Noise comparison: 35/4:2:0 vs. 180/4:2:2

[Gints Klimanis]

Piotr,

"Shimmering" highlights are a problem with native EX1 footage from component output, 12MP images from D3 on Sony PS3 or even BlueRay displayed on my Sony Bravia 52" LCD - state of the art in 2007. I was unable to find any TV controls that would eliminate the artifact, even when TV sharpness was at a minimum. Scenes with detailed highlights, such as from sunny sand or animal fur, would shimmer in panning or movement, and the artifacts were annoying. This visual trash is a product of the display, but still, it must be considered.

While there is a certain amount of pride involved in displaying the perfect image on a consumer TV, consider that you will be the only one to view your high quality footage unless you provide it to a production house. A better comparison would be to author two 4:2:0 BluRay disks or 10-20 MBps MP4 files and provide two input sources that allow you to pause and swap between them.

Also, do you see noise on a static image on your TV? I see some when I pause a BluRay disc played by a Sony PS3.

[Tim Polster]

Originally Posted by Piotr Wozniacki

Hi Tim,

Thanks for you offer of inspecting some of my raw footage; what length clips (in seconds) do you think would be enough for me to upload (as I mentioned, my Internet connection - particularly upload - is low speed)?

Cheers

Piotr

Piotr, I think a good thing would be to shoot a little footage and let me know the settings used. I can then shoot a simlar scene with my setup with the settings applied and compare to two. As far as file sizes, I guess as much as you are comfortable to send. enough to play a bit in realtime would be nice to see in terms of video instead of still frames. Do you have a website to upload them to? Maybe here at DVInfo?

If you can send the internal and the nano of the same footage that would help me with the comparisons.

[Piotr Wozniacki]

Originally Posted by Dan Keaton

Dear Piotr,

The Sony Long-GOP consists of I Frames, B-Frames, and P-Frames.

They are not all created equal in terms of quality.

Dear Dan,

I'm well aware of the Long-GoP structure, and realize that the I, P, and B frames cannot be the same quality by definition. All I'm saying is that the difference between the best quality frame, and the adjacent one which happens to be of the worst quality, seems too huge to be right in my nano's L-GoP files at 100 Mbps (and higher). Also, it's much greater than that between the native EX1 L-GoP frames (relatively speaking of course, as the two formats are not directly comparable).

Interestingly, at 50 Mbps (for which the Sony chip inside the nanoFlash was designed originally AFAIK), the I, P and B frames are much closer in their quality, and hence the shimmering of noise is waaay less noticeable. This confirms my observations throughout the entire thread.

Several posts ago, I asked you for providing us with some specification of the nanoFlash L-GoP structure (like the number of frames per GoP for PAL, using closed GoPs, etc.) - it could enable me to be more specific as to which frame is which. Also, if you could provide me with some official technical information from Sony explaining why the higher bitrates would result in more quality difference between adjacent frames of the Long-GoP structure - I'd happily accept my nanoFlash as working up to specs...

Thanks,

Piotr

[Piotr Wozniacki]

Originally Posted by Gints Klimanis

Piotr,

"Shimmering" highlights are a problem with native EX1 footage from component output, 12MP images from D3 on Sony PS3 or even BlueRay displayed on my Sony Bravia 52" LCD - state of the art in 2007. I was unable to find any TV controls that would eliminate the artifact, even when TV sharpness was at a minimum. Scenes with detailed highlights, such as from sunny sand or animal fur, would shimmer in panning or movement, and the artifacts were annoying. This visual trash is a product of the display, but still, it must be considered.

While there is a certain amount of pride involved in displaying the perfect image on a consumer TV, consider that you will be the only one to view your high quality footage unless you provide it to a production house. A better comparison would be to author two 4:2:0 BluRay disks or 10-20 MBps MP4 files and provide two input sources that allow you to pause and swap between them.

Also, do you see noise on a static image on your TV? I see some when I pause a BluRay disc played by a Sony PS3.

Gints,

I'm not sure why you are bringing a "shimmering highlight" subject here - I'm talking about shimmering grain (or noise, fluctuating from from to frame).

To answer your question: yes, I can see noise on static images; however - as I said just a couple of posts ago - it's the moving video where the noise becomes really disturbing, due to the said shimmering / fluctuation.

