V1U grabs
 

I'll post some footage up in the next few days. But for now here are one or two grabs. Not exactly spectacular subject matter though, sorry!

The first is a shot from my driveway. I didn't take any time to set this particular shot up because I just wanted a quick comparison of progressive scan vs interlaced. The results aren't pretty! Take a look at the house roof, trees, and car number plate in the background and compare the two.

http://www.simonwyndham.co.uk/Sony%2...Interlaced.png
http://www.simonwyndham.co.uk/Sony%2...rogressive.png

http://www.simonwyndham.co.uk/Sony%20V1U/trains.png
http://www.simonwyndham.co.uk/Sony%20V1U/Person.png
http://www.simonwyndham.co.uk/Sony%20V1U/Cranes.png
http://www.simonwyndham.co.uk/Sony%20V1U/Clifton_2.png

The grabs look promising. Can't wait to see your footage!

The camera certainly can handle a wide dynamic range.

I see what you mean about the progressive, there is quite a noticeable drop in resolution.

I'd be interested to know just how much in-camera control one has over the image if you have the time.

Many thanks Simon.

TT

I don't know what to make of these, other than to say they are interesting travel photos.

In other words, hard to judge without something side-by-side. The fringing clearly remains. I'm inclined to say the detail is better. I'd like to compare a few of my Z1U grabs to these. The images have a slightly processed look.

What were the in-camera sharpening settings?

Pretty country...

Sharpening settings were left at default. For once I felt that the default settings weren't intrusive, so I left them as is.

There is some fringing. Although there doesn't appear to be as much as other cameras and lenses that I have seen.

Dear Simon, I think that the very noticeable resolution drop in the progressive vs interlaced shots is all so important to us potential buyers - if the progressive mode is just a gimmick, claiming to give the "film look" but at the expense of resolution that evident - many will be better off to go for FX7 rather than the V1 (cheaper, and still offering the revolutionary exposure latitute). So, are you SURE both clips were shot in the same camera settings? No AF?

BTW, some CA is also noticeable in both interlaced and progressive:-(

Piotr, I make a living from video, so getting the settings exactly the same in such comparisons is something I make sure that I do. I never use auto functions on a camera ever. So rest assured, both those grabs were from EXACTLY the same camera setup. All I did was change between progressive and interlace. This is not an isolated incident

You will notice that the problem isn't simply one of lost resolution. Detail takes on a really weird effect that looks very similar to one of those filters in Photoshop that makes a picture look like it was painted.

One thing I didn't try was turning the camera off and then back on again. On the pro cameras Sony recommends switching the camera off for 10 seconds after switching modes. Although I have never come across this type of problem with the picture before.

Regarding CA, this will always be apparent in most cameras. In fact I haven't come across one that doesn't exhibit this, including high end broadcast zooms.

thanks for posting that Simon. interesting shots!

just one thing... surely living in PAL land, this cam was the V1E and not the V1U? The thread title says "V1U".

Simon,

What process did you use to take the grabs? Could the grabbing itself have affected the output? It's hard to imagine how the image could degrade in progressive mode on a camera that has truly progressive chips :(

The person who compared the V1 to his JVC (the PAL version of the 100) said the progressive mode looked a little softer.

DSE's paragiding clip at 24p did not show those "oil paint" style artefacts but that was held in a 60i stream. I wonder if 30p softens?

Could Sony be crippling progressive recording? I ask that because those grabs of 50i are fantastic and if the same resolution was available at 25p Sony might think it was a little too good at the price point. Is that just stupid conspiracy talk???

Or is there an incompatibility with capture equipment causing the loss of resolution? How should the 25p captured? Would one capture the 25p footage as 50i? Having looked at the grab in Shake, the progressive image looks like it has been subjected to a quite severe noise reduction routine. Is there a noise reduction setting engaged in progressive mode?

I really do like these images, really quite un-Sony like...

Anyone?

TT

The grabs were taken direct from a Vegas 7 timeline. All project properties and clip properties were set as appropriate to each clip to make sure they were shown at their absolute best.

Trust me, the progressive mode IS softer and less detailed than interlaced. I viewed images on a 32" high def CRT and the difference was noticeable switching between the two modes.

Rest assured I'll be getting as much info as I can from Sony regarding this.

Interesting...

I'm not sure what to think of these at this stage. They look clean and more photographic in the areas that are exposed well but the blown out areas sure do look fake to me with very high contrast edges. Actually all of the images look kind of soft to me. II would have thought your detail level was set to min and not the default level.

