New Sony PXW-FS5

[Christopher Young]

Originally Posted by Ron Evans

Right now the DVX200 looks to be a better choice.

Ron ~

Right now I'm holding my powder dry after spending an afternoon with the DVX200.

The one we looked at was the latest firmware version prior to release we were told. First things first. I had heard about noisy shadows and blacks so I was on the lookout for this. Sure enough I could see noise in the dark areas at 0dB. What I DIDN'T expect to see as I wasn't even looking for it and it was the furthest thing from my mind was fixed pattern noise in the blacks. To see pronounced FPN at 0dB on a late model camera release surprised me. Something else I didn't expect to see and wasn't even thinking about looking for it until it hit me was moire... moire on T shirt material even? These three issues all raised their heads in the first thirty minutes of just trying the camera out without any specific tests in mind.

The other aspect I found a little surprising was that I was led to believe that the 200 had a similar zoom function to Sony's Clear Image Zoom. A zoom that once it reached the end of its optical travel would then continue on from the long end. On the x70 for example you go from full wide to 12 x and then on to 24 x with pretty acceptable results. On the 200 this is not the case it is actually a doubler as we are used to in B4 lens terminology. The HD wide end is 28mm in FF terms so this becomes 56mm and the long end goes from 365.3 to 730.6mm. So no zooming from full wide of 28mm with the 2x engaged. I can see uses for this but the loss of the wide end really cramps ones style if you want the extended reach at the same time.

In all fairness to Panasonic this was a pre-production model so I will wait to see a full production roll out model before I draw any further conclusions. The DVX200 camera has masses of appeal on paper with the specs it offers but unless the image characteristics improve somewhat I think there will a few people asking questions. No doubt more and more opinions will start to surface in the near future as these cameras hit the stores. I'm just surprised to see to see these issues raising their heads so close to release date.

Chris Young
CYV Productions
Sydney

[Christopher Young]

Originally Posted by Jamie Roberts

I'm tipping many buyers will also be shelling out for a new computer in order to edit the footage the FS5 produces without pulling all their hair out!

Jamie I know where you are coming from. Long GOP formats like XAVC-L are always harder for CPUs. to decode. What I am doing with all my XAVC footage is batch converting it to Cineform 4:2:2 Medium which is a wavelet based codec. It plays super smoothly edits nicely and can carry a fair bit of correction and effects before it even looks like dropping below a full 50 frames per second and that sort of performance can be seen on a 1st gen i7 CPU. The trade off is that the clips will be about 3.6 times the size. Quality? I cannot pick the difference so it's a no brainer for me.

Chris Young
CYV Productions
Sydney

[Tom Gresham]

Any updates on delivery date? With two on order, I'm counting the days. I just don't know how many days to count. ;)

[Terence Morris]

I'm excited about this camera more for HD at 4:2:2, 10-bit with an efficient codec onto cheap XCSD cards off a super 35mm sensor, plus interchangeable lenses. Don't really care about 4K just now, but love the idea of having more "cinematic" look potential and more robust post-production tweak-able footage - all in a gun-and-run format. Kind of like a PXW X70 with a bigger chip AND I can swap lenses. I don't think anything else out there ticks all of those boxes. 240 FPS slow-mo is icing on the cake really. Will seriously consider buying this when available for cinema verite type projects.

[Nick Fotis]

Maybe a stupid question, but... would it be compatible with the HXR-PMU128?
If yes, that would be a boon (and would permit high rate 4K recording)

N.F.

[Terence Morris]

Originally Posted by Terence Morris

I'm excited about this camera more for HD at 4:2:2, 10-bit with an efficient codec onto cheap XCSD cards off a super 35mm sensor, plus interchangeable lenses. Don't really care about 4K just now, but love the idea of having more "cinematic" look potential and more robust post-production tweak-able footage - all in a gun-and-run format. Kind of like a PXW X70 with a bigger chip AND I can swap lenses. I don't think anything else out there ticks all of those boxes. 240 FPS slow-mo is icing on the cake really. Will seriously consider buying this when available for cinema verite type projects.

I guess I rather irreverently spilled my guts, but would anyone care agree or disagree with my take on this camera. Thanks.

-Terence

[Mike Watson]

Originally Posted by Tom Gresham

Any updates on delivery date? With two on order, I'm counting the days. I just don't know how many days to count. ;)

Eric from Pro Video and Tape said I was one of the first few on the pre-order list, so I assume I'll be one of the first to get one. Like you, I am chomping at the bit. They were saying shipping in November but I'd heard mention of "perhaps by the end of October", so I have my hopes up. My beta 300 was serial number #10012, the 12th Beta 300 to roll off the line. I am curious to see if I can have a similar accolade for the FS5, although certainly it won't be as iconic.

Originally Posted by Nick Fotis

Maybe a stupid question, but... would it be compatible with the HXR-PMU128?.

The FMU128 would require molding of the body that the FS5 does not seem to have.

