Get 9.5 Stops of dynamic range from your DVX100

pertierr said:
The gain on the DVX is not a DSP function. It is wired directly to the A/D converters, so the gain switch affects both the Andromeda output and the DV output. This test was exposed for DV and we took no care with the Andromeda side specifically to make it as unbiased as possible. Indeed this is a "best case scenario" as far as the DV dynamic range, and a worst-case for the Andromeda side.

If you're not convinced, I can easily point out in the DVX100 schematic how the gain is wired. But please don't label this as misleading if you do not know how it works.

Cheers,
Juan
Click to expand...

I don't mean the actual gain function of the camera, but the digital processing that goes on after that. I would be interested to see the schematic if possible, thanks.

I am not trying to knock the product, I just find some of the pictures on the website to be misleading. On some, the "DV comparison" is clearly over-exposed and the uncompressed isn't. Therefore there must be some addition increasing of the levels before it is recorded to DV. It is clearly not only an effect of increased lattitude, the DV stuff IS overexposed. I have seen other people asking similar questions both here and on other boards.

Don't get me wrong, the product looks great, and I am not trying to discrediting your claims, just some of the stuff on the website comes across a little biased. :)
 
mmm:

The argument that the "DV side is overexposed" is a result of misunderstanding what is going on and what the two images represent.

The statement would be correct if we were comparing two different cameras. However, the DV and Andromeda images on the website are not two different cameras, but the same camera, and furthermore they are the exact same CCD data, just recorded at different points on the processing chain.

This is very simple:
1.The CCD's receive light and output an analog signal.
2.The A/D converters take that Analog signal and convert it to digital with 12-bit precision. This digital signal is what Andromeda records. At this point, Andromeda is done.
3.The output from the A/D's is processed internally by the camera, decimated, resized, compressed, etc to be put in DV format.
4.This DV stream is either recorded to tape or sent out over firewire.

For the purposes of testing dynamic range reduction on the DV side, it DOES NOT MATTER if either the DV or Andromeda side is over exposed or under exposed. Why? Simply because both outputs start from the ~same data~, coming from the same CCD's. Andromeda captures everything that comes out of the A/D converters, and the DV output is just a processed version of the data that Andromeda captures. Therefor, it is ~physically impossible~ for there to be more information on the DV output that is NOT on the Andromeda output.

Simply put, whatever data there is at step 2 above, that's all there is. At step 4, you can only have an equal or less amount of information that originated from real optical information. This is mathematically true, and provable.

Furthermore, just about ANY wide dynamic range image recorded simulatenously via Andromeda and DV, can be used to show a reduction in dynamic range on the DV side, because simply any information that is visible on the Andromeda image (A/D stage) but not visible at the DV output, was simply cut out by the internal camera processing.

Thus, the Andromeda/DV frame pairs in which the DV image looks overexposed, still show a reduction in dynamic range, simply because the details that you can see in the Andromeda image are indeed captured by the CCD's, but the internal camera processing cut it out.

I hope this helps. You're right in that i've heard a few other people say it before, but it is a complete misunderstanding of what those images represent. I understand that initially one might treat this just as if one where comparing two cameras, but this is not the same thing at all.

In any case, the latest dynamic range images are just about the most thorough and accurate dynamic range comparison one can make. They were exposed for DV, and Andromeda still shows a lot more dynamic range.

Let me know if there is something you want me to explain better. Feel free to email me at Juan@reel-stream.com and I will be happy to go into as much detail as you want.

Cheers,
Juan
 
That is a sweet mod.

If I had another 3 large to spend on a camera tho, then I think I would sell my DVX and buy the HVX.
 
pertierr said:
The statement would be correct if we were comparing two different cameras. However, the DV and Andromeda images on the website are not two different cameras, but the same camera, and furthermore they are the exact same CCD data, just recorded at different points on the processing chain.
Click to expand...
Juan,
Just because you are using the same camera and the exact same view and pulling the exact same footage at the same time doesn't mean the normal DV version isn't REALLY overexposed.

