capture component out? was: B&H ships FX1

Anhar Miah · November 8th, 2004, 07:43 PM

Sorry to butt in,

But i know this is not conclusive but:

Why not see if you can still view a "live" video when there is no tape inside the FX1, if it does it MAY indicate that no MPEG2 compression is goining on i.e. video feed live from the CCD's.....

Michael Garrett · November 9th, 2004, 07:28 AM

<<<-- Originally posted by Darren Kelly :

I personally think reording to the mini dv tape will be fine for 99.999999% of all shooting, but if you are convinced you need to record from the component out.

Frankly, there is no one who bought this camera that needs the extremely small percentage of improvement this will give you. Remember, HDCAM records to MPEG2 compression, so does Varicam. Seems the Networks think it's OK.

If you want the best in quality, buy a Viper System. If you don't know what a Viper System is.....chances are you don't need one and can't afford one.

The Viper is what Lucas uses to do Star Wars.

DBK -->>>

The topic of this thread is of interest to me, and to others here, who do serious compositing work and are interested in the possibilities of HDV as a low-cost HD acquisition medium. The fact that it is sampled at 4:2:0 obviously makes it much harder to pull a nice key, but if you factor in MPEG 2 compression, this is a not trivial addition. I gather that MPEG compression is funkier than the more DCT-based DV or HD CAM compression because it is interframe, thus decimating the footage further. Isn't it generally considered to be a delivery codec?
I have seen Varicam footage and the compression is much milder (even if it is an MPEG variant?), as well as the image being sampled at 4:2:2. This is probably the practical entry point for shooting HD against a green screen, but is much more expensive. So enquiring minds want to know if uncompressed HDV is possible and provides a big enough boost in raw quality.
Also, I believe that Lucas used the the Sony HD CAM SR cameras for the final SW movie (the 4:4:4 RGB variant of HD CAM), not the Viper cams.

Rgds

Michael

Joel Corral · November 9th, 2004, 10:04 AM

i always thought that as soon as images hit the CCD's the video is compressed. isn't that true? if so how can any footage be played out as uncompressed at a better picture quality?

please some enlight me on this, i am very interested to understand the capabilities of the HDR-FX1, i have one on order and should be here this week.

thanks

joel

Carlos E. Martinez · November 9th, 2004, 10:19 AM

<<<-- Originally posted by Darren Kelly :
If you want the best in quality, buy a Viper System. If you don't know what a Viper System is.....chances are you don't need one and can't afford one.

The Viper is what Lucas uses to do Star Wars.

-->>>

A much more impressive Viper application can be seen in Collateral.

Being the film all shot at night, it's a remarkable achievement. There's a scene in total darkness, with actual night-lit LA at the background, through glass windows, where you can see the shadows of the characters moving. I remember thinking at that moment that nothing like that could ever be achieved with video, so it was a surprise when I read of Michael Mann telling about the Viper they used.

Carlos

Joshua Herrell · November 9th, 2004, 01:12 PM

Okay, I am not to sure how component out works to deleiver HD, to my knowledge there are two types of component, the old yellow, red, white, and the wierd one, Yb, Yr, whatever. If you can take the HD