Hi there

Bit of an odd question but..

Shot a test HDX900 1080i/25p half-shutter Nano 50mbs MOV. lastest firmware..
A white lamp shade.. against white silk curtain,.. was giving me a red/pink line along the shade edge..against the white curtain..

Everything else fine..

Viewing on Mac book.. Not pro.. Quicktime..

Any idea,s ?

Thanks

Hi Robin:
There are sooo many things which could be coloring the image you are seeing on your MAC laptop that it's sad. Do you own, or have access to an HD monitor ? Check playback on one of those first, then you will know for sure if something is amiss.

Hi Mark

Thanks for your time.. yes not so apparent on HD monitor.. Pana 900A .. see it much more on Mac book

Re shot just now .. same scene.. and seemed to have gone away.. the curtains have a very fine grain pattern in the material.. I think it depends on the folds.. some sort of optical thing I think.. ??

Thanks again

Dear Robyn,

"Red/Pink" may be magenta, and it may be chromatic aberration.

Is there a slight green line on the other side of the shade?

Is the "Red/Pink" line towards the edge of the image as opposed to the center?

Which lens are you using?

If you have a good HD-SDI or HDMI monitor, is the "Red/Pink" line visible when playing back the footage directly from the nanoFlash?

Hi Dan

Yes I would say magenta.. other side of shade is against a green curtain ! so not sure about that.. :)

And yes the line is right on the edge of the shade..

Yes good glass.. HD Fujinon.. 4.5 wide angle zoom.. much less pronounced on Pana HD 900A on play back from Nano.. but still there a tiny bit..

Iam sure its to do with the very fine grain in the curtain material ..chromatic aberration.. what causes it to be magenta?

Thanks for your time.. again !

Dear Robyn,

Technically it may not be chromatic aberration.

With certain cameras, under some contidtion, when going from one color to another, there is sometime a green line on the left and a magenta line on the right.

Are you recording at 50 Mbps or above, thus 4:2:2?

Hi Dan

Looks like lateral aberration .. from a quick look on the net.. edge against a very bright background.. actually magenta strip was on the left of the shade.. not an aperture settings or even a shot I think I would ever do on a real shoot.

Shooting 50Mbs Long GOP MOV.. had the lens a while and no complaints under many different circumstances .. but nearly all of that,99% has been shooting DVCpro HD.. not with the nano.. so in theory 100Mbs.. maybe just something to watch for with that lens?

Dear Robyn,

If you can recreate the shot, please do so while recording to tape also.

When you run the tests, record some at both 50 Mbps and 100 Mbps Long-GOP.

I assume that you recorded the scene originally in 1080i50.

The DVCPro HD is far less resolution than the nanoFlash.

DVCPro HD reduces the resolution before compression, the nanoFlash does not.

nanoFlash
1280 x 720 for 720p
1920 x 1080 for 1080

DVCPro HD
960 x 720 pixels for 720p,
1280 x 1080 for 1080/59.94
1440 x 1080 for 1080/50i

(Specifications for DVCPro are from Wikipedia)

DVCPro HD may be 100 Mbps at certain frame rates, but drops down dramatically at other frame rates, such as 24p. The nanoFlash always gives you the requested bit-rate regardless of the frame rate.

Hi Dan

Ah ok interesting.. what Mbps does DVCpro go down to at 25p...? so would higher resolution give more chromatic lateral aberration ? or no effect?

Thanks again!

Dear Roby,

Wikipedia sates that at 24 frames per second DVCPro goes down to 40 Mbps.
It does not state the bit rate at 25 frames per second.

We do not know if this is chromatic aberration or not.

A carefully controlled test, with your camera, your lens, running both tape and the nanoFlash at various bit rates would be very desirable.

Remember, the nanoFlash is just more faithfully recording the images from the camera.

Hi Dan

Interesting I didnt know that the 100mbps fell to 40 at 24p.. presumably 25p isnt much more..

So shooting at 25p with the nano at 50mbps is actually a higher bit rate than going to tape..???

Looks to be purple fringing.. thing.. cant think what else it is.. but yes could shoot some more tests.. as long as I have the curtains and the sun still shines :)

Thanks again

Dear Robyn,

The difference is far greater than the 10 Mbps between 40 Mbps and 50 Mbps.

First, the nanoFlash records the full resolution provided to us.

Second, the nanoFlash, at 50 Mbps, is Long-GOP, and much more efficient than I-Frame Only at these bit-rates.

Good luck in your tests.

I highly recommend that, if you do the tests, example the images carefully on a good monitor, before or while you are recording. If you can see the fringing, the nanoFlash will record it.

Hello Dan,
While you on the subject of Long Gop Vs I-Frame do you have a comparison chart that you could post for the various bit rates.