[Piotr Wozniacki]

Originally Posted by Dan Keaton

If you want the highest quality, we recommend 220 Mbps or 280 Mbps I-Frames Only. Our 280 Mbps I-Frame Only has been thoroughly tested by a major network for one of their high production value prime-time shows using the most sophisticated of test equipment.

The quality met their specifications and needs.

Dear Dan,

Just to let you know that I do agree with the above statement of yours; my reservations only apply to the high bit rate (100 Mbps and more), Long-GoP nanoFlash format.

In one of the very first posts in this thread, I have explicitly stated that neither 220 Mbps I-Frame Only, nor Long-GoP clips of up to 50 Mbps, show the increase of shimmering noise, as compared with the native EX compression.

Although later on, I guess I denied this original assessment of the I-Fo format being OK - that was a false impression on my side, due to my difficulties at some stage to properly synchronize the EX1 and nano clips with the single frame accuracy; I was evidently wrong at that stage - sorry for that (I realize it further complicated this convoluted investigation of mine).

So to recap - yes, the I-Frame Only format offers all the benefits of nanoFlash high bitrates and 4:2:2 color resolution, while -unlike the L-GoP format - NOT introducing excessive shimmering noise .

I hope we're clear now; please address my kind requests in the previous post.

Thanks again,

Piotr

[Piotr Wozniacki]

Originally Posted by Tim Polster

Piotr, I think a good thing would be to shoot a little footage and let me know the settings used. I can then shoot a simlar scene with my setup with the settings applied and compare to two. As far as file sizes, I guess as much as you are comfortable to send. enough to play a bit in realtime would be nice to see in terms of video instead of still frames. Do you have a website to upload them to? Maybe here at DVInfo?

If you can send the internal and the nano of the same footage that would help me with the comparisons.

Hi Tim,

I wonder if Dan would make access to CD's ftp available for me to upload such 2 files - it'd have the advantage that both yourself, and the CD engineers, could analyse them and help me make sure whether my nanoFlash is up to specs in L-GoP mode.

Dan, what's your word?

[Dan Keaton]

Dear Piotr,

I will send you the FTP File Upload instructions.

[Gints Klimanis]

Piotr, thanks for the enlightenment. We need more education on how to decide between Long-GOP and I-Frame-only. I certainly would not want to collect a couple years of event footage at 100-140 MBps LongGOP only to learn that additional detail is lost in using using noise reduction in time-consuming post-production. If I have to switch from 140 MB Long-GOP to 220 I-Frame to get the benefits of the Nano, so be it. CD, please write a guide.

[Mike Schell]

Originally Posted by Gints Klimanis

Piotr, thanks for the enlightenment. We need more education on how to decide between Long-GOP and I-Frame-only. I certainly would not want to collect a couple years of event footage at 100-140 MBps LongGOP only to learn that additional detail is lost in using using noise reduction in time-consuming post-production. If I have to switch from 140 MB Long-GOP to 220 I-Frame to get the benefits of the Nano, so be it. CD, please write a guide.

Hi Gints-
Take a look at the Video Quality Vs BitRate Chart (http://www.dvinfo.net/forum/converge...-bit-rate.html) which measure video quality vs bit-rate (using the Video Clarity System). It show as general trend of higher bit-rate equals better video quality, but with diminishing returns. 100 Mbps Long-GOP continues to be the sweet spot of quality vs storage requirements. Although not shown in the chart, our tests indicate that above 140 Mbps there is little difference in Long-GOP vs I-Frame only compression. At 100 Mbps and lower, Long-GOP clearly wins.

I am writing a more detailed explanation (with help form Video Clarity) to more fully explain the tests and results.

Best-

[Piotr Wozniacki]

Originally Posted by Dan Keaton

Dear Piotr,

I will send you the FTP File Upload instructions.

Thanks Dan.

I have uploaded 2 pairs of 5-7 secs clips (each pair consisting of the EX1 native and nanoFlash 100 Mbps L-GoP):

- back-lit (even though the sky is almost blown-out, the wooden surfaces and tree barks are darkish (up to 40%), and noisy)

- in bright sunshine, but contrasty (realistic exposure, with the wooden surfaces and tree barks in up to 70% range; still noisy in the shaded areas)

Please blow up a 720 crop (typical), and inspect any GoP - frame by frame, to see how much they vary in the contained noise for the nanoFlash version as compared with that from the EX1. Please note that in 180 Mbps clips, this phenomenon is even more pronounced (I haven't uploaded 180 Mbps clips due to their size, and my slow and unreliable connection).

