New Sample Footage

[Thomas Smet]

Steve while you as one single person may not see any water color effect in the footage of the church does not mean it may not be there. Are you going to say that I do not know what I am talking about? The water color effect is there but is is very small and very subtle. I just do not think you can see it. Thats fine. It is very small details. I have tried in the past to point it out but you refuse to listen to anybody. I never realized we had to take your word as the final say on the subject.

Is the watercolor bad? no not really. If you are ok with it then it is fine. It is not something that overpowers the image.

Is it there? Yes I do think so. Having compositors eyes I am a little bit more used to seeing image details because we have to look out for things like this in a composite so we get clean results.

Again I'm not really sure why you are being so negative. You ask to see proof and when we show you you just ignore it and say it isn't there. If you do not see it and it doesn't bother you thats fine but why sit there and try to claim that we are trying to pull all this out of thin air. This is not a conspiracy against SONY or you. I for one have tried to not say anything negative to you or about you on the subject but yet you are quick to put down anybody else who you do not agree with.

Again, I like SONY cameras and own a DSR-300, DSR-370, DSR-500, Z1 and a HC1 so this is in no way a way to slam SONY.

I agree with SONY on the subject that the sharpness needs to be turned down. This should get rid of any other noise reduction that may be causing the water color in certain areas which will get rid of the chattering noise. (see my post above) I only bring up all of this so people with the V1U can learn how to avoid this. What I bring up is no different then trying to help people to get better keys by telling them to reduce the detail to reduce the edge enhancement.

You can try to tell me until you are blue in the face that the noise reduction water color effect isn't there but I know what I see and I am no fool. I do not think you are either and I rather respect a lot of the stuff you have done. This one time I must disagree with you and I'm sorry for that. I have to trust my eyes however.

Now if you can show me proof that it isn't there and that I am doing something wrong I would be glad to listen just like I have with everything else so far.

[Steve Mullen]

Originally Posted by Thomas Smet

Now if you can show me proof that it isn't there and that I am doing something wrong I would be glad to listen just like I have with everything else so far.

It doesn't work that way. Anyone who claims something exists needs to prove it does. Simply repeatedly claiming you see "it" isn't good enough.

I made a very reasonable request:

1) provide a link to the clip (and to be a realiable claim, you really need multiple clips).

2) Provide a quote from the shooter that the P video was shot with Sharpness = 5 which is what Sony recommends. (This rules out 3 of Brett's clips.)

3) Describe what you see and where it is. Saying it's "on the church" isn't good enough.

In other words, you need to be as responsible as a professional reviewer would be if you are going to make claims about an effect that "... is there but is is very small and very subtle."

At this point -- not even ONE report of the alleged watercolor meets these standards for public verifiability. And, with each day going by with no new clips from the thousands of V1U buyers -- and Brett's clips disqualified -- those who make such claims need to "put up, or ... ."

In fact, if a problem is "very small and very subtle" and is reported primarily by those who don't even own a V1U and isn't reported by 99.9% of those who do own a V1U -- one could say there is no need for a debate. The issue is so trivial as to not be worth considering.

And, it's amazing that you have been knocking the V1U (based upon some compression theory) for 3 months -- claiming it won't be good for compositing and yet have never used a V1U. That's the classic behavior of a "troll."

Now I know you aren't -- and I think your compression theory is true. But to not test a V1U to verify your worries -- does seem a bit strange for a "pro" who was really considering buying a V1U.

[Bob Grant]

Isn't one of the features of the V1x that you can save cameras setting to a file on a memory stick?
Would it be too much to suggest that those who own or have use of a camera and want to post screenshots or footage include that file?
There's so MANY tweaks in this camera I doubt anyone is going to write them all down or remember them, this might help others to repro the problem, offer meaningful advise etc.
Perhaps this community can then move onto to what should be one of it's primary roles of advising how to get the best out of the camera.

[Mikko Lopponen]

Originally Posted by Steve Mullen

Blend is the correct deinterlace method for static video. It provides the maximum vertical resolution. You need to learn more about deinterlacing.