For instance.

If in order to get the best quality of exposure for your Andromeda image it meant that you HAD to run the signal "hot" for the normal DV video... then of course the DV that is coming out at the same time your getting a great picture on the Andromeda output would not only LOOK over exposed but would BE over exposed.

The real point is that you should be comparing the best exposure for the Andromeda output against the best exposure for the DV output. (And I think the Andromedea will still look better and have much greater clarity anyway.)
 
Oh, I agree, i'm not saying the DV footage(on the old images on the website, not the latest comparison) is not over exposed. All I'm saying is that for the purposes of evaluating if the Andromeda footage has a greater dynamic range than the DV output, a shot of any scene with a wide range of light intensities will work, regardless of whether it is clipping or not.

I'm addressing the argument that because the DV footage is overexposed, the two frames don't show that the Andromeda image captures a wider dynamic range. This is simply not true. The proof is in that more dynamic information made it to the Andromeda image than to the DV image. It's simple.

Once you determine that the in-camera processing cuts off X stops on the high end, that remains true for the same settings, regardless of what the scene is.

Cheers,
Juan
 
I was one of those arguing that the DVX version was simply overexposed, but I've spent a lot of time e-mailing back and forth with Juan, and the simple idea is that no, it's not overexposed, it just lacks the ability to retain the latitude at the top of the scale. I pointed several examples out to Juan that showed what I considered to be underexposed DVX shots, and he said that it's all due to the lookup table (LUT). I was still unconvinced until we saw this latest round. Here the exposure is clearly the same, and the dark levels are all very, very close between the regular DVX and the Andromeda. The only significant difference is what happens when the light gets bright -- Andromeda retains the image detail, but the regular DVX clips it.

I don't think the existing examples on their site do a very good job of explaining it, they certainly led me to conclude that it was an exposure error. But this latest example looks conclusive.
 
So my next question is... how much hard drive space are we talking about for one hour of uncompressed video??
EDIT: Just found this: "A 100GB drive will give over 1 hour of uncompressed 4:4:4 footage."
http://forum.reel-stream.com/viewtopic.php?t=249


And I read on your site: "Allows for HD output up to 1540x990 from the DVX100 by utilizing the pixel shift between the CCD's. "
http://www.reel-stream.com/andromeda

Does this mean your talking High Definition output??
 
I have a question regarding noise. Is noise generated by the DVX in low light a problem at the CCD level, the A/D level, or at the DSP and/or DV codec level? In other words, would darker areas of a scene (shadow detail) have as much noise with the Andromeda?
 
Seeing those snapshots really makes me want to get the Andromeda instead of the HVX. Just have to figure a way to make the capturing rig more transportable. A special backpack maybe?
 
Fred,

That's an excellent question. Even though the in-camera processing perform operations that can amplify the noise, the noise mostly is generated in the CCD's so overall the noise will be there. However, given that the DV processing always outputs 8-bits in a pre-defined manner, being able to define the RGB LUT(Curves) from the original 12-bits allows you to capture regions of that 12-bit x 3(36bit) data that you never see in the DV output.

About the HVX vs Andromeda, i like to tell people that they are different solutions. As far as I know, Andromeda installed on the DVX100 is the only Full bandwidth 4:4:4 Uncompressed acquisition system for under at least $65,000. However, it is a different setup than having a completely self-contained camcorder.

I tell people that it's all about what they want to do. If you can deal with compressed footage, and the look out of the camera, then the HVX will most probably work great. However, if you have color-correction in mind to give a specific look, chroma keyeing, effects shots, or a need for increased color precision, full color sampling and dynamic range, then Andromeda+DVX100 offers something that no system in its price range can do.

Let me know if you have any other questions...