Originally Posted by Terence Morris

I'm excited about this camera more for HD at 4:2:2, 10-bit with an efficient codec onto cheap XCSD cards off a super 35mm sensor, plus interchangeable lenses. Don't really care about 4K just now, but love the idea of having more "cinematic" look potential and more robust post-production tweak-able footage - all in a gun-and-run format. Kind of like a PXW X70 with a bigger chip AND I can swap lenses. I don't think anything else out there ticks all of those boxes. 240 FPS slow-mo is icing on the cake really. Will seriously consider buying this when available for cinema verite type projects.

If you had hard drive space to burn, I suppose you could shoot HD 4:2:2 10-bit - I'm showing you'd get about 7 minutes on a 64GB card. Can your edit system handle this bandwidth? What do you have for a monitoring solution that can take advantage of this extra data? How are you outputting? Will your audience notice a difference? To me, this is not the advantage of this camera. But I am not the camera gestapo - but it for whatever reasons you wish.

[Terence Morris]

Quote: If you had hard drive space to burn, I suppose you could shoot HD 4:2:2 10-bit - I'm showing you'd get about 7 minutes on a 64GB card. Can your edit system handle this bandwidth? What do you have for a monitoring solution that can take advantage of this extra data? How are you outputting? Will your audience notice a difference? To me, this is not the advantage of this camera. But I am not the camera gestapo - but it for whatever reasons you wish.

Regarding the card space, XAVC-L compresses HD to 50 MB/s - so that's not a problem. Pro Res conversion (for example) will blow that up quite a bit, which I think is what you maybe were referring to (7 minutes/64 GB)? I use external hard drives for all my projects, which are relatively cheap enough now. Regarding the quality, you may be right that most of the audience will not notice. But I do, and there is no rationalization beyond that it bugs me. I'd put up money for competing in blind trials to pick 4:2:0 over 4:2:2. Poor me - I could save so much money. Having the X70 has been a kind of vindication. No more greenish greens, reddish reds. And I have much more latitude in post with 10-bit vs 8-bit. I mainly deliver for internet or small screen charity events, so no great shakes, but I much prefer 4:2:2/10bit to work with. Horses for courses I guess.

[David Heath]

Originally Posted by Terence Morris

Regarding the card space, XAVC-L compresses HD to 50 MB/s - so that's not a problem.

It's 50 Mb/s, NOT MB/s - a very big difference! :-)

Originally Posted by Terence Morris

I'd put up money for competing in blind trials to pick 4:2:0 over 4:2:2. Poor me - I could save so much money.

Well, I'd be happy to take your bet! Of course, it depends on what the trial is, but on straightforward shooting/viewing - even on a very high quality monitor - I would be very, very surprised if anyone could tell the difference without extreme pixel peeping.

And to an extent depends what the format is. The whole rationale behind the existence of 4:2:2 stems from interlace, and in such a system there are indeed valid reasons for not subsampling vertically, at least at the acquisition end of the chain.

For a progressive system all that goes away and it is far more sensible to have a symmetrical system - 4:2:0 or 4:4:4.

Originally Posted by Terence Morris

And I have much more latitude in post with 10-bit vs 8-bit. I mainly deliver for internet or small screen charity events, so no great shakes, but I much prefer 4:2:2/10bit to work with. Horses for courses I guess.

Ah, now 10 bit is a different matter. For most material 8 bit is fine, but it's when you get into the world of log curves and such that 10 bit really gets important.

It's also worth saying that 4:2:2/10 bit codecs have traditionally gone hand in hand with higher end codecs with higher bitrates, so less overall compression - and the latter can make a huge difference to how material may stand up in post. But such doesn't have the "headline" nature of "4:2:2", even though practically it may be the compression factor that is really making most difference.

[Andy Wilkinson]

New Sony FS5 footage...

VideON | First Look At What The PXW-FS5K Can Do | XDCAM HD Products

[Terence Morris]

Originally Posted by David Heath

It's 50 Mb/s, NOT MB/s - a very big difference! :-)

Well, I'd be happy to take your bet! Of course, it depends on what the trial is, but on straightforward shooting/viewing - even on a very high quality monitor - I would be very, very surprised if anyone could tell the difference without extreme pixel peeping.

And to an extent depends what the format is. The whole rationale behind the existence of 4:2:2 stems from interlace, and in such a system there are indeed valid reasons for not subsampling vertically, at least at the acquisition end of the chain.

For a progressive system all that goes away and it is far more sensible to have a symmetrical system - 4:2:0 or 4:4:4.

Ah, now 10 bit is a different matter. For most material 8 bit is fine, but it's when you get into the world of log curves and such that 10 bit really gets important. ----agreed, but I can still find I can push around the x70 XAVC-L 10-bit image much more, even in "standard" settings, with less chance of it falling apart.

It's also worth saying that 4:2:2/10 bit codecs have traditionally gone hand in hand with higher end codecs with higher bitrates, so less overall compression - and the latter can make a huge difference to how material may stand up in post. But such doesn't have the "headline" nature of "4:2:2", even though practically it may be the compression factor that is really making most difference.

It's 50 Mb/s, NOT MB/s - a very big difference! :-)

---I actually know this, so hold my head in shame, and thanks for the correction; see what occurs when typing something late into the evening after a couple of glasses of merlot.

