low light versus pd150 at DVinfo.net

Tung Bui · July 19th, 2003, 11:42 AM

Has anyone had a chance to compare the dvx with the pd150/vx2000? I've read mixed messages regarding its low light compared to the sony. Some say it loses out abit in sensitivity and that the gain causes abit more grain. The dvx is abit thin on the ground here in Australia so I cant rent one to compare.
Has anyone had real life experience with both? I often do shoot at outdoors events at night and I've been tossing up between these two for awhile.

Barry Green · July 19th, 2003, 12:32 PM

I've used both extensively. The DVX100 is a little bit more sensitive, but the PD150 seems to have a bit cleaner gain signal. If you're shooting with no gain on, the DVX100 will give a cleaner picture. If you're relying on gain, the PD150 may take the lead.

The DVX100's auto gain circuit only goes to 12db, whereas the PD150's auto gain goes to 18db. The DVX100 can do 18db but you have to manually select it in the menus and assign it to one of the user buttons.

Really, though, the low-light issue between these cameras is splitting hairs. They can basically be considered equal in low light, so let other factors decide your choice (i.e., whether you need 24P, or DVCAM, etc).

There are some good comparison pictures between the VX2000 and DVX100 at:
http://www.bealecorner.com/dvx100/compare/index.html

(but notice that he only used +12db on the DVX100, not +18db, so even though the DVX100's "gain" picture is darker than the VX2000's, it didn't have to be, he could have used +18db).

Peter Jefferson · July 19th, 2003, 01:04 PM

on top of that it has a built in wide angled lens and as mentioned, true progessive..
the PAL version uses 25fps true progressive as opposed to the 24p

Peter Jefferson · July 19th, 2003, 01:06 PM

forgot to mention the PAL verisona also has a higher resolution and it prolly alil deeper in colour depth than the NTSC model..

Andre De Clercq · July 19th, 2003, 03:23 PM

Every cam can be made to have a "good picture" at zereo gain with a given minimun lux spec. "Zero gain"however means reference gain and with that gain goes allways a certain amount of noise... As long as the S/N and is not being specified (just like in pro cams) cam manufacturers can play all kinds of games with the low light figures. For consumer cams there is even more "playground" involved in order to get low light specs.. Tricks like reducing saturation levels at low light, reducing chroma bandwidth, reducing luma bandwidth and contrast, coring..and much more, make it very difficult to compaire. I never got any answer here:
http://www.dvinfo.net/conf/showthrea...&threadid=3785

Jarred Land · July 19th, 2003, 08:31 PM

you can see some low light fireworks footage here with the dvx:

http://www.dvxuser.com/cgi-bin/DVX/Y...num=1058503100

that is no gain and 24p...

I know its not the best comparison as it is a light source, but its nice to see no grain and good exposure.

Andre De Clercq · July 20th, 2003, 02:14 AM

Nice pictures,but slightly overexposed (discoloration in the bright parts). Fireworks is indeed not a "low light" application...the sky is dark, but lights are rather bright. Most camera's (also the cheaper ones) have S/N ratio's of 40 db at zero gain (=allmost invisible noise in the pictures). The better cams have 60 and more db which make that in real low light situations they can add 20 db extra gain and still look "noisefree" The 40 db S/N cams however will end up with 20 db S/N if the extra 20 db would be applied. This would result in noisy pictures. The best way to compare two cams on low light sensitivity (if both have the same min aperture) is to put the lenscap on, set both at say +12 db (or more) gain and compare the visible noise in the pictures.

Stephen van Vuuren · July 20th, 2003, 09:39 AM

Andre:

That comparison would not apply to the DVX100 since it does not have gain in progressive modes.

Andre De Clercq · July 20th, 2003, 01:28 PM

Both cams can be compared in interlace mode (with up to 12 db gain). This test will also indicate which cam has the best S/N at zero gain. Prog mode will not change S/N much.

Tung Bui · July 21st, 2003, 07:03 AM

Um, interesting. I own a mx500 and I can only say I was quite apalled at its low light. Considering how much it cost here in Australia I dont really want to get burnt again.
I'm interested in hearing about real life footage not s/n ratios, though I understand it is the objective way to do things. Unfortunately numbers can as easily disinform as well when manufacturers lie about them.
Adam Wilt did a review and amongst other things he said in real life situations the both cameras look better under different situations. If under most circumstances you cant notice a difference in brightness and grain filming at night time then I 'd go for the dvx just for its adjustability. Otherwise I'd prefer to get a useable image in the dark because no matter what you do to it you cant improve the grain or the darkness later in post.

Jarred Land · July 21st, 2003, 10:40 AM

how low light do you want to go?

Andre De Clercq · July 22nd, 2003, 12:37 PM

Tung, S/N is a basic property and when it come to real low light performance, this specification, together with the min lens F-number really specify what you can get in low light situations. (unless very advanced signal processing like motion adaptive frame averaging) is being applied Unfortunately S/N specs are mostly delibarately hidden away in consumer cams' specs..because they are too explicit...not too much commercial fantasy would be possible anymore. Because of all the tricks which are allowed in consumer cams it can indeed turn out that if two cams have comparable sensitivity that the subjective image quality is slightly image content dependent.

