create their progressive footage. It makes sense, since 1080/24p is available also via HDSDI.

Petr

Shannon Rawls
January 7th, 2006, 09:45 AM
Don't beleive Shannon's Chart. It's incorrect. The Canon will yield a better resolution then what that chart shows. I know Shannon personally. Probably better then anybody here, and he told me to tell everybody not to use that chart as scientific data.

Just passing the message

- Wendell R.

Kristian Indrehus
January 7th, 2006, 09:50 AM
There is absolutely no evidence that the Canon is THE camera choice for theatrical release. a) it is highly unlikely and will be very rare that any of the footage from these cameras will end up in a theatrically released 35mm print, and b) IF that does happen both the HVX200 and the JVC will be equally suitable choices. Consider that 24f 1080p is about 20% less resolution than 1080i, which is about the same resolution as 720p for one thing.

I agree with John, and I just bought the XL-H1. I love the camera but I may go buy the HVX as well. But not because of any higher resolution.
I think we should not be too hung up in these specīs. They all seem to have more then decent resolution. Iīm not saying the discussion isnīt important and that specīs doesnīt count. But lets not get fanatic about them. There are so many other equally important things to look for, like what happens when you start to move these things, and colors etc.

Iīm not into moviemaking for the big screen myself, but I had a part in one many years ago. $4 mill. budget. (Thatīs pretty much as high as it goes around here) It was shot on 35mm but I was very disappointed with the quality when I saw it again a few months ago. Now, I had a guest-part in a more recent movie. It was shot on digibeta and then transferred to film, and that looked quite impressive. Here they say in three years it will all be done in HD. (In Norway that is) That certainly does not mean it will be done on prosumer-camīs. Though it might happen occasionally.

If I had to choose between pro-sumers for the big screen, I would have chosen the one with 1920x1080 SDI uncompressed out and recorded to deck. Wait, thatīs not entirely true. I would have done test transfers of the camīs that I found most pleasing to the eye. I certainly would not choose looking at res.charts.

For my own work, I picked the H1 because itīs so versatile. All the inīs and outīs you need, it can be used as a shoulder cam for short periods without any hassle, decent lens, sdi out for studio recording, tc in/out for multi-camera work, and very important to me; both the look of progressive and interlaced compatibility for TV work. What more can you ask for in one package and in this price range? Dare I say it? My customers are used to seeing big shoulder cams, but they go WOW about this little fellow.

Petr Marusek
January 7th, 2006, 09:57 AM
Don't beleive Shannon's Chart. It's incorrect. The Canon will yield a better resolution then what that chart shows.

Shannon, you've been saying this, implying that the camera will do 800Hx800V in 24F, if I understand right. Does it mean that the German magazine also talks about progressive 800Hx800V resolution?

Since you have the camera, could you possibly shoot the resolution chart in 1080i and 1080p and post some frame grabbs, please.

Pierre Barberis
January 7th, 2006, 10:22 AM
Comparing on different targets pictures in different rezsolutions is jus NOT SERIOUS.
People who do this without following a documented process and describing the production thread they have been using are just an embarrassment, and they lead new candidates to erroneous choices.

Try by yourself and make judgement on Movies !!

Some Moderator should do his job and kill such ( suspect ? ) posts.

Petr Marusek
January 7th, 2006, 10:41 AM
Comparing on different targets pictures in different rezsolutions is jus NOT SERIOUS.
People who do this without following a documented process and describing the production thread they have been using are just an embarrassment, and they lead new candidates to erroneous choices.

Try by yourself and make judgement on Movies !!

Some Moderator should do his job and kill such ( suspect ? ) posts.

Excuse me, but the frame grab resolution was approx. 1920x1080 for all the cameras except the JVC, where the resolution was approx 1280x720, which is OK for 720p camera. Since the cameras all had significantly lower resolution than 1080/720, the resolution of the frame grabs was always more than enough; there was no need to have uniformity beyond what was presented. Where the chart did not fill the screen properly, I compensated for that.

I think that the figures are as accurate as you can get. Another site claimed 600 lines for the Canon and 700-750 for the Panasonic. Maybe they are the ones who need to kill their posts!

Gabriele Sartori
January 7th, 2006, 11:54 AM
Here is a resolotion shart shot with hand-held JVC HD10:

Peter I checked the charts and although is probably a test that we can't consider scientific I agree with your observations. Now a question since I have an HD1. What is the resolution of the HD10 in your opinion? From the chart it seems to me not to different from the other cameras although we all agree the problems of the old JVCs are not about resolution.

Thanks a lot
Gabriele

Petr Marusek
January 7th, 2006, 12:27 PM
Now a question since I have an HD1. What is the resolution of the HD10 in your opinion?
If I understand the question right, the resolution on HD1 and HD10 is 600Hx450V, judging from a hand-held chart shoot. It could be more with a tripod mounted camera, but resolution is not everything.

