HVX200 vs XL-H1 Highlight Blowout Analysis
 

I have been on the fence between the XL-H1 and HVX200, so I decided to revisit my highlight blow out tests. I am still amazed that the HVX200 with DVCPRO HD has highlight detail beneath apparently blown out highlights, while the XL-H1 just simply does not. Once that blow out is on HDV, it appears to be really in there.

Below are comparisons between HVX200 and XL-H1 source footage brightness reduction using FCP color correction (whites).

HVX200 Examples
Example 1 - Before and After
Example 2 - Before and After
Example 3 (My favorite) - Before and After

XL-H1 Examples
Example 1 - Before and After
Example 2 - Before and After
Example 3 - Before and After

I think I'm back on the HVX200 train. :P

-steev

I want to share another test here.

I did another test which isn't perfect but still proved something important. I took a frame from raw HVX200 footage, recompressed it both to DVCPRO HD and to HDV. Then I brought those clips back into FCP to see how the CC brightness tweaking would behave. My theory was that the HVX footage encoded to Mpeg2 will lose the ability to handle highlights the way DVCPRO HD does.

And this test is showing exactly that.

http://www.holyzoo.com/content/hvx20...OHD_vs_HDV.mov

Use your left and right arrow keys to toggle back and forth to analyze the differences - there are many.

None of this is surprising, DVCPro HD is a more forgiving codec than HDV. If you've got and H1 the very simple answer is to make sure you expose correctly!

Or capture the HDV to an intermediate like CFHD or Canopus HQ(both of which are better than DVC-Pro HD IMHO) and get better or equal lattitude.

Steve, could you post before/after with the FCP waveform monitor as well, so we can see what's happening with regards to the highlights?

Graeme

However, what I'm seeing is once you've committed to Mpeg2 (HDV) there's no getting it back, regardless of intermediate post codec. I tried the test from HDV to Uncompressed as well. The codec latitude is not there. Admittedly, this is very subtle stuff, and I'm being contested on DVXuser that it's a insignificant moot point. :) Would it prevent my from buying one over the other based on this alone? No, but it adds to the equation.

Graeme, here's another .mov with 4 .jpgs in it with Waveform and Color Correction filter included. 2 with before/after using Mpeg2, 2 with before/after using DVCPRO HD.

http://www.holyzoo.com/content/hvx20...OHD_vs_HDV.mov

Thanks Steev. The movie is great, but not as elucidating as I'd thought it would be.

Graeme

Well folks, I hate to waste everyone's time with Codec latitude speculation, but I believe true scientists will admit when their findings swing away.

I tried bringing an image from a 12MP in RAW format, downres'd in FCP then exported both to HDV and DVCPRO HD. Then I applied the same CC test, and I don't see a bit of difference!! Wow. Okay, so well, hmm, I don't know!

However, here's the good news. My decision has been baked, formalized, and completed. Because.... I called BHphoto to cancel my preorder from Oct 22nd, and they said the system shows that 1 out of 9 HVX200s they just received is coming my way!

WOOOOOO F**N HOOOOOOO!!!!!!!!! Two 8GB Package. If this aint true, I'm gonna go so postal.

Very good testing.....

Good stuff, Steve. When you get your HVX please give us some appetizers to nibble on. I'd love to see more real-world HVX stuff.

Nice stuff Steev.

I hope you get your camera soon.

I further into the preorder list.
I'm probably lucky if I see it within a month.

Note: the above analysis has nothing to do with the codec, but how the NLE you used handles super-whites in YUV space. It seems that the NLE used is broken for HDV handling as it is clipping YUV to 235. All cameras (HDV included) will encode luma levels above 235, yet some poorer quality NLEs will use cgRGB processing which will clip Y levels to the 235. On the PC only Vegas defaults to full range YUV processing (by using vsRGB), Premiere Pro with Aspect HD also handles this correctly (although PPro alone clips the YUV space.) Which NLE was used?

Looks from the downloads and Steev's profile like he is using FCP.

As I said on DVXuser, this has nothing to do with CODEC, the XLH was improperly setup. The knee was not set to low (you can tell by looking), the footage was overexposed anyway and the master pedestal and setup level were not adjusted at all...


ash =o)