Stephen van Vuuren · July 22nd, 2003, 12:53 PM

Andre:

There's no point in comparing the cams in interlaced mode for most DVX100 users. Many users, like myself, never shoot interlaced and thus never use gain.

Furthermore, interlaced is fundamentally different from progressive image capture, so right now in progressive mode there is no comparison cam (progressive with 3 1/3" CCDs.

Andre De Clercq · July 22nd, 2003, 01:59 PM

Stephen, I know very well the differences between interlaced and progressive capture and related noise behaviour and artifacts. Once again, if one of the cams is better at +12db both in interlace. the perceived noise at zero gain (regardless for the DVX in interlace or pro) will be simular. I feel as if you try to limit the cam's noise spec only to prog scan...this whole thread however, starting with Tung's and Barry's questions include "gain" as an element, also Adam Wilt...maybe they all try to use the cam the wrong way...in interlace! Why does Panasonic specify the cam at +18 db if nobody uses gain.
If you know how S/N is really being measured, you will agree that the procedures are exactly the same for i and p mode, and the slight differences in CCD integration times/processing will only make 1 or 2 db difference.

Stephen van Vuuren · July 22nd, 2003, 03:38 PM

Andre:

We're missing each other here. Let me clarify. My point it that I (and other DVX100 users who shoot only progressive) are completely uninterested in what happens when gain is engaged on the DVX100 as it will not happen.

At 0db gain in interlaced vs a PD150, I agree, that would reveal results worth looking at. But once the gain circuitry is engaged on the DVX100 in interlaced mode, the resulting image is of little use in evaluating low light progressive mode on the DVX100.

In fact, when I first got a DVX100, I shot in a dark room with a single small candle. The difference between the images captured in interlaced at 12db gain and 24p progressive were so extreme as if they were shot by different cameras.

July 19th, 2003, 11:42 AM	#1
Tung Bui Regular Crew Join Date: Dec 2002 Location: Sydney, Australia Posts: 101	low light versus pd150 Has anyone had a chance to compare the dvx with the pd150/vx2000? I've read mixed messages regarding its low light compared to the sony. Some say it loses out abit in sensitivity and that the gain causes abit more grain. The dvx is abit thin on the ground here in Australia so I cant rent one to compare. Has anyone had real life experience with both? I often do shoot at outdoors events at night and I've been tossing up between these two for awhile.

July 20th, 2003, 09:39 AM	#8
Stephen van Vuuren Space Hipster Join Date: Oct 2001 Location: Greensboro, NC Posts: 1,508	Andre: That comparison would not apply to the DVX100 since it does not have gain in progressive modes. __________________ stephen v2 www.insaturnsrings.com

July 22nd, 2003, 12:53 PM	#13
Stephen van Vuuren Space Hipster Join Date: Oct 2001 Location: Greensboro, NC Posts: 1,508	Andre: There's no point in comparing the cams in interlaced mode for most DVX100 users. Many users, like myself, never shoot interlaced and thus never use gain. Furthermore, interlaced is fundamentally different from progressive image capture, so right now in progressive mode there is no comparison cam (progressive with 3 1/3" CCDs. __________________ stephen v2 www.insaturnsrings.com

July 22nd, 2003, 03:38 PM	#15
Stephen van Vuuren Space Hipster Join Date: Oct 2001 Location: Greensboro, NC Posts: 1,508	Andre: We're missing each other here. Let me clarify. My point it that I (and other DVX100 users who shoot only progressive) are completely uninterested in what happens when gain is engaged on the DVX100 as it will not happen. At 0db gain in interlaced vs a PD150, I agree, that would reveal results worth looking at. But once the gain circuitry is engaged on the DVX100 in interlaced mode, the resulting image is of little use in evaluating low light progressive mode on the DVX100. In fact, when I first got a DVX100, I shot in a dark room with a single small candle. The difference between the images captured in interlaced at 12db gain and 24p progressive were so extreme as if they were shot by different cameras. __________________ stephen v2 www.insaturnsrings.com

July 19th, 2003, 12:32 PM	#2
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	I've used both extensively. The DVX100 is a little bit more sensitive, but the PD150 seems to have a bit cleaner gain signal. If you're shooting with no gain on, the DVX100 will give a cleaner picture. If you're relying on gain, the PD150 may take the lead. The DVX100's auto gain circuit only goes to 12db, whereas the PD150's auto gain goes to 18db. The DVX100 can do 18db but you have to manually select it in the menus and assign it to one of the user buttons. Really, though, the low-light issue between these cameras is splitting hairs. They can basically be considered equal in low light, so let other factors decide your choice (i.e., whether you need 24P, or DVCAM, etc). There are some good comparison pictures between the VX2000 and DVX100 at: http://www.bealecorner.com/dvx100/compare/index.html (but notice that he only used +12db on the DVX100, not +18db, so even though the DVX100's "gain" picture is darker than the VX2000's, it didn't have to be, he could have used +18db).

