XDCAM and low light
 

It appears that the EX is a noticeable step up from the 1/3" HVX200 cam with low light.

HVX200:
GAIN 0 dB
Lens set for equal frame composition as XDCAM EX.
Aperture: Wide open
http://www.slashcam.de/images/texte/...-HVX200Low.jpg

XDCAM EX:
GAIN 0 dB
Lens set for equal frame composition as HVX200.
Aperture: Wide open
http://www.slashcam.de/images/texte/...Sony-LowEx.jpg

Thats Killer...........

I'd almost like to see the test done with a PD170....

Wow.. thats all I can say.

Wow - both shot 24p 1/48th shutter? What iris?

It came from the German test link on the HVX200 vs EX thread.
The only reference I can see was the frame composition and gain.
That's a good question regarding frame rate and shutter speed.
It may be in there ,but I can't read german.
Having said that, I would believe they were set up the same.

In all honesty, and with personal experience with the HVX, I HIGHLY doubt this comparison. Something's just not right about it.

The HVX is about a stop shorter than a DVX100a.

With a lamp shot like this, I truly doubt that the HVX would have THIS MUCH trouble getting a decent image..

I don't know. All I say is, don't count your chickens just yet. Wait until a decent comparison is done.

I hear you; though I can't imagine why this test would of been fudged.
It's not hard to do.
Set up both cams for the same frame composition.
Set bot cams to the same frame and shutter rate.
Set both to 0 dB gain.
Open the apertures on both and watch the show..

Now having said that, 0dB gain is a relative measurement within each cam.
Now, if the test was done correctly, what would the HVX200 look like with +3 to +6dB gain. If it produces the same luminosity and has the same amount on noise, then that's a different story. On the otherhand, by cranking the gain to +3 or +6 on the HVX200 and it matches luminosity of the EX, but the image is noisy.. Well, you know where I'm going.

You make a good point thought. We will have to wait and see. Although, we are hearing it fairs well in low light ( 1/2"), so I would not be surprised that it's better than the HVX200.

You would think the HVX would hold well here since it's true CCD capture prior pixel shift is 960x540. I guess it's possible that the signal processing is a miss on the HVX. I like the HVX, but I've found out your really need to light with this cam or you can end up with some busy backgrounds.

The actual gain factor must have played a dominant role in the output of this comparison - with the EX1 scale starting with the "negative" gain of -3dB, its 0dB does not necessarily correspond to that of other cameras.

Good point Piotr.
I forgot that the EX offered -3dB.
But, if its 0dB reference is free of noise, I see why it maybe labeled as such.
Unless one designed the gain circuit in the cam, It' hard to know how the gain is configured and referenced.

This is true but we are talking about different chip sizes so its simple physics that its going to be considerably better in low light.

Daniel, I never suggested it isn't - but in practical terms, I don't think the difference is as big as the pictures suggest.

Also you must remember that each individual pixel is of a much smaller size on the EX1 than on the HVX, because of its number.

Ssswwwweeeeeeeeeettt!
 

WOW.

Regardless of pre-set gain settings all I can say is WOW!!!

Super clean and there could have easily been less lit and the image would be usable. That just sealed it for me. I can't wait to buy one, now and maybe make this the workhorse cam!

What a joy it would be to need very little or no on camera lighting. Thanks for the comparo!

I would love to see this done with a few different cams too.

Z1, PD, V1

Any takers?

Mike

That's how I feel about it. If it's clean and can offer this much light under those conditions, I'm a happy camper.

Have a look at some more news from the Australian roadshow:
http://xdcam.com.au/modules/news/

There's a guess that the ASA may be around 800.

Hi,

We are impressed with this camera. We have several on order. Even with our REDs coming this December!

This test however, does not quite compute. I agree with the septics here. You can achieve the same thing even with the light loving Z1, so long as you put it on a tripod, lower the shutter speed to..... and don't move a muscle.

longer shutter speeds are great for certain effects, however it is no measure of the cameras true low light capabilities. The two photos here do not contain enough information about the shoot to make any real conclusions.

Looking forward to more detailed tests.