Ash Greyson
January 7th, 2006, 01:55 PM
For film out resolution is the top priority, that is why you see just about everyone with film dreams choose an F900 over a Varicam



ash =o)

Petr Marusek
January 7th, 2006, 03:58 PM
http://www.dvxuser.com/articles/shoot3/FX1-charts/FX1-CamAlign-CF30.jpg

http://www.dvxuser.com/articles/shoot3/FX1-charts/FX1-CamAlign-60i.jpg

http://www.dvxuser.com/articles/shoot3/XL2-charts/XL2%204x3%20EIA1956.JPG

http://www.dvxuser.com/articles/shoot3/DVX-Charts/DVX-EIA-24p-thin-4x3.jpg

http://www.dvxuser.com/articles/shoot3/

I found this article and these charts. DVX chart vertical result needs to be decreased by 25% for 16x9 letterboxed aspect ratio, since the CCD's are 4:3, unless you are going to shoot in 4x3 aspect ratio or with an anamorphic adapter. Some other minor adjustments need to be done.

I've viewed some of the charts in MS Photo Editor with 400% zoom, despite the fact that my 19" monitor has 1600x1200 resolution and is not even set for this resolution, so the equivaelent resolution increase for the close study was about 600%.

I conformed again the H1 progressive resolution as 800Hx650V. Sony CF30 resolution is lower in both vertical and horizontal compared to the Canon in 24F. Sony CF25 resolution should be identical to CF30.

XL2 resolution is 525Hx525V; DVX is about the same, but in letterboxed 16x9 aspect ratio it becomes 525Hx395V. Both of these cameras are progressive.

Chris Hurd
January 7th, 2006, 09:35 PM
People who do this without following a documented process and describing the production thread they have been using are just an embarrassment, and they lead new candidates to erroneous choices. Try by yourself and make judgement on Movies !! Some Moderator should do his job and kill such ( suspect ? ) posts.I'm inclined to agree with you, but I've recently been too busy dealing with trolls, trolling posts, vulgarity, profanity and other inappropriate mayhem to bother with this one yet. Sorry,

Petr Marusek
January 8th, 2006, 05:08 AM
There is nothing wrrong with analyzing charts from different sources. On the contrary, it is better to have the information confirmed from more than one person. The dvxuser charts indicate similar resolution on the Sony camera as Martin's chart does; that is why I did not repeat the numbers from the dvxuser chart. You only have to compensate for the way the charts were shot and some charts from these various sources were not shot properly, but as long as it's clear, what has to be corrected, it's no problem.

On another forum there are posts that HVX has higher resolution than H1. It is not true. I've looked at the Canon chart at 600% magnification and i stand by the 800Hx650V resolution figures in the progressive mode.

A. J. deLange
January 9th, 2006, 10:23 AM
There's definitely nothing wrong with analyzing charts as long as you keep their limitations in mind and understand that they only convey some of the information necessary - particularly in the way we use them. For example, everyone looks to see where the lines blend or aliasing begins and call that the resolution but no one looks at the amplitude difference between black and white which is also an important part of resolution (this is moving towards measurement of modulation transfer function). For example my analysis of Putora charts for the XL-H1 stock lens and the 3x wide angle show the two to have about the same resolution where they clearly do not in any real world picture (the 3x was reputed to be "soft" in the XL2 days). I'll bet MTF plots of the two would show a clear difference.

It is also important to remember that the ultimate suitability of a particular camera or lens for a particular purpose depends on a subjective judgement that A is better than B irrespective of what res chart or MTF plots may show.

Shannon Rawls
January 9th, 2006, 10:51 AM
It is also important to remember that the ultimate suitability of a particular camera or lens for a particular purpose depends on a subjective judgement that A is better than B irrespective of what res chart or MTF plots may show.

A.J., I totally agree....

But how come these things weren't taken into consideration when the tables were turned early last year when measuring the Sony HVR-Z1U? Back then, resolution chart scores were ALL the nay-sayers pointed at. They sang the "low resolution chart" song so loud, that that's all many people payed attention to when trying to make purchase decisions.

But Now that the HVX-200 "MIGHT" have lower resolution then its competitors, all of a sudden resolution charts are no longer important. Now it's all about intended use and picture quality. What's up with that?

Furthermore, in light of the "ALLEGED" noisy image the HVX-200 produces, HVX lovers are starting to now brainwash people into believing "NOISE IS A GOOD THANG!" and trying to convince themselves (and potential buyers) that having a noisy camera is beneficial to all so you can achieve a more film-like image.

Sadly, people are listening and BELIEVING all this hogwash. *sigh* It's funny & disturbing at the same time.

- ShannonRawls.com

Petr Marusek
January 9th, 2006, 11:17 AM
When you're buying 1080p camera, do you expect it to resolve 100% more lines than Canon XL2 480p camera, or just 10% more, or it somehow does not matter any more? Should 1080p HD camera really resolve less vertical lines than PAL XL2 SD camera? Is something wrong here? If not, explain it to me, please, pleeeeeeeeeease!