July 19th, 2003, 01:04 PM	#3
Peter Jefferson Inner Circle Join Date: Apr 2003 Location: Aus Posts: 3,884	on top of that it has a built in wide angled lens and as mentioned, true progessive.. the PAL version uses 25fps true progressive as opposed to the 24p

July 19th, 2003, 01:06 PM	#4
Peter Jefferson Inner Circle Join Date: Apr 2003 Location: Aus Posts: 3,884	forgot to mention the PAL verisona also has a higher resolution and it prolly alil deeper in colour depth than the NTSC model..

July 19th, 2003, 03:23 PM	#5
Andre De Clercq Major Player Join Date: Sep 2002 Location: Belgium Posts: 804	Every cam can be made to have a "good picture" at zereo gain with a given minimun lux spec. "Zero gain"however means reference gain and with that gain goes allways a certain amount of noise... As long as the S/N and is not being specified (just like in pro cams) cam manufacturers can play all kinds of games with the low light figures. For consumer cams there is even more "playground" involved in order to get low light specs.. Tricks like reducing saturation levels at low light, reducing chroma bandwidth, reducing luma bandwidth and contrast, coring..and much more, make it very difficult to compaire. I never got any answer here: http://www.dvinfo.net/conf/showthrea...&threadid=3785

July 19th, 2003, 08:31 PM	#6
Jarred Land DVX User Join Date: May 2003 Location: Vancouver Posts: 281	you can see some low light fireworks footage here with the dvx: http://www.dvxuser.com/cgi-bin/DVX/Y...num=1058503100 that is no gain and 24p... I know its not the best comparison as it is a light source, but its nice to see no grain and good exposure.

July 20th, 2003, 02:14 AM	#7
Andre De Clercq Major Player Join Date: Sep 2002 Location: Belgium Posts: 804	Nice pictures,but slightly overexposed (discoloration in the bright parts). Fireworks is indeed not a "low light" application...the sky is dark, but lights are rather bright. Most camera's (also the cheaper ones) have S/N ratio's of 40 db at zero gain (=allmost invisible noise in the pictures). The better cams have 60 and more db which make that in real low light situations they can add 20 db extra gain and still look "noisefree" The 40 db S/N cams however will end up with 20 db S/N if the extra 20 db would be applied. This would result in noisy pictures. The best way to compare two cams on low light sensitivity (if both have the same min aperture) is to put the lenscap on, set both at say +12 db (or more) gain and compare the visible noise in the pictures.

July 20th, 2003, 01:28 PM	#9
Andre De Clercq Major Player Join Date: Sep 2002 Location: Belgium Posts: 804	Both cams can be compared in interlace mode (with up to 12 db gain). This test will also indicate which cam has the best S/N at zero gain. Prog mode will not change S/N much.

July 21st, 2003, 07:03 AM	#10
Tung Bui Regular Crew Join Date: Dec 2002 Location: Sydney, Australia Posts: 101	Um, interesting. I own a mx500 and I can only say I was quite apalled at its low light. Considering how much it cost here in Australia I dont really want to get burnt again. I'm interested in hearing about real life footage not s/n ratios, though I understand it is the objective way to do things. Unfortunately numbers can as easily disinform as well when manufacturers lie about them. Adam Wilt did a review and amongst other things he said in real life situations the both cameras look better under different situations. If under most circumstances you cant notice a difference in brightness and grain filming at night time then I 'd go for the dvx just for its adjustability. Otherwise I'd prefer to get a useable image in the dark because no matter what you do to it you cant improve the grain or the darkness later in post.

July 21st, 2003, 10:40 AM	#11
Jarred Land DVX User Join Date: May 2003 Location: Vancouver Posts: 281	how low light do you want to go?

July 22nd, 2003, 12:37 PM	#12
Andre De Clercq Major Player Join Date: Sep 2002 Location: Belgium Posts: 804	Tung, S/N is a basic property and when it come to real low light performance, this specification, together with the min lens F-number really specify what you can get in low light situations. (unless very advanced signal processing like motion adaptive frame averaging) is being applied Unfortunately S/N specs are mostly delibarately hidden away in consumer cams' specs..because they are too explicit...not too much commercial fantasy would be possible anymore. Because of all the tricks which are allowed in consumer cams it can indeed turn out that if two cams have comparable sensitivity that the subjective image quality is slightly image content dependent.

July 22nd, 2003, 01:59 PM	#14
Andre De Clercq Major Player Join Date: Sep 2002 Location: Belgium Posts: 804	Stephen, I know very well the differences between interlaced and progressive capture and related noise behaviour and artifacts. Once again, if one of the cams is better at +12db both in interlace. the perceived noise at zero gain (regardless for the DVX in interlace or pro) will be simular. I feel as if you try to limit the cam's noise spec only to prog scan...this whole thread however, starting with Tung's and Barry's questions include "gain" as an element, also Adam Wilt...maybe they all try to use the cam the wrong way...in interlace! Why does Panasonic specify the cam at +18 db if nobody uses gain. If you know how S/N is really being measured, you will agree that the procedures are exactly the same for i and p mode, and the slight differences in CCD integration times/processing will only make 1 or 2 db difference.