Cheers,
G

Thanks for posting that Steven.

Two things jump out at me. Memory cache recording not available for this model. I have heard elsewhere that although not available at initial release it will be a firmware upgrade. I'm going to double check that with Sony if I can.

The second is the claim that it takes 3 minutes to download 1 hour of "footage" from the SxS card. At 3x speed it would be 20 minutes to download 1 hour of video (recorded at 35mbps). I can't help someone got confused between 3x and 3 minutes.

At -3dB gain the EX1 provides correct exposure rated at 800ASA. Shots of a model exposed at +/- 2 stops corrected well, without noise being visible in the 2 stops under material I saw projected (unfortunately using a data projector). 2 stops over and the high lights were burned, but the soft graduation looked natural. Don't know the gamma curve selected. Was impressed by images in a poorly lit auditorium (at -3dB) where the Z1 would have needed 18dB. This was a default gamma and looked even better when I dialed in black stretch. I didn't have time to try out other cinegamma curves.

I hope this low light comparison is accurate...

That would be simply awesome...

But, the surface area gained by going to 1/2 inch chips does not seem intuitively to explain these kind of impressive results...

To me, these results seem to defy physics...

Again I hope it's true, but I'm a little skeptical...

Thanks Serena,
Wow... It looks like the EX will be a decent perfomer in low light. This is VERY good news! This makes me believe the sample images at the start of this thread are looking better and better....

Here is a quote from a published review (http://dvuser.co.uk/content.php?CID=171):

"I suspect Sony have left Proxy files from the EX1 for a number of reasons. First, to save space on the cards to allow for more recording time of luscious HD footage, and also, the actual import time for full HD files from the SxS cards can only be described as 'blisteringly fast'. It is getting close to 20x real time. 1-hour of full 1920x1080 HQ HD footage can be imported via the XDCAM EX transfer software in approximately 4 minutes flat! "

20x the transfer speed...? Whoa! That's amazing if true!

3x Transfer speed is for XDCAM HD disk based systems. SxS is much much faster. Rough tests I did at IBC suggested transfer speeds of around 4 to 5 mins for an hours worth of HQ material.

Sorry for misunderstanding the transfer rate.

!!!! I'm sold!!!!

The ability to offload and even verify the material will not be much longer than a quick coffee break during a shoot.

I can't help but think this camera is going to be a viable ENG camera as News moves into HiDef.

Alister, I assume you were offloading to a laptop with SxS port (MacBook or Vaio), no?

Is this transfer speed referring to transfer of files

1) from the camera itself connectd directly to a pc, or
2) from an express port on a laptop, or
3) from the USB card reader connected to a pc?

Will the 3 different modes of transfer mentioned above have any effect on the transfer rate? I've noticed that the USB express card reader neds to be connected to a mains supply - rather strange since all current card readers draw the necessary power from the USB itself. It seems that there are some more electronics inside there doing soemthing. I wonder...

The transfer from SXS is limited by by the speed of the computers hard drive! The speed is the same from the camera or USB reader and just a tiny bit faster if you have an express slot on the computer. As I said the main bottleneck is the the speed of the computers hard drive as the SxS card can stream at 800Mb/s.

As others have noted it apparently is quite fast, which makes sense given that the maximum stated transfer rate of 800 Mbps is more than 20 times the maximum recording bandwidth of 35 Mbps. So 3+ minutes to transfer 60 minutes of footage is plausible.

And yet when I asked the Sony Rep at HD World in NYC today (they had 2 EX1 on hand) he said he had tested the previous day on a recent Sony Vaio and found the transfer speed to be about 3 times real time (although he seemed a bit unsure).

Returning back to the original title of this thread, I've just noticed that B&H are stating that minimum illumination for the EX1 is 4 lux at f1.6 and 18dB!! From where did they get this figure? This is not what I call a low light capable camera at these figures! I've noticed also that they are stating that sensitivity is not quoted by manufacturer - in fact it was! I hope that it is a typo error... Even though the original official min illumination is stated as being 0.14 lx (typical) (1920 x 1080/59.94i mode, F1.9, +18 dB gain, with 64-frame accumulation), if one had to switch off the 64-frame accumulation and take in cosideration that it is a F1.9 lens, I do not think the min illumination will end up as 4 lux. .....I hope......