Steven Thomas
January 9th, 2006, 11:32 AM
Shannon does bring up a good point.

Where's the benifit for 1080P with resolution that may not even
benefit for 720?

Maybe I'm missing something here.
Please help me understand?

Well, we'll all soon know after the camera comparison test kicks off soon.

Steve

Shannon Rawls
January 9th, 2006, 11:32 AM
When you're buying 1080p camera, do you expect it to resolve 100% more lines than <snip> 480p camera, or just 10% more, or it somehow does not matter any more?
Great question Petr.

Remember, there's more to resolution then measuring it Up & DOWN....let's not forget LEFT & RIGHT as well.

To answer your question, I guess I would expect an NTSC 1080p camera it to perform twice as good as an NTSC 480p camera. Both vertically & horizontally. And in color rendetion for a potential film blow-up quality.

This is one reason why I won't use 720p. That format makes no sense to me. (save your breath on trying to convince me, i've studied & read it all)

So now that we know what I expect of a 1080p camera vs. a 480p camera.......is it actually happening???

Is the XL-H1 twice the horizontal & vertical resolution of the XL2?
Does the picture of the XL-H1 look twice as good as an XL2 when blown-up to the best 35mm film stock you can buy?
Will the XL-H1 make a BLU-RAY/HD-DVD look twice as better then an XL2 make a BLU-RAY/HD-DVD look?
Would the picture of the XL-H1 create a prettier and more "hollywood-like" movie on a regular plain' ole DVD then a XL2 would?
Will my executive producer give me TWICE AS MUCH MONEY to make a movie if I were using a XL-H1 instead of an XL2?

likewise

Is the HVX200 twice the horizontal & vertical resolution of the DVX100?
Does the picture of the HVX200 look twice as good as an DVX100 when blown-up to the best 35mm film stock you can buy?
Will the HVX200 make a BLU-RAY/HD-DVD look twice as better then an DVX100 make a BLU-RAY/HD-DVD look?
Would the picture of the HVX200 create a prettier and more "hollywood-like" movie on a regular plain' ole DVD then a DVX100 would?
Will my executive producer give me TWICE AS MUCH MONEY to make a movie if I were using a HVX200 instead of an DVX100?

I have no idea, let's test it and see. Resolution chart tests are cool...but THAT'S the answers I need to know!

- ShannonRawls.com

Kenji Kodama
January 9th, 2006, 11:48 AM
Canon XLH1 about 800 lph (unconfirmed !) (1080i)
Sony HDR-FX1/Z1 about 700 lph (1080i)
Panasonic HVX200 about 600 lph (1080p)
JVC HD 100 about 500 lph (720p)


And,

Sony HDR-HC1/A1 about 800 lph(1080i) in good light condition(1000lx).
(same as H1!)

But FX1 looks better.

Petr Marusek
January 11th, 2006, 02:41 PM
http://www.pixelmonger.com/hd_assets/cam27V.jpg

This is a resolution chart of Panasonic Varicam, which shoots 720p. It is overall similar to H1 in progressive 1080p (24F). The Panasonic resolves slightly less horizontally, slightly more vertically than the Canon, which resolves 800Hx650V, per chart that was provided by Shannon, who claims that when viewed on a monitor, the resolution was higher. The Canon chart data was adjusted for not shooting the chart in proper size.

H1 1080i resolution (interlaced) is 800Hx800V, per German magazine test that was presented here.

HVX-200 resolution is 625Vx600H @1080p, per tests dome by Kaku from Japan, at various sharpness settings.

Resolution naturally is not everything, but resolution-wise the cameras stack as follows:

1.
Canon @ 1080i, which is the most comon mode, at which great majority of HDTV broadcast is done worldwide.

2.
a/ Varicam in 720p. Varicam shoots 60p and 720/60p is one (minor share) of the HDTV broadcast standards in the US. This broadcast does not exist anywhere else in the world. HDTV broadcast in Europe and Asia (Japan, etc.) is 1080i only. Varicam shoots 720-60p at 100 Mbps; its 24p derived stream is 40 Mbps.

b/ Canon @1080p has overall similar level of resolution as Varicam.

3.
HVX at 1080p. The resolution is significantly less than Varicam at 720p but more than NTSC Canon XL2. I'd say it's in the same class as PAL XL2.

The resolution is based on lines per picture height, which is the standard way of describing it.

Chris Loomis
January 11th, 2006, 02:53 PM
how does the JVC hd100u rate in this profile?

Pete Bauer
January 11th, 2006, 10:00 PM
Wrangler Note:
The thread "Resolution: Panasonic HVX and Varicam, Canon XL-H1" also started by Petr was MERGED with this thread, since it was an extension of the same topic. Merged posts appear in order based on the date/time posted. Just trying to reduce the clutter.