And it will also smooth noise. If you don't know why is that I'll tell you. Because you are effectively blending two images together. Think about it. If you take two images with a moderate amount of noise and blend them together then the noise will go down. Ofcourse. That fact is also used in many noise reduction algorithms.

Maximum vertical resolution would be provided by weave. Especially for static video.

Bottom-line, when you add display variability into the different levels of Sharpness and Gain -- reports of noise simply can't be trusted.

I think that the biggest problem for lcd's is color reproduction, gamma and black levels. There is one setting that could possibly cause an "oil paint" effect and it's not deinterlacing but brightness. Some lcd's start dropping brightness digitally instead of lowering the backlight after a certain value and then we will see some shades going all wonky.

But that would cause an oil paint effect everywhere.

[Stephen van Vuuren]

Originally Posted by Steve Mullen

Really, only HDTV's with ATSC resolution (1280x720 or 1920x1080) connected by component or HDMI qualify. Otherwise, your display device becomes the weak link in the test.

I would disagree. Any professional, calibrated device hooked into the appropriate device and/or software should be able to view 100% of the signal (perhaps only a portion of the frame if at 100%). My fine pitch calibrated CRT computer monitors show far more of any image in both detail than any other device in my studio. Plus, if you know how to use NLE, scopes, channels and levels, you can diagnose most issues even on crappy displays.

I'm working on a 4K uncompressed 32-bit float/HDR project right now. I don't have a 4K Float/HDR display nor will budget for one. But even many high end studios don't either and you simply have the required to setup to look at portions and proofs as well as using analytical tools to verify your images.

More importantly, all the reported V1 artifacts are clearly visible on all my machines and displays, even my laptop LCD. Obviously lower contrast, uncalibrated displays will affect the level of artifacting you notice and in some shots it's more subtle, but it's there and in the video signal - just take this thread footage and channel separate it - you can see the various in the resolution loss in the color channels.

[Mikko Lopponen]

Originally Posted by Steve Mullen

2) on the static INT>PROG sample -- at the switch point the branches get noisey. In particular, the horizontal branches. A display should switch to "blend" on static video. When I switch VLC to blend, the noise is eliminated.

That's like saying "when I switch on noise reduction, the noise is eliminated".

No software dvd-players use blend. Powerdvd, windvd, theatertek etc, all use weave or bob depending on the situation. Weave for progressive, bob for interlaced. And there are very good reasons for that too.

[Steve Mullen]

Originally Posted by Mikko Lopponen

Maximum vertical resolution would be provided by weave. Especially for static video.

You are 100% correct. VLC offers "blend" which, for some reason, I took to be equivilant to "weave." That's the problem with software deinterlacers -- their terminology doesn't correspond to the same terminology used in hardware. For example, VLC offers "X" mode.

Weave, not blend, is correct for static video. Weave would not reduce noise, while blend would.

Sorry for the confusion I caused.

Of course, we still don't know what Sharpness Brett used for the "shed" clip. If it wasn't 5 then we don't have a clip of trees at 5. It would be nice if he would post what settings he used.

Guess I'll have to shoot some trees.

[Steve Mullen]

Originally Posted by Stephen van Vuuren

Any professional, calibrated device hooked into the appropriate device and/or software should be able to view 100% of the signal (perhaps only a portion of the frame if at 100%)


More importantly, all the reported V1 artifacts are clearly visible on all my machines and displays, even my laptop LCD. .

1) I've already removed the request for display info.

2) When you say "just take this thread footage and channel separate it -- you can see the various in the resolution loss in the color channels" you are confirming the issue we all agree exists: Softer luma and chroma in P-mode. Can you quantify the loss?

3) "All the reported V1 artifacts" is repeating the existence of artifacts without stating what "they" are and which clips show them.

The reason I'm being such a pain in the ass is that the specifics are critical as the issues with the V1 are very different than, for example, the Split Screen Effect with the HD100. With the HD100, there were dozens upon dozens of reports from all over the world. Many included screen-shots. The posts typically included specific gain levels. The problem was so well defined that we could simply say "SSE" and everyone knew what we meant.

