NV-MX300A Low Light Test
 

I've been bothered for some time now about the numerous LUX doubts and criticism concerning the MX300. I presumed its high pixel 1/4" CCDs were a detriment. So I decided to do some tests to see how bad or how good the minimum lux requirement is. I've tried to be as scientific as possible, though I had to be subjective in judging the playback footage. What I couldn't do is compare low light footage with several cams. What I did do was compare the low light footage between the GR-DVL9500U and the NV-MX300A, under the same conditions. For those of you who don't know the DVL9500, it was the clear low light winner in PC Mag's 1999 cam review. It was only beat out by the XL1. ("The JVC GR-DVL9500 posted strong scores on our Low-Level Image Analysis tests, and our jury ranked it second only to the $4,500 Canon XL1. This JVC camera is an excellent choice." ---PC Magazine)

http://www.zdnet.com/pcmag/features/.../features.html

The cameras tested were the:

Canon ZR
Canon Vistura
Canon Optura
Canon XL1
JVC GR-DVL9500
Panasonic PV-DV910
Sharp VL-PD1U Slimcam
Sony DCR-PC1
Sony DCR-TRV900

Note: in Webster's Third New International Dictionary, lux is defined as: "a unit of illumination equal to the direct illumination on a surface that is everywhere one meter from a uniform point source on one candle."

* Hallway Low light comparison/analysis

The hallways in my apartment building are dimly lit with 2 overhead lights, about 14 feet apart. The bulbs used are 60 watts, but are encased with yellowish light covers. The light is very poor, as a result. (I have cheap landlords). I slowly walked through the halls, like a nut, with the MX300 in one hand, the DVL9500 in the other. I recorded in auto mode with both, but I had the DIS shut off with the DVL9500. (Zoom was set at minimum.) I then dumped the footage on a VHS tape and viewed the footage on an ordinary 20" color TV. (Of course, I had to convert PAL VHS to NTSC VHS via pass-through with the AIWA MX100 converting VCR.)

Results: the results were near the same, except that the DVL9500 was a little brighter. Also, the image was much clearer (sharper) with the MX300. The auto focus and white balance could not hold a firm picture with the DVL9500.

* Low light in elevator comparison/analysis:

Results: the same.

* Underground apartment parking lot comparison/analysis:

It was dark down here, with only the odd fluorescent tube. Again, the results were near the same, except that the MX300 was a little brighter. Also, the image was much clearer (sharper) with the MX300. The auto focus and white balance could not hold a picture with the DVL9500. I also noted that when I pointed the cams at a fluorescent tube, there was tremendous flicking with the PAL MX300. There was no flicker with the DVL9500.

* True lux test (still subjective):

I used the smallest candles I had for the test, "birthday candles." I recorded 1 meter from the candles. All other lights were shut off. I did this at night, of course.

Again, same results with the DVL9500's auto focus and white balance unable to hold a picture. I started with 1 candle, then 2 etc, until I had 5 burning at once. When I had 3 candles lit, I noticed a considerable improvement with the DVL9500's picture, because the auto sensors seemed to have enough light without getting confused---jumping around all over the place. However, the MX300's footage also improved as I added more candles.

* Conclusion:

I was quite surprised at the results. The MX300 beat the DVL9500 hands down at every turn. The quality of the MX300 video was that much better, brighter, sharper and cleaner. I also shot with the DVL9500's by setting the white balance, exposure and focus manually. This greatly improved the quality of the video, and drastically lowered the lux requirement---perhaps beating out the MX300 with lux. However, this was difficult to determine and the MX300 footage was still cleaner. Anyways, I hope this helps answer the lux concerns of the MX300. And by the way, I still love the DVL9500, though I'll have to use the MX300 a little more often! (except under 60 cycle fluorescent lighting)

One final note, no cam engine is picked up with the MX300 during recording; whereas the DVL9500 is one noisy bugger.

http://www.dvfreak.com/lux.htm
http://www.dvfreak.com/pana_mx5.htm

Camcorder guru John Beal posted a low light comparison
between the MX3000 and several other 3-chip cams, like
VX2000, TRV900, & GL-1. The MX3000 looked cleaner than
both the TRV900 or the GL-1, but not as good as the VX2000.
But, in my opinion, the VX2000 is in a different class. I paid
$1600 for my MX, you'd be hard pressed to get a brand new
VX2000 for even $1000 more. MX is the best image for the
money. Judge for yourself:


http://www.bealecorner.com/trv900/lowlight3/index.html

Thanks for the link. The test I did was with auto mode only, and the MX300 was the clear winner. I plan to re-do the test with running these cams in manual, matching the shutter speeds, on a tripod. You only paid $1600 for your MX3000? That was a deal! Did you buy it from Tim?

This weekend I'm going to do the same test, but in manual mode---with these 2 cams. I'll post my findings. Will the DVL9500 perform better? I don't know yet. I wonder if the MX300 performs better in low light because the auto sensors work better?