The 0.14 lux is an imaginative statement and is of interest if filming comets (quite useful, actually). The other figure is f/10 on a white card (90% reflectance) at 2000 lux, presumably at -3dB gain, 1/60 second ; from which I calculate 8 lux at 18dB gain at f/2.0 at 1/60sec and about 4 lux at 1/33 sec. So it's possible. Going the other way: 64 x 0.14 = 8.96 lux. I think I've seen the 8 on a presentation table, so that might be right.

edit: when the EX is put in the "low light" category that is relative to other HD cameras, so if one has been shooting with a Z1 then 8 lux is good.

Thanks Serena for the technical reply. The problem is that I'm presently used to filming at very low light situations with an SD Sony camcorder (DSR-300) rated at F11 at 2000lux - and it does wonders. I know that SD and HD sensors differ but I'm just curious to know how much will I loose with a F10 at 2000 lux sensitivity? Will it be just one stop or could it be more?

Confused: is that "8 lux is good" statement supposed to be compared with the quoted "4 lux" rating given in the Z1 spec?

f/11 is a 1/3 stop smaller than f/10 -- the same, in practice.

In fact given as 3 lux for the Z1, but I think that is much the same sort of figure as the 0.14 lux stated for the EX; 1/3 sec shutter, 18 dB gain, f/1.6
Generally the Z1 "speed" is rated about 250 ASA (0dB), whereas the EX is rated as 800 ASA at -3dB and looks to have a stop in reserve (in fact 2 stops from some shots seen at the roadshow). Rating the Z1 at 250 ASA, then at 1/60 sec shutter at f/1.6 it needs 175 lux (reading off my Lunasix meter).

Of course that is taking an incident light reading which is calibrated to be the same as reading of reflected light from an 18% grey card. So the actual equivalent figure is probably about 32 lux for the Z1.

To really measure low light performance isn't exactly a trivial task, the quoted figures mean little without a S/N figure and even then there seems to be no agreed standard for how that is measured. I've seen (and shot myself) footage from some HDV cameras that it's claimed show 'acceptable' low light performance and yet to me looking at the amount of noise I have a problem considering it acceptable.
I think it also worth considering that Sony have been using a bit if digital magic in their recent camera such as level based dynamic noise reduction and that can have some undesirable side effects. So there's an even greater challenge with the EX1, not just getting a comparative set of figures against its peers but also checking for any unwanted artifacts when it's stressed.

From what i hear, the EX1 is a lot better in lowlight than a Z1.
Someone who is used to Z1s has played around with an EX1 and its a lot more sensitive for sure.

The specs at the NYC store are usually correct, but not always. In this case, i would suggest that the 4lux figure is an error. And as someone previously pointed out, the senstivity IS provided by manufacturer, which contradicts the NYC stores statement on their specs page.

Since the specs look decent and we're hearing from a couple hands-on review that it fairs well in low light, I'm willing to believe it's decent.

I've contacted B&H regarding this query. Their reply which was received a few minutes ago is as follows:

"SONY PMW EX1 specifications

from the Sony website

http://www.sonybiz.net/biz/view/Show...DCAMCamcorders

Sensitivity (2000 lx, 89.9% reflectance)
F10 (typical) (1920 x 1080/59.94i mode)

Minimum illumination
0.14 lx (typical) (1920 x 1080/59.94i mode, F1.9, +18 dB gain, with 64-frame accumulation)

S/N ratio
54 dB (Y) (typical)

from the B&H website

http://www.bhphotovideo.com/c/produc...AM_EX_SxS.html

Sensitivity Not Specified by Manufacturer

Minimum Illumination 4 Lux f1.6 at 18 dB

Vertical Smear Not Specified by Manufacturer

The way I read the specs the light sensativity of this camera is not very good.

The Sony site shows the the sensativity at F1.6
while the B&H site is showing the sensativity at 1.9

The bench mark for illumination camparison is taken at F11, the limitation of the camera lens and appeture do not allow that reading, so technically you would want to take the sensativity measurement at the highest stop.