I will, however, try to wait for the right weather, and shoot the same scene well lit, and flat (with no shaded areas). When I upload the clips, you will see how even very brightly lit surfaces of uniform color still shimmer with noise!

Of course I realize that the input (from the EX1 camera) is noisy in the first place; however - as said before - I'm afraid it's not right for the high bitrate L-GoP nanoFlash files to augment this noise so dramatically, and in such a fluctuating fashion (which intensifies the shimmering effect).

Comments welcome.

[Tim Polster]

Great. Let me know where to get the files from and I will have a look and test with my camera.

[Piotr Wozniacki]

Dear Dan,

Any comments on the uploaded clips?

Piotr

[Piotr Wozniacki]

Originally Posted by Mike Schell

Hi Gints-
Take a look at the Video Quality Vs BitRate Chart (http://www.dvinfo.net/forum/converge...-bit-rate.html) which measure video quality vs bit-rate (using the Video Clarity System). It show as general trend of higher bit-rate equals better video quality, but with diminishing returns. 100 Mbps Long-GOP continues to be the sweet spot of quality vs storage requirements. Although not shown in the chart, our tests indicate that above 140 Mbps there is little difference in Long-GOP vs I-Frame only compression. At 100 Mbps and lower, Long-GOP clearly wins.

I am writing a more detailed explanation (with help form Video Clarity) to more fully explain the tests and results.

Best-

Dear Mike,

I'd be very grateful if - apart from the "detailed explanation" announced above - you answered my kind requests to Dan; quote:

Originally Posted by Piotr Wozniacki

Several posts ago, I asked you for providing us with some specification of the nanoFlash L-GoP structure (like the number of frames per GoP for PAL, using closed GoPs, etc.) - it could enable me to be more specific as to which frame is which. Also, if you could provide me with some official technical information from Sony explaining why the higher bitrates would result in more quality difference between adjacent frames of the Long-GoP structure - I'd happily accept my nanoFlash as working up to specs...

Thanks,

Piotr

I hope CD is not going to leave all the questions, doubts and theories, contained in this thread, unanswered. Like for example this observation of mine:

Originally Posted by Piotr Wozniacki

Interestingly, at 50 Mbps (for which the Sony chip inside the nanoFlash was designed originally AFAIK), the I, P and B frames are much closer in their quality, and hence the shimmering of noise is waaay less noticeable. This confirms my observations throughout the entire thread.

Thanks and Regards

Piotr

[Gints Klimanis]

Originally Posted by Mike Schell

Although not shown in the chart, our tests indicate that above 140 Mbps there is little difference in Long-GOP vs I-Frame only compression. At 100 Mbps and lower, Long-GOP clearly wins.

Mike, do you recommend using I-frame Only at 140 MBps and above to avoid MPEG "shimmering"? Stepping through the frames in Sony Vegas, I can usually identify a higher quality frame every six frames or so, but I would guess that Long GOP length is 15. Is it 6 or 15?

Did CD need to tune the Sony encoder differently for higher bitrates? Also, is it possible to expose the bit allocation ratio between I and P frames to the user?

[Mike Schell]

Originally Posted by Gints Klimanis

Mike, do you recommend using I-frame Only at 140 MBps and above? Stepping through the frames in Sony Vegas, I can usually identify a higher quality frame every six frames or so. But I would guess that Long GOP length is 15. Is that correct?

Also, is it possible to expose the GOP length or bit allocation ratio between I and P frames to the user?

Hi Gints-
Yes, I would definitely recommend I-Frame-Only at bit-rates of 140 Mbps and above. We don't see any improvement with Long GOP at these data-rates and I-Frame is a little easier to edit.

I suspect you are looking at the I or P frames when you identify the slightly higher quality frames in the GOP. We saw similar results with the Video Clarity tests. The I-Frame-Only quality was very smooth, while the Long-GOP quality varied from frame to frame (I-Frame was best, followed by the P-Frame and then the B-Frame). There was not a huge difference from frame to frame, but it was noticeable.

I'll speak to our engineers about the possibility of opening up the GOP length and bit allocation. But, I can't make any promises.

Best-