Cheers,
Juan
 
Flintstone said:
Seeing those snapshots really makes me want to get the Andromeda instead of the HVX. Just have to figure a way to make the capturing rig more transportable. A special backpack maybe?
Click to expand...
That is my concern as well. How have some of the people who have bought this figured this out? What might you suggest to make this more feasable?

This looks like a great product. You really need to be out front on suggestions on how to make this more 'real world' possible and you will sell a lot more of them.
 
If you look at any professional movie shoot, there is at the very least one thick cable attached to the camera. For example, if you have ever seen the "Behind the scenes" for the movie "Collateral", you'll see at least 4 cables coming out of the camera, and not going to a powerbook or a mac mini, but rather refrigerator-sized raid arrays and multiple monitors.

Last time I checked, they shot it in the real world. :)

I guess my point is, if you will be running around with a camcorder in your hand, then you will have to decide if uncompressed footage is really worth a change in your operations. Because if it isn't, I don't know of any other uncompressed solution that is more portable, or cheaper than Andromeda. If you work on movie/spot sets, then having cables coming out of the camera will not be anything new.

We understand that the perfect system would be all self-contained, but this would be a brand new camera, and cost a lot more than what our system costs. Our goal was to make Uncompressed 4:4:4 10bit, 12bit video available to everyone.

Cheers,
Juan
 
Flint, I have built a portable "Monitor" that has a built in mac mini for recording, it works well..I will post some pics asap..it runs for 3 hours without 110v AC..we just did a greenscreen shoot in our studio with the rig. no real issues..
 
obin said:
Flint, I have built a portable "Monitor" that has a built in mac mini for recording, it works well..I will post some pics asap..it runs for 3 hours without 110v AC..we just did a greenscreen shoot in our studio with the rig. no real issues..
Click to expand...
This sounds interesting. What about swapable hard drives out of the min mac? Since you can only get about an hour per 100GB of drive... then you will need lots of them to complete even a days shoot.

EDIT: After reading this the mini is NOT going to allow swapable hard drives!!
http://www.macintouch.com/macmini/review.html

You must be running the mini off of a battery converter?? (for field shoots)
 
Obin can give you the details, but I just wanted to add that you do not need to record to the internal drive, you can record to firewire drives as well.

We've tested 7200RPM notebook drives in enclosures that take power from the firewire bus, so there is only one cable between the drive and the mac mini/powerbook.

A 100GB FW drive can be used as a 'cartridge' and holds over an hour of 4:4:4 uncompressed footage, plus it's a lot cheaper than even the smallest P2 cards which only holds a few minutes of compressed footage.

Cheers,
Juan
 
Thanks for the info Juan. Could you post some comparative pics eventually?

Obin, looking forward to see those pics of your Mac Mini and screen rig. Do you have a touchscreen, or are you using the mouse to navigate in Sculptor?
 
You mean comparisons between the Uncompressed Andromeda output and DVCPROHD? We are planning to post comparisons as soon as we can test an HVX. However, Obin is a beta tester for us and he frequently uses the Varicam on shoots. He might be able to better express how the Uncompressed output compares to DVCPROHD from the Varicam.

He also shot the Varicam frame in this image:
http://www.reel-stream.com/maycap/and_vs_vari.jpg

This is not meant a comparison, and we only shot the Andromeda image, Obin shot the Varicam.

Cheers,
Juan
 
Before now, I just couldn't see the benefit. But, I agree with Barry Green, Adam Wilt, and others. This proves that Andromeda does something remarkable. I can almost imagine making this work because the DVX just doesn't have enough latitude in some circumstances (theatrical spot lights, for example) and even though I hate, hate, hate the idea of tinkering with the internals of a perfectly good camera, the benefits are there. It's obvious.

RealStream:
Thanks for figuring out a way to sell me on something I never, ever wanted to buy!

Obin:
I'm sorry for giving you a hard time.
 
You must log in or register to reply here.
Back
Top