The B&H website offers the specifications that are available at the time of the publishing of the webpage.

The specifications are offered as a convienience to our customers and should not be considered as control tested by B&H

It is always best to defer to the manufacturers website whenever possible

Just keep in mind that the manufacturers will offer the specs that make their products look the most desirable."

I admit I didn't really understand this reply. If the sensitivity is a standard benchmark (as opposed to the highly debatable lux ratings), than theoretically a sensitivity of F10 is very very good. As far as I know the best SD cameras have a sensitivity of F11. However having said that I'm not taking in consideration the S/N figure as Bob Grant has pointed out.

I'd like to know exactly where B&H got the 4 lux figure because I don't see that number in Sony's literature, and B&H didn't state the sensitivity which is clearly listed in Sony's specifications. Looks to me like someone at B&H pulled a lux rating number out of thin air, and many people are quoting that instead of referencing what Sony says.

The Sony Z1U is widely rated at 3 lux and the EX1 rating should be lower than that; more likely around 1-2 lux.

P.S. According to the following review the EX1 is a hair more sensitive than the F330/F350 cameras: http://digitalcontentproducer.com/ca...ex/index2.html

What does B&H video have to do with real actual specs?
They are a store. They sell camera supplies.

If you're worried, you best bet is not to buy the camera until you can try it, or wait for more reviews.

So far the hands-on pre model review are positive for sensitivity.

Looks Accurate for the HVX200
 

I'd hate to break it to some of you HVX200 lovers out there but the camera IS NOT GOOD in low light. The frame grabs in the first post look very similar to tests I've done using 0 dB, 1/60, 1080/60i settings on the HVX200. While I love the Panasonic in every way, it requires a decent amount of light to look good. Anything over 6 dB of gain on that camera looks like garbage in HD. 1/3 inch chips just can't cut it at the moment for HD acquisition in low light.

The setup for the photo looks like a living room with no couch lighting. The only lighting is from that lamp and it's not pointed at the couch. If the Sony camera can do that good with little to no noise it will be a stellar camera. The Digital Content Producer magazine preview mentioned that it's a tad faster than the PDW-F330/F350 cameras in low light. Sensitivity of F10 at 2000 lux, whereas the PDW-F330/F350 are F11 at 2000 lux.

On a side note, I really hate how tests are conducted on independent sites. Those two pictures aren't worth the bandwidth used at 600pixels wide. Where are the full raster 1920x1080 pictures at? And then you have screen grabs that will be released when the camera comes out that people color corrected. BOO! Raw data please at practical settings. 24p at 1/24 shutter speed does not count (people were using this to show HVX200 sensitivity pics).

Let's use BitTorrent to host the jpeg files (or PNG files) if bandwidth limits are a concern!

If this camera lives up to the hype, Panasonic is making a huge mistake by not updating their chip for the next iteration of the HVX200. Or for that matter, not upgrading their LCD screen to HD.

Let's keep up with the data here: Sony specifies the minimum illumination for the Z1 as 3 lux and the minimum illumination for the EX as 0.14 lux. What do these figures mean for people wanting to shoot in low light? Not much because those figures do not represent minimum illumination in normal video acquisition modes. The most useful figures to have in mind are 250 ASA (or ISO) and 800 ASA, the former being established in studio evaluations shooting under normal acquistion modes. This says the EX is about 2 stops more sensitive than the Z1. A conclusion reached some posts above by working at it from various directions.
How far can that be pushed on a particular shoot? Much depends on the scene (high key or low key) and mood. As Bob points out, the theoretical maximum speed may not be useful in a particular shoot.
The roadshow demonstrated latitude in the EX with a high key scene, which made it very difficult to see what was happening in the blacks (because there were hardly any) when pushed two stops. I think it safe to assume that at an ISO (ASA) of 3200 the EX will be noisy, although I couldn't see it in the corrected footage.

I'll add that my Canon EOS 5D starts to show noise at 1600 ISO and 800 is a good working maximum. That has a full 35mm frame sensor.