Well, I'd be happy to take your bet! Of course, it depends on what the trial is, but on straightforward shooting/viewing - even on a very high quality monitor - I would be very, very surprised if anyone could tell the difference without extreme pixel peeping.

----Well, if there is ever a world tournament :-)

And to an extent depends what the format is. The whole rationale behind the existence of 4:2:2 stems from interlace, and in such a system there are indeed valid reasons for not subsampling vertically, at least at the acquisition end of the chain.

For a progressive system all that goes away and it is far more sensible to have a symmetrical system - 4:2:0 or 4:4:4.

----I don't pretend understand much of that in any depth (but promise to educate myself accordingly) - and with respect, I do trust what my eyes tell me; whenever I have seen footage 4:2:2 vs. 4:2:0 even crushed somewhat on Vimeo, it positively pokes me in the eye. Or I could be seriously deluded. I do intend to do a series of blind comparisons on myself, since I respect your expertise. (I also have perfect pitch apparently, and a slightly mistuned piano causes much pain while every one else merrily bobs along oblivious to any discord.)

Ah, now 10 bit is a different matter. For most material 8 bit is fine, but it's when you get into the world of log curves and such that 10 bit really gets important. ---agreed, but I can still find I can push around the x70 XAVC-L 10-bit image much more, even in "standard" settings, with less chance of it falling apart.

It's also worth saying that 4:2:2/10 bit codecs have traditionally gone hand in hand with higher end codecs with higher bitrates, so less overall compression - and the latter can make a huge difference to how material may stand up in post. But such doesn't have the "headline" nature of "4:2:2", even though practically it may be the compression factor that is really making most difference.

----Thank you for your guidance. I have relied on DVinfo so much in the past to set me right, including your good self David, but I think I would to enjoy this camera, if only for the super 35mm sensor, 240 FPS slo-mo and swappable lenses. Fool and his money etc etc.

[Andy Wilkinson]

This was posted yesterday on Vimeo by someone in Italy. Sony FS5 on a Ronin M. He says it was shot in SLog2 (not yet sure if it was 1080p 10-bit or QFHD 8-bit).

Sony PXW-FS5 very quick test with Ronin M on Vimeo

[Andy Wilkinson]

I see this has now become a private video.

[David Heath]

Originally Posted by Terence Morris

----I don't pretend understand much of that in any depth (but promise to educate myself accordingly) - and with respect, I do trust what my eyes tell me; whenever I have seen footage 4:2:2 vs. 4:2:0 even crushed somewhat on Vimeo, it positively pokes me in the eye.

But the question I have to ask is have you ever seen a 4:2:0 v 4:2:2 comparison *with all else equal*? Same compression factors, obviously same camera ( :-) ), same everything else? Because if the answer is "no", then whilst you may indeed have seen a difference, then how can you definitively say it was due to colour space and not one of the other factors?

This is relevant as typically 4:2:2 may be more likely to come with lower overall compression and other "good" factors. So that codec may well show a real benefit, but not necessarily down to colour space. It's a bit like going to my favourite restaurant, who (unknown to me) buys all his lamb from a local high quality source which he has a good deal with, and all his beef from a more anonymous source. I don't eat meat anywhere else, so my conclusion is that "beef is a better meat than lamb". It's valid on the basis of my experience....... but.... :-)

And - was what you saw interlace or progressive? It is undoubtedly true that 4:2:2 is more suitable in the interlace world, though I'd expect that to be more the case further down a chain than on source material.

You have to ask yourself why, if it's OK to halve chroma resolution horizontally, then why is it not OK to halve it vertically? Because that's what is implied by demanding 4:2:2. And the answer, quite simply is interlace. Halving the samples is a bad idea when you think of the line structure of interleaving lines of the alternate fields. Move to progressive imagery and all that goes away.

Originally Posted by Terence Morris

........ but I think I would to enjoy this camera, if only for the super 35mm sensor, 240 FPS slo-mo and swappable lenses. Fool and his money etc etc.

Don't get me wrong, I'm not trying to put you off this camera, I've seen it and basically was pretty impressed by it. But the 4:2:2 aspect is pretty well down my list of reasons for thinking that....... It's certainly good if you're thinking of shooting HD interlace, but I can think of many more important factors to buy it for.

(Actually, I think I'd still overall go for the FS7 personally, but it's horses for courses and I do like the variable ND of the FS5. Very impressive. But if sheer small size and the ability to strip it down even smaller is of key importance then yes, it has to be the FS5.)

[Terence Morris]

David - You have given me a lot to think about. I do appreciate you taking time to enlighten green-horns such as myself. There is probably some definitive Digital Video Engineer's Handbook I should be consulting, although Wikipedia is pretty good for most things. Anyway, I foresee an interesting learning curve ahead. The bottom line is, no, I haven't done a controlled comparison with all parameters being equal, as best as that may be possible to achieve. And I had not even factored in interlaced vs progressive, but I see your point in light of the way horizontal rastering works. Hmm...

Thanks again,
Terence