We need well shot video and then screengrabs that show everyone the problem.

[Philip Williams]

Originally Posted by Steve Mullen

<snip>
The reason I'm being such a pain in the ass is that the specifics are critical as the issues with the V1 are very different than, for example, the Split Screen Effect with the HD100. With the HD100, there were dozens upon dozens of reports from all over the world. Many included screen-shots. The posts typically included specific gain levels. The problem was so well defined that we could simply say "SSE" and everyone knew what we meant.
<snip>

I think one of the problems here is that A/B comparisons are really where the issue becomes apparent.

Lets take that red shed as an example. If a guy videos that at 24P and looks at it later, he'll just assume that's the best image and detail the camera can produce. He might be a bit unhappy about some of the blockiness in some of the wood grain, and maybe assume that the camera wasn't able to record the fine details in the red siding on the front of the shed. He may not be that critical at all of course and be 100% happy with the footage.

BUT, if he shot it in interlaced and then progressive and A/B compared the shots, he'd see that some of the wood grain that turned to blocks in P is rendered in interlaced. He'd also see that there's all kinds of detail in the front of that shed in interlaced, but in P its turned into a blurry red wall.

So some of the progressive funkiness going on with the V1 is definitely not going to jump out at people like JVC's split screen. You definitly didn't have to A/B roll JVC footage where SSE displayed itself.

Anyway, I'm sure this will be clarified in more detail over the coming weeks.

[Brett Sherman]

Originally Posted by Steve Mullen

3) The three clips where you used 9 obviously have significant EE. And that raises the question, with the EE so boosted, it's possible that this may be introducing other problems. Frankly, I think these should be removed from your site because they were shot incorrectly. Yet they are being used for many claims. And, that raises yet another question -- for the INT>PROG sample -- did you use 5? Because if you didn't, then this too is a bad sample.

I'm convinced that there is no watercolor artifact. There's certainly nothing in hours of 24p India footage that looks like a watercolor artifact.

I think it is time to stop repeating the claim of a watercolor artifact in posts with no reference to any evidence. Repeating a false claim has a way of making it seem true.

I did make clear in a later post that I had the Sharpness on 9. However I realize there may be some confusion about the settings, so I'm posting those on the web page too. www.geekstudios.com/demos I'm not sure how you decided anything but 5 is "bad". Where does Sony say you should set Sharpness on 5? The default is 7. To me 5 is simply too soft, but I've never complained about the EE or noise at higher settings. I haven't yet arrived at my optimal sharpness setting. It will take more time and playing around. I think reasonable, talented shooters can arrive at different conclusions.

If you don't like my footage you can post your own. I never claimed that my footage was the end all and be all of V1 footage. It's just one sample. And I think it's useful for people to see.

I'm also not quite sure what you mean by properly shot. All of my stuff is shot with a tripod and properly exposed with 1/60th shutter. Also, I reject the notion that everything must be absolutely perfect for you to judge a camera's quality. Most people want to know what the camera will look like in a variety of situations. There are times when everyone will need to shoot with gain up and iris wide open, so it's important to know what it will look like in those situations. So long as it's disclosed what the settings are, I don't see what the problem is.

As far as the watercolor effect. Below is a cropped photo from the workshop scene in both interlaced and progressive that to me shows a watercolor effect. It was shot with sharpness at 7, Sony's nominal setting. Now you may call it "softer" and I call it "watercolor" or "paint." But there is definitely color banding going on in the progressive that is not happening in the interlaced. I make no judgement for other people whether or not this is acceptible for them. I just point it out. When I have time I will recreate the situation that convinced me I wasn't imagining things.