From these images I would have to say they look more like the HVX-200 with dull colors. They look very natural but very soft as well kind of like how the HVX-200 looks. The progressive image is clean but I see exactly what you are saying about it being softer.

I would also have to say that a lot of these stills look a lot like the footage from my SONY HC1. They colors are more accurate from th V1 but the overall soft natural look looks kind of close to me. This is not in any way a bad thing. I happen to like this type of look and think a natural look is much better than resolution.

With that said however these shots seem to be missing that sense of texture you get with HD footage where you can feel the grain in the wood or a roughness to a rock or stone wall. I will send these images to a HDV tape and see how they look on my 50" HDTV.

I am satisfied that Simon is careful and precise in his methods. The grabs are illustrative and representative of the V1E. There remains some doubt about it being exactly alike with the V1U. I see pretty good detail in these images.

Thank you Simon, nice work.

Of course the native images from the m2t stream are 1440, these are upsampled to 1920, presumably so they don't display squashed, but some interpolation has been added in that transformation.

Using Shake to deinterlace, apply a gaussian blur (2px) and then applying a median filter to the interlaced image I can get a reasonably close approximation to the progressive frame.

I'm finding it difficult to see any benefit in the 25p mode over a high quality deinterlacer and 50i footage.

I'm interested to know what Sony have to say.

TT

Just as they will be when displayed on a 1920x1080 screen anyway.

The scaling is only horizontal and doesn't change the effect that I have illustrated.

The same upscaling can be said for the interlaced video as well since it also uses 1440x1080 pixels. Both progressive and interlaced would be recorded on tape as 1440x1080 so using the argument that the images were upscaled is not a valid point at all since they were both uprezzed from 1440 to 1920.

I cannot for the life of me figure out why a camera that is supposed to run at a native 1920x1080x60p has a progressive frame with a softer image. The only thing I can think of is a low pass filter that they figured was needed because extra high detail could cause some flickering due to recording progressive video in a interlaced stream. Since most of the time the video would be played back as interlaced they wanted to avoid any flickering issues that may have come up with really high detail. I would have thought the same level of filter would have been used on interlaced as well since the progressive source in the DSP is still the same. With interlaced the I guess the low pass filter could be lower since you may not get the same level of flickering with interlaced video since it is meant to be shown one field at a time anyways.

I would really like to see more samples of the same exact locked down shot recorded as interlaced and progressive. Perhaps a landscape shot with lots of small details in the distance. This will really show us how much softer the progressive really is.

I did see a 'flicker reduction' option in one of the menus, but I had it switched off (I couldn't see any discernable effect when I tried it)

The thing is that people are commenting that the progressive image is slightly softer than the interlaced one, but seem to be missing the point that there is that weird 'paint effect' thing going on. That bugs me much more than any softness.

I do see the softness in Simon's P grabs.

These are from one sample of the V1E.

We will not get production models until December. So you've got to ask -- given there will be no more than a dozen or two V1's in the USA for over a month -- how can anyone in the UK have a "production" model now.

My guess this is a communication error by Sony UK and this is one of the IBC test models that may not have the lastest firmware. There was a firmware update after IBC.

Remember the SSE in the HD100 that was released in Europe before the USA. Clearly, JVC Europe was selling camcorders that were not as fully "finished" as the ones we got 45-days later in the USA.

I measured no difference nor could I see ANY difference between I and P directly feeding a monitor via HDMI. Nor, could I see any difference in recorded test shots I made after reading about the difference. And, I can't think of any process reason for any difference.

DSE -- do you have I and P shots and rez tests. It would be nice if HDV buyers didn't have to go through an SSE type scare again.

Like Steve suggests, maybe there is something wrong with Simon's particular camera. Ironically, Mikko's V1E footage from IBC doesn't seem to show any paint type effects on 25p. I also don't really see any resolution drop in Mikko's footage between 50i and 25p.

The thought that it might be something to do with firmware etc occurred to me. I hope you are right!

Simon's footage looks like there's some kindof noise reduction going on in the progressive frame. That's why it looks kinda like a painting (smoothing algorithms tend to do that) and has less detail.

The progressive pic has more details in certain places (tire rims but not much) but less on smoother surfaces which seem to indicate some sort of noise reduction going on. Why?

I've did some CC to have some idea of latitude, it's also deinterlacced:
http://xs107.xs.to/xs107/06412/Person.jpg

I can't see this. Take a look at the car tire in the foreground. You will see on the interlaced version that the tire has a wafer of rubber sticking out of the tread. In the progressive scan image this is completely gone.