Many thanks

Dear Robyn,
nanoFlash
1280 x 720 for 720p
1920 x 1080 for 1080

DVCPro HD
960 x 720 pixels for 720p,
1280 x 1080 for 1080/59.94
1440 x 1080 for 1080/50i

(Specifications for DVCPro are from Wikipedia)

DVCPro HD may be 100 Mbps at certain frame rates, but drops down dramatically at other frame rates, such as 24p. The nanoFlash always gives you the requested bit-rate regardless of the frame rate....Hey Dan: Oh crap ! I didn't know that ! I assume this is also an accurate statement covering the Flash XDR ? 1280 x 1080 = 1080/59.94i = SUCKS ! Less pixels at *any* data rate is still less pixels.

Dear Mark,

Sorry, but I do not understand your comments.

The Flash XDR and nanoFlash always records full raster.

Full Raster for 720p = 1280 x 720 (1280 horizontal pixels, 720 vertical pixels).

Full Raster for 1080 = 1920 x 1080

The DVCPro HD is the codec that throws away resolution, then encodes the video stream.

The nanoFlash does not throw away any resolution.

Dear Lance,

Here is a chart that we created from our tests of the nanoFlash using the Video Clarity system.

Note: A lower number equals a better image based on Video Clarity's modeling of human perception of image quality.

Lower numbers plot higher on the chart, thus as the line gets closer to the top of the chart, the footage is higher in quality according to Video Clarity.

Our tests were performed using a high quality test file of a football game and this is not representative of all types of footage.

Just click on the chart to enlarge it, otherwise it i not large enough to actually read.

Dear Mark,

Sorry, but I do not understand your comments.

The Flash XDR and nanoFlash always records full raster.

Full Raster for 720p = 1280 x 720 (1280 horizontal pixels, 720 vertical pixels).

Full Raster for 1080 = 1920 x 1080

The DVCPro HD is the codec that throws away resolution, then encodes the video stream.

The nanoFlash does not throw away any resolution. Hi Dan: Wasn't I clear enough ? I was commenting on how surprised I was to learn that DVCPRO 100 throws away resolution. I was asking if what you wrote about the Nano always giving the Full Raster Maximum Resolution also was the case for the XDR ? Sorry you misunderstood my post. :-(

Dear Mark,

Yes, I completely misunderstood your post.

You wrote "I assume this is also an accurate statement covering the Flash XDR?"

So I thought the following statement applied to the Flash XDR. But now I know that it didn't.

Yes, both the Flash XDR and nanoFlash record full raster and apply the full bit-rate (per frame) to each frame recorded.

For example, if one shoots 100 Mbps and 1080i50.94 which is actually 29.97 frames per second, then the bit-rate per frame is 100 Mbps / 29.97.

If one shoots 100 Mbps and 1080p23.976, then the bit-rate per frame is 100 Mbps / 23.976, a slightly higher bit-rate per frame.

On the other hand, DVCPro HD, dramatically reduces the bit-rate, down to 40 Mbps (according to Wikipedia), when shooting at 24p.

Hi Dan:
Specifically, I knew about XDR & Nano giving full raster, but what surprised me was the DVCPRO 100 spec was throwing away resolution to make 1080i 59.94 ! Panasonic outta know better than that !

Mark, as I understand it, this all goes back to the Varicam tape cameras, which ALWAYS recorded 60 frames per second. If you were using a Varicam for 24P, it would still record all 60, but only use 24 of the frames. If you have the overall recording being at 100 Mbps, then at 24 you are actually using (24/60)*100Mbps = 40Mbps.
Therefore when you record on a P2 card at 24PN, you are also only recording at 40Mbps, and thereby extending your record time significantly, while keeping the same compression ratio. (And when you record at 24P using the full 100Mbps, 60% of your data is going into unused frames.
It's actually amazing that at the time the Varicam seemed to look so good; but as time has marched forward the Long-GOP compression system has proved far more efficient than the old DVCproHD, and of course DVCproHD in my opinion has really never been up to the task of recording full raster 1920x1080.

Hi Dave:
The thing which caught my attention for a big wow- was how Dan was pointing out how the DVCPro 100 format actually threw away temporal resolution to accomplish 1080 59.94 i. It seems like an unreasonable compromise to drop from full raster HD to 960 x 1080 Ugghhhhhh ! Yuk ! Who cares about compression rate at this point, when you throw away pixels out of the raster.

Dear Lance,

Here is a chart that we created from our tests of the nanoFlash using the Video Clarity system.

Note: A lower number equals a better image based on Video Clarity's modeling of human perception of image quality.

Lower numbers plot higher on the chart, thus as the line gets closer to the top of the chart, the footage is higher in quality according to Video Clarity.

Our tests were performed using a high quality test file of a football game and this is not representative of all types of footage.

Just click on the chart to enlarge it, otherwise it i not large enough to actually read.Hey Dan: I was checking out the chart you posted, and I couldn't help but notice a significant gain in quality between 100 Mbps and 280 Mbps in both the Chrominance and Luminance signals.