I'm generally pleased with the V1. However, I do have some disappointments with its progressive performance. It seems like Sony is making some false claims about the camera. On their website they claim it is sharper than competing cameras, however it is doubtful that that is the case in progressive scan mode. When I was deciding between the V1 and the A1, progressive scan was a factor. However, it appears that it should not have been, since A1s frame mode probably has the same or better resolution than V1s progressive mode. A native progressive scan camera like the V1 SHOULD be just as sharp in progressive as it is in interlaced. If there is some Noise Reduction going on reducing resolution, it would be nice to have some control over it.

Now I have no bias against the V1 in this discussion. I understand how our british brethren might by slightly biased against the V1 because of the trouble they went through. If anything, I'd be biased towards the V1, since I've bought one and have no options for returning it. In general, people like to think they made the best purchasing decision. Even if I had the option, I think I'd keep it anyways, since there are things I don't like about the A1 and I've bought a fair amount of accessories. Now I have to ask you, Steve, do you have any financial relationship to Sony? Including receiving free, reduced price or loaner cameras or equipment? If so, you should disclose that.

[Brian Rhodes]

Did any one notice the chromatic aberration on Steve's Pics 0-3. The food case and chairs have a purple outline.

[Todd Giglio]

Thanks for the images Brett. I can see the 'smoothing' effect in the p image on the side of the shed over the i, but I also think the roof over the door in the p grab seems cleaner with more defined detail (where the i image seems blurry or soft).

Over all, I'm in the same boat as you. I bought the V1U and I will live with it. This 'effect' doesn't really seem a major problem to me and I don't find it too distracting. It would be nice if the i and p images were identical (and maybe Sony will have a firmware update that can do this), but since I'm shooting with a 35mm adapter (DOF), I think I can work around these issues. Once I get a chance (I have two young boys and they CAN BE a handfull!), I'll try to shoot some trees/roofs and post A/B samples with 24p/30p/60i. I also have a Z1U and maybe I'll try to shoot the same images and post them too.

Todd

[Thomas Smet]

Originally Posted by Steve Mullen

It doesn't work that way. Anyone who claims something exists needs to prove it does. Simply repeatedly claiming you see "it" isn't good enough.

I made a very reasonable request:

1) provide a link to the clip (and to be a realiable claim, you really need multiple clips).

2) Provide a quote from the shooter that the P video was shot with Sharpness = 5 which is what Sony recommends. (This rules out 3 of Brett's clips.)

3) Describe what you see and where it is. Saying it's "on the church" isn't good enough.

In other words, you need to be as responsible as a professional reviewer would be if you are going to make claims about an effect that "... is there but is is very small and very subtle."

At this point -- not even ONE report of the alleged watercolor meets these standards for public verifiability. And, with each day going by with no new clips from the thousands of V1U buyers -- and Brett's clips disqualified -- those who make such claims need to "put up, or ... ."

In fact, if a problem is "very small and very subtle" and is reported primarily by those who don't even own a V1U and isn't reported by 99.9% of those who do own a V1U -- one could say there is no need for a debate. The issue is so trivial as to not be worth considering.

And, it's amazing that you have been knocking the V1U (based upon some compression theory) for 3 months -- claiming it won't be good for compositing and yet have never used a V1U. That's the classic behavior of a "troll."

Now I know you aren't -- and I think your compression theory is true. But to not test a V1U to verify your worries -- does seem a bit strange for a "pro" who was really considering buying a V1U.

Steve,

I would gladly rent a V1U if somebody near me rented it out but sadly that is not the case. Also all the projects I currently have do not require me to rent a camera so I'm not about to spend the money to do so until I have a need to.

I do not have any of my own shots from the V1U but I didn't realize this thread was only for owners of the camera. I do agree with you that it would be best to use the camera myself but at the same time I know when I see certain things. I'm sorry for the terms used here but like you I'm not really sure what to call it yet. Just because it doesn't have a proper name doesn't mean I am crazy.

I also started this in the hope of trying to find a way to reduce the effect. I do not think SONY would be that thick to design a camera that is flawed. In order to understand how to reduce the issue however we must know what it is and what causes it. I do not exactly know whats going on because I am not an engineer. That doesn't mean that I wouldn't like to figure it out so I can reduce it in the future if I do ever need to rent a V1U or buy one.

I agree with SONY that the sharpness needs to be turned down. Of Course I have no way to test this but I think if the sharpness is turned down most if not all of the problems will go away. I agree with one of your statements that it is kind of a tradeoff between sharpness and a clean image. This is pretty much thecase with every HDV camera to a certain degree. The V1U just has a different look when it is too sharp.

The V1U I'm sure has a certain default sharpness so most casual users will find the images to be sharp and detailed. That however may not mean the images are of the best quality. Yes reducing the sharpness makes the image soft but we must remember that all HDV cameras are using chips of not so optimal resolution. SONY came up with an interesting way to sharpen the images from it's 960x1080 chips without making it look too much like EE.

For compositing I have always used even SD cameras at greatly reduced sharpness. This really helps reduce the EE and makes for a much cleaner key. I'm sure the V1U is no different it is just that one camera has EE while the other has funky artifacts. My bet is that if you reduce the sharpness to 5 or maybe even 3 all artifacts will go away and the image will look perfectly clean.

One test I like to do with the church shot is to use a crop to make it SD and then watch this SD cropped version on a HDTV or computer monitor. In this case you can clearly see some dancing shimmering details. I know most people don't crop to SD but some people do and a lot of people use the Z1 for this type of shooting.

I would love to see some progressive footage with reduced sharpness to see if that helps clean up everything.

[Stephen van Vuuren]

Originally Posted by Steve Mullen

1) I've already removed the request for display info.

2) When you say "just take this thread footage and channel separate it -- you can see the various in the resolution loss in the color channels" you are confirming the issue we all agree exists: Softer luma and chroma in P-mode. Can you quantify the loss?

We don't have footage yet to quantify - just verify.

3) "All the reported V1 artifacts" is repeating the existence of artifacts without stating what "they" are and which clips show them.

In this thread, detail loss and gain in different areas of image with Brett's footage. In Tony's images, his footage. etc.

The reason I'm being such a pain in the ass is that the specifics are critical as the issues with the V1 are very different than, for example, the Split Screen Effect with the HD100.

We are in the same stage as early reports of HD100 issues - but while skepticism is always a great attitude, the evidence is in. I recall your responses to the first SSE reports and often you blame the shooter as the cause despite footage to the contrary. This attitude accomplishes nothing positive IMO and is in conflict with your claimed skepticism. If you argue there is not enough evidence for or against the existence of camera issue/defect, there is not enough evidence to point to shooter "skill" as the cause for the observed issue.

We need well shot video and then screengrabs that show everyone the problem.

While there is plenty of video & grabs to indicate a clear problem, I agree that footage shot in controlled fashion with the goal of shedding more light on the V1 would help. That's why I have requested it (and hopefully the V1 shooter down the road from me & I can get together to make it happen).

[Tom Roper]

Originally Posted by Steve Mullen

If DNR has removed detail in an area -- and if the monitor uses "bob" deinterlacing -- the paint effect will be increased. Since the majority of HDTVs and softer de-interlacers use "bob" -- I'm not surprised some are seeing a more of a problem than others.

The majority of HDTV monitors do not use "bob" they use vertical filtering which weaves together the two fields but filters or softens the whole image, even the static part to lose the combs on movement. The softening is inherently not good and ghosts but it's hardly noticed usually because the video is downrezzed to the native 720p resolution of the panel.

The more egregious sin is committed when the technique is used on 1080p monitors.

The vertical filtering should not cause blurring in specific areas of the screen, but rather throughout the whole of it.

A more desirable deinterlacing technique is the per-pixel motion adaptive which weaves the static parts and bobs (interpolates) the moving parts.

The most sophisticated method is direction vector which weaves the fields, but horizontally shifts areas to null the combs in the moving areas without softening. It's used in high end video processors.

I think the only HDTV monitors to use "bob" were the 1080i rear projection crt's that would display 720p by scaling the frame to 1080, then bobbing it to 540i.