I agree that this does look like extremely harsh noise reduction.

My recollection was that the article clearly stated the V1E was a pre-production model.

I do, but they're for a magazine article due out next month, so it's not appropriate for me to post them publically.
What Simon is showing isn't quite my experience, and apparently not yours, either. I dunno what the state of Preproduction V1E vs V1U is, but...when I received the prepro V1 that I had, I also saw 3 others in various states of development, all significantly older. So...it could be that Europe has the same thing.
for me, I shot cascades on a canopy in both P and I, and there is no discernible difference on a 1920 x 1080 monitor, both at 12' and 20" displays.

Progressive video is often filtered (DVX100: Thin & Thick modes) and it's possible the wrong parameter value was set for the V1E.

He said that he got the image captures out of Vegas. Could Vegas be treating the progressive stuff as interlaced and doubling fields, since its encoded in interlaced stream? It would be nice to know the actual vertical and horizontal resol of what he gots. I also noticed that the images are not 1920 by 1080 so maybe the down cropping filter was lossy.

Vincenzo

Off-topic but useful Shipping info : Production V1E's are due to be in 1st customer's hands in UK during 1st week of November. So 4 weeks time, more or less.

By the way, the FX7 is delayed (for sure in Europe and i'd imagine USA too)and is currently running about 3 to 4 weeks behind the V1, so i hear.

Draw your own conclusions why Sony appear to have 'reversed' their normal release of a new cam, with the 'pro' model coming out *before* the 'consumer' model..

I guess I'm blind - on A/B switching the progr looks slightly sharper to me and since it's a static shot, that's to be expected.

But since the shots are not at the same moment in time, various factors could be affecting the results much less firmware.

What's more exciting is the nice latitude that keeps showing up in these shots - very impressive indeed.

To me, this looks IDENTICAL to Canon's "F" mode. You can only really discern it from grabs, much harder to see in real motion. Same is true of the HVX as well. Something about the motion seems to mask the resolution loss.



ash =o)

Why would there be a resolution loss going from interlaced material to progressive? I understand why the Canon F mode is slightly lower in resolution because it is being derived from an interlaced CCD. Since the V1 has CMOS chips that are inherently progressive it must take a special type of balls up!

Did anyone record 30p in the US? If so how did that compare to the sharpness of 60i?

No difference. Strobing (less) and NO judder frames. 30P was definitely nicer. But, we haven't seen 24p without pulldown.

No. As I said at the beginning of the thread Vegas was set up correctly for each shot (ie project properties upper field first, NO deinterlacing, clip properties set to upper field first for the interlaced shot, and project properties progressive scan, and clip properties progressive scan for the progressive shot). I know how to use Vegas thanks. :-)

I think you're blind! ;-) Either that or you've muddled them up.

Quite how you could say this when A/B switching them I don't know. Look closely at the tire detail in the foreground for one. There is one HUGE detail missing from the progressive scan image that seems to have been filtered out (a piece of rubber hanging out of the side near the top of the tire). Look at the drive detail. In the progressive image there's huge swathes of detail missing. Yes, actual detail. This isn't about softness or sharpness, its about actual detail in the picture.

As Steve Mullen mentioned, it could be that the filtering is set incorrectly.

Simon, the progressive is in fact so much inferior, there has to be some reason other than the way it really works in the V1E - the difference is just too big to be true.

Or should I say - different? Doing away with this tyre piece completely must be some filtering in action, not just softening/resolution loss.

Now I must say I'm completely lost: I cut out small windows around the tyre detail from both the interlaced and progressive grabs, saved as separate files - and guess what! The rubber detail, so obviously missing from the progressive frame when displayed in whole, is back there!

? Static shots should have the same resolution. Moving shots would give the resolution advantage to progressive.

I don't think it's as simple as that.

I believe that with interlace, sensor rows are "paired" to improve sensitivity and reduce twitter. Such that even field lines may be made from sensor lines 1+2, 3+4, 5+6 etc, odd field lines from 2+3, 4+5, etc.

My question to Simon is what do Sony have to say about all this? Until we get an answer to that, speculating about technicalities is all very well, but it may just turn out to be an early (and unrepresentative) individual model.

With all due respect, with most 24p footage in Vegas, field order is set to "none". With HDV, interlaced footage is set to "upper field first" but with this cam I assume you would also want none though progressive HDV support for this cam is not yet built into Vegas which may be the isssue here (incorrect pulldown removal).


I did not have them muddled up but to me it looks like the depth of focus has changed - very close and far looks softer but in focus looks slightly sharper.