Hey Frank,

If you haven't already performed the low light tests, may I
suggest an other pair of parameters to add to your test images?

The MX300(0) has two auto-settings for low light situations:

1) Gain-Up mode

2) Low Light mode


To see both of the above settings side by side with images
from "Full Auto" and "Best tweaked" manual would be VERY
interesting. When are you going to post the results ?

I haven't yet had time to perform the manual mode setting test. I'll see about your suggestions. What I want to do is use all manual settings on both cams with matched shutter speeds---a sort of shoot out, and give me an idea which cam has the best low light (in manual mode). I don't know what the results will be, until I complete the test. However, I have a feeling the DVL9500 will have the better low light footage, but I could be dead wrong. I'll try and do it this weekend (I know, I promised it for last weekend---sorry).

Hey Frank,

We know that you're a busy man, but how about posting those
Low Light test results? What I'd REALLY like to see is a low light
shoot out between the MX3000 and the MX5000, just to see if
there is any truth to the speculation that the MX3000's 1/4" CCDs
would render better images in dim conditions compared to the
MX5000 1/6" (mega-pixel) CCDs.

The re-test in manual modes will take me about 4 hours (with viewing). I haven't forgotten. They'll be coming in a few days. I can't do it tonight. Probably tuesday or the following day.

I'm sure the MX500/0 requires more light, but won't be able to do a test between the MX3 and MX5 until I own a MX5! Hey, I'm not a rich guy. Right now I'm working on a small article and building a computer for video editing. Money money money! (more more more...please---pretty please?)

I will try to get information from Panasonic regarding the amount of lux it requires to match the minimum lux required for the MX3. This is very subjective though.

http://www.dvfreak.com/lux.htm

I shot more footage in manual mode/tripod and dumped it on a VHS tape. I just have to view the footage, and write/post the findings. That'll be Monday (tomorrow).

Both cams were mounted on a tripod, OIS/DIS off, stutter speed at 1/100th of a second. manual focus, no iris lock (yes, you can lock the iris with the GR-DVL9500's), manual white balance, manual exposure.


Hallway Low light comparison/analysis:
Results: the results were near the same, except that the DVL9500 was a little brighter. Also, the image was much clearer (sharper) with the MX300. Both cams held a firm picture.

Low light in elevator comparison/analysis:
Results: the same.

Underground apartment parking lot comparison/analysis:
Again, the results were near the same, except that the MX300 was a little brighter, But the DVL9500 image was cleaner. (There was that strange flicking with the PAL MX300 image when pointed to the lighting. (There was no flicker with the NTSC DVL9500). I guess I'm comparing an apple with an orange.

Truer lux test:
Using "birthday candles." I recorded 1 meter from the candles. All other lights were shut off.

Starting with 1 candle, then 2 etc, until I had 10 burning at once. (Time consuming, and I had to replace the candles a few times.)

When I had 4 candles lit, I noticed a fair image with the DVL9500. To be expected, as more candles were added, the image improved.

A similar image was achieved with 10 candles for the MX300. The footage also looked sharper, whereas the 4 candle recording with the DVL looked more yellow and duller (less sharp).

Conclusion:
The results were reversed. The DVL9500 beat the MX300 with the candle test. However, The quality of the MX300 video was much better overall: brighter, sharper and cleaner. Final word: do not view this as a scientific test. But this will give you an idea of how these cams work in lame light.

Great work, Frank! You're the man!!!

Thanks for doing the hard work & sharing your results with us.

I am curious to know how many candles the MX300 required to look as good as the DVL9500? As in, giving as much detail as the DVL9500 with 4 candles?

The MX300 is rated at 8 lux if I remember correctly. What's the official lux of the DVL9500?

Looks like the MX500 should have an unofficial 10 lux low light specs instead of the official 15 lux. Was the MX500's picture very grainy at 10 candles?

Thanks in advance!

The MX300 took about 10 candles for near the same image at 4 candles for the DVL9500. BUT, overall the MX300 image always seemed to be sharper, even when the image was darker. This was very subjective, though. I did not have a lab with a bunch of experts viewing the playback.

This is why I'll keep my DVL9500's and my MX300. Besides, they're great for run n gun---now only if I can run.

How many candles did the MX300 need to get a viewable image at all ( where you could recognize someone, though with some noise and grain ).
Dumb question: What is an mx300 and when was it made? Is this a different kind of animal than my 852?

- Larry

A viewable image is subjective. The DVL9500 needed less light but with just a bit more, the MX300's was way sharper. With both cams I could see an image with 1 candle, but it was a poor image.

The MX300 is about 5 years old. The MX350 replaced the MX300 and the MX500 replaced the MX350, then the GS100 came along. Of course, I have missed PAL and NTSC models because the MX350 only came in PAL and the GS100 only in NTSC.