F35: A few questions from a new owner

M

MikalZ

Well-known member
Hi there - I've recently put together a very minimalist F35 rig, and I should say upfront that getting these beasts up and running would not be possible without the contributions from the pioneers on this board. :thumbsup:
My aim with the rig is to be as stripped down as I could possibly make it, no interface box, no 4:4:4, just the brain with a plate and battery shooting 10bit.

Without having recorded anything but a few test clips at this early stage, I had some initial questions:

1: The Atomos I'm recording to is showing the input as 23.97 with a 2:2 pulldown. Is this to be expected, or is there is a better way to send real progressive 24p out? I'm in CINE mode.
2: I've run the auto black balance but still have a stuck pixel. What are the strategies here...just continue running the balance until it goes away?
3: After turning off the VF menu I've still got a safe area type overlay that's making it's way onto my recording (again no IF box in my setup). What's the best way to turn this overlay off?
4: Any other general tips for someone dipping their toes in here?

Thanks!
 
I don't say it to be rude, but reading the operators manual is very informative. It's a very quick Google search.

Unless your primary output is theatrical there's no good reason to record true 24p. 23.98p is far more useful for most destinations. I'm not sure if that's what you were asking, or if we're just discussing the 29.97 pulldown to 23.98. Knowing the frame rate setting of the camera itself would be useful in this context as well. The body SDI ports are considered monitor ports, so they may not support true 24p (though they should support 23.98).

Under some circumstances stuck pixels aren't fixed by a black balance. So is life. First, update to latest firmware, and then follow the RPN procedure, as outlined here: http://www.dvxuser.com/V6/showthread.php?325895-How-to-Fix-Stuck-or-quot-Hot-quot-Pixels-on-the-F35

Finally, if you get to post without catching hot pixels, you simply need to mask that hot pixel with interpolated data from surrounding pixels. There are a variety of ways to do this depending on your software package. Fortunately, for any given camera, the hot pixels stay fairly consistent, so you can reuse the filter/plugin on all the footage as needed.

The overlay is considered a marker. It's controlled in the Operations Menu, on the Marker Setting page. Look for the SAFETY settings. (as in, "safe area")

Tips: get an IF box. It'll save you from costly blunders.
 
Viddovation - not rude at all, and your response is appreciated (in particular the Marker settings location). I have gone through the manual, but as I'm sure you will attest from your first days with the system it's rather cryptic to parse. I'll be combing over it a few hundred more times I'm sure.

Very good point on the monitor outs possibly not supporting 24p, in which case I won't worry about it.

I'm going to give auto black a few more attempts before attempting the dreaded RPN procedure....masking in post is not ideal.

And buy an IF box you say, as if that is a trivial thing to find on it's own :P
 
Quick update: after a few consecutive ABBs, the stuck pixels look to be gone. And with the safe area overly off (thanks again), I'm recording wonderfully gritty F35 footage.
Also, looks like the monitor outs can send 24p no problem.
 
Trust me, I was just as surprised as you will be once you realize it's true: The F35 will not output 24P or 23.98P, it is always PsF. If you want to be certain, look at the video info in quicktime.

Look around for a bit you will find a IF box for around 1k. I have purchased 2 additional boxes since I originally bought my F35.

I completely understand going the light easy route, that has always been my intent with my F35. I would say the F35 with 7Q up to 50P is just too damn versatile, and this will always be my F35 recorder due to the support CD has put forward making the F35 formats work. It is very exciting that firmware updates are still improving the capability of this camera.

Hunt around for a bit and you will find a 7Q without sony license for around $1500 used. Now they can use Samsung drives, although I just bought 2 more 512gb drives for my 7Q for 330 a piece.

There are deals to be had, in the mean time keep shooting and learning the F35 system.
 
yachacha said:
Trust me, I was just as surprised as you will be once you realize it's true: The F35 will not output 24P or 23.98P, it is always PsF. If you want to be certain, look at the video info in quicktime.

Look around for a bit you will find a IF box for around 1k. I have purchased 2 additional boxes since I originally bought my F35.

I completely understand going the light easy route, that has always been my intent with my F35. I would say the F35 with 7Q up to 50P is just too damn versatile, and this will always be my F35 recorder due to the support CD has put forward making the F35 formats work. It is very exciting that firmware updates are still improving the capability of this camera.

Hunt around for a bit and you will find a 7Q without sony license for around $1500 used. Now they can use Samsung drives, although I just bought 2 more 512gb drives for my 7Q for 330 a piece.

There are deals to be had, in the mean time keep shooting and learning the F35 system.
Click to expand...

yachacha - The monitor outs are always PsF?? This is really interesting and I've not read this previously. I've been wondering why my Atomos is always showing 2:2 pulldown, even when I switched to 24p; and this might explain it.
If this is true, what is the downside that you have found to shoot footage this way? The footage looks fine when played back...
 
Yes, it is always pSf out of the monitor out, I have tested every possibility. 50P out of the IF box is true progressive. There is no difference although if the pSf is handled incorrectly on export from something like final cut, the footage becomes grainier. I know there have been a few conversations about this, I am not as good at finding references as Viddo.

Sony uses a unique pSf to send the 444 signal over two sdi lines, this is why 7Q can see the dual link correctly and others might not be able to.

The Aja LUT box is another device which will remove the pSf and convert to true progressive.

I usually run footage through compressor and convert to true progressive before I edit.
 
PSF is "progressive segmented frames"

It's progressive data in an interlaced container stream. Think of the situation for video production facilities say 10 years ago, with hundreds of thousands if not millions of dollars just spent on brand new high definition wiring, decks, cable runs, routing and so forth. Imagine that via routing you can have engineering send a desired video signal from any place in the building to any other point. Imagine that all of your very expensive equipment only understands interlaced SDI streams - not progressive. If you try and send a "true progressive" SDI stream, it's going to fail to route, because the intermediate routing equipment doesn't recognize it as a valid video stream.

Now imagine, that you can store your progressive frame data in a container stream that the routing equipment sees as a valid interlace stream. The routers now do their job and forward the stream as desired. The final receiving equipment still needs to actually recognize the data as progressive, but you've still drastically reduced the overall expense of supporting progressive data in a large facility.

That's an extremely simply explanation of the what and why of psf. As for your quality concerns - as long as your recorder recognizes it properly as progressive, it is no drop or difference in quality. The Final Cut bug referenced is not one I have personal experience with. Forthe RGB 444 the F23/F35 do use a more proprietary interlacing structure across both dual link cables, and not all recorders will support it. The O7Q does.
 
yachacha said:
Yes, it is always pSf out of the monitor out, I have tested every possibility. 50P out of the IF box is true progressive. There is no difference although if the pSf is handled incorrectly on export from something like final cut, the footage becomes grainier. I know there have been a few conversations about this, I am not as good at finding references as Viddo.

Sony uses a unique pSf to send the 444 signal over two sdi lines, this is why 7Q can see the dual link correctly and others might not be able to.

The Aja LUT box is another device which will remove the pSf and convert to true progressive.

I usually run footage through compressor and convert to true progressive before I edit.
Click to expand...

Thanks yachaha. Since your post I've done some reading and found that the HD-SDI spec does not allow for true progressive and is always PsF, so its to be expected that the camera would output that. The docs on my Atomos say that it accepts segmented frames without issues and converts them to true progressive. So I think I'm good here, phew. That said, if you happen to see an IF show up let me know!
 
Even the F65 only outputs PsF. It is a legacy thing. it is slowly being dropped.

Not sure if you were aware but you need to make sure you perform an extended black balance to do a proper dead pixel check. Hold the button down for longer... I may be confusing myself with the F900. The F35 may do that automatically. My F35 did have a few I had to zap manually.

The camera only checks either red, green or blue channels for dead pixels each cycle. hence it can take a few goes.

Enjoy the camera ;)
 
Also for those curious I believe PsF made life easier on tape decks - breaking the picture up into half frames eased their workload considerably.
 
Viddovation said:
PSF is "progressive segmented frames"

It's progressive data in an interlaced container stream. Think of the situation for video production facilities say 10 years ago, with hundreds of thousands if not millions of dollars just spent on brand new high definition wiring, decks, cable runs, routing and so forth. Imagine that via routing you can have engineering send a desired video signal from any place in the building to any other point. Imagine that all of your very expensive equipment only understands interlaced SDI streams - not progressive. If you try and send a "true progressive" SDI stream, it's going to fail to route, because the intermediate routing equipment doesn't recognize it as a valid video stream.

Now imagine, that you can store your progressive frame data in a container stream that the routing equipment sees as a valid interlace stream. The routers now do their job and forward the stream as desired. The final receiving equipment still needs to actually recognize the data as progressive, but you've still drastically reduced the overall expense of supporting progressive data in a large facility.

That's an extremely simply explanation of the what and why of psf. As for your quality concerns - as long as your recorder recognizes it properly as progressive, it is no drop or difference in quality. The Final Cut bug referenced is not one I have personal experience with. Forthe RGB 444 the F23/F35 do use a more proprietary interlacing structure across both dual link cables, and not all recorders will support it. The O7Q does.
Click to expand...

Thanks for the solid explanation here, Viddo. After playing around with the camera more last night, I'm confident that the Atomos seems particularly equipped to handle these interlaced-wrapped progressive streams. The prores files that get saved to the SSD are clean, true progressive.
 
JMtheDP said:
Even the F65 only outputs PsF. It is a legacy thing. it is slowly being dropped.

Not sure if you were aware but you need to make sure you perform an extended black balance to do a proper dead pixel check. Hold the button down for longer... I may be confusing myself with the F900. The F35 may do that automatically. My F35 did have a few I had to zap manually.

The camera only checks either red, green or blue channels for dead pixels each cycle. hence it can take a few goes.

Enjoy the camera ;)
Click to expand...

Interesting about the F65, James. My understanding at this point is that it's really an HD-SDI spec thing and hence these cables will always send PsF (This would mean the Alexa outs are segmented too)? I'm not exactly clear on how the IF box's dual outs get around this. I guess this is one argument for newer specs like HDMI (and thunderbolt?), although I'll take the locking SDI plugs anyday.

And so far so good on stuck pixels, after a few consecutive ABBs, but good to know there are more options if 'cosmic rays' cause any to return!
 
MikalZ said:
Interesting about the F65, James. My understanding at this point is that it's really an HD-SDI spec thing and hence these cables will always send PsF (This would mean the Alexa outs are segmented too)? I'm not exactly clear on how the IF box's dual outs get around this. I guess this is one argument for newer specs like HDMI (and thunderbolt?), although I'll take the locking SDI plugs anyday.
Click to expand...

It has nothing to do with the cable. HD-SDI cables are simply very high quality coax shielded 75 ohm cables. You can use the same cable to run composite video if you like.

Composite video is an analog voltage signal though, whereas all flavors of SDI are digital. That means an SDI signal is a very rapid serial stream of 1s and 0s encoded as logic highs and lows. The cable has to be very high quality because otherwise the rapid digital stream can get interrupted.

As you might imagine, as far as the cable is concerned you can send any order of 1s and 0s that you like, at any speed that you like. This is why 3G, 6G and 12G connections using a single SDI cable are possible - you're just sending the data faster. But the equipment at the receiving end needs not only to be able to operate at the correct speed, (and actually - the newer equipment has to detect the incoming speed) but has to have been programmed to understand and decode the format being used.

As I said in my prior post, a great deal of early very expensive HD-SDI equipment was only programmed via hardware and firmware to understand interlace formats. Today most equipment is designed for user firmware upgrades, but this wasn't nearly as true 10 years ago, and even so you couldn't count on every manufacturer of every piece of gear in your facility providing those updates in a timely manner. Sending progressive data in an interlaced container stream was a backwards compatibility measure.

The reason the monitor ports of the F35 are set via hardware to use PsF and the IF box uses true progressive is because the engineers anticipated the monitor port needing to be backwards compatible with older interlace only equipment. But again, it's not the cable. It's the equipment hardware/firmware.
 
Viddovation said:
It has nothing to do with the cable. HD-SDI cables are simply very high quality coax shielded 75 ohm cables. You can use the same cable to run composite video if you like.

Composite video is an analog voltage signal though, whereas all flavors of SDI are digital. That means an SDI signal is a very rapid serial stream of 1s and 0s encoded as logic highs and lows. The cable has to be very high quality because otherwise the rapid digital stream can get interrupted.

As you might imagine, as far as the cable is concerned you can send any order of 1s and 0s that you like, at any speed that you like. This is why 3G, 6G and 12G connections using a single SDI cable are possible - you're just sending the data faster. But the equipment at the receiving end needs not only to be able to operate at the correct speed, (and actually - the newer equipment has to detect the incoming speed) but has to have been programmed to understand and decode the format being used.

As I said in my prior post, a great deal of early very expensive HD-SDI equipment was only programmed via hardware and firmware to understand interlace formats. Today most equipment is designed for user firmware upgrades, but this wasn't nearly as true 10 years ago, and even so you couldn't count on every manufacturer of every piece of gear in your facility providing those updates in a timely manner. Sending progressive data in an interlaced container stream was a backwards compatibility measure.

The reason the monitor ports of the F35 are set via hardware to use PsF and the IF box uses true progressive is because the engineers anticipated the monitor port needing to be backwards compatible with older interlace only equipment. But again, it's not the cable. It's the equipment hardware/firmware.
Click to expand...

You clearly know a lot about this, Viddo. I'm in new territory here and may have gotten bad internet info, going off the following page which says:
"You cannot output true progressive over HDSDI, the standards do not allow for it, so you use PsF which is progressive inside an interlace stream instead."

http://www.xdcam-user.com/2013/03/what-is-psf-or-why-does-my-camera-output-interlace-in-progressive/

So just to clarify, your saying that the HDSDI spec does indeed allow for progressive - and it's just a matter of hardware implementation?
 
MikalZ said:
You clearly know a lot about this, Viddo. I'm in new territory here and may have gotten bad internet info, going off the following page which says:
"You cannot output true progressive over HDSDI, the standards do not allow for it, so you use PsF which is progressive inside an interlace stream instead."

http://www.xdcam-user.com/2013/03/what-is-psf-or-why-does-my-camera-output-interlace-in-progressive/

So just to clarify, your saying that the HDSDI spec does indeed allow for progressive - and it's just a matter of hardware implementation?
Click to expand...

Alister Chapman is a far wiser man than I. That said, you're quoting a response he made to a very specific question about the Sony PXW-FS7's outputs, and if you read the response in that context it takes a more limited nature.

A better quote to pull would have been from the begininng of the article the comment was in relation to: "the early standards for HDSDI did not include a specification for 25 or 30 frame per second progressive video."

Specifically, SMPTE292M was the HD-SDI standard adopted in 1998. It defined a set of 720p and 1080i standards. These standards set forth the timing expectations and how the different bits of data (audio, video, meta-data, closed captioning, etc) were to be encoded. There was in 1998 no standard for progressive 1080p.

However, in 2002 SMPTE372M was adopted defining Dual Link SDI with 1080p variations. The IF box on the F35 is a Dual Link port, and complies with the Dual Link (372M) specifications. The monitor ports are not dual link (there are two ports, but they're independent of each other) and they comply with the SMPTE292M specification, but with the added support for PsF, which is a cheat that lets progessive frames travel in an interlace (292M) stream.

You ought to understand that a specification like HD-SDI is comprised of multiple elements - it may define electrical requirements, it may define physical connectors and wiring requirements, and perhaps most importantly, it defines how the data itself is encoded. Those sort of documents tend to rely heavily on sets of statements that go "device MAY" and "device MUST". This is how you get components from multiple manufacturers to play nice with each other.
 
Viddovation said:
Alister Chapman is a far wiser man than I. That said, you're quoting a response he made to a very specific question about the Sony PXW-FS7's outputs, and if you read the response in that context it takes a more limited nature.

A better quote to pull would have been from the begininng of the article the comment was in relation to: "the early standards for HDSDI did not include a specification for 25 or 30 frame per second progressive video."

Specifically, SMPTE292M was the HD-SDI standard adopted in 1998. It defined a set of 720p and 1080i standards. These standards set forth the timing expectations and how the different bits of data (audio, video, meta-data, closed captioning, etc) were to be encoded. There was in 1998 no standard for progressive 1080p.

However, in 2002 SMPTE372M was adopted defining Dual Link SDI with 1080p variations. The IF box on the F35 is a Dual Link port, and complies with the Dual Link (372M) specifications. The monitor ports are not dual link (there are two ports, but they're independent of each other) and they comply with the SMPTE292M specification, but with the added support for PsF, which is a cheat that lets progessive frames travel in an interlace (292M) stream.

You ought to understand that a specification like HD-SDI is comprised of multiple elements - it may define electrical requirements, it may define physical connectors and wiring requirements, and perhaps most importantly, it defines how the data itself is encoded. Those sort of documents tend to rely heavily on sets of statements that go "device MAY" and "device MUST". This is how you get components from multiple manufacturers to play nice with each other.
Click to expand...

Man, we have gone down the rabbit hole on this one! Thanks for the info. At this point I'm content knowing there is no degradation in the monitor outs, compared with sending 4:2:2 10bit from the IF.

What the IF *does* allow is 4:4:4 and I'm still on the fence about how much of a win this actually is, when factoring in the added weight and larger disk footprint.
 
dhessel said:
It also allows for 50P if I am not mistaken.
Click to expand...

I need to comb through the big F35 thread again to the part where this is discussed. Is the bottom line that the monitor outs are only good for 24p? The manual shows available "PsF" rates only up to 29.97, so sure looks like it.

Also recall there being talk of 50p softening the image, not sure where that ultimately landed.
 
You simply can't get the 50p datastream on a 1.5Gbps HDI-SDI stream. 50p requires a dual link or 3G (as in 3Gbps) port. The monitor ports don't do this. Afterall, even if you're recording 50p you don't really require a 50p stream to monitor on set - only to record.
 
You must log in or register to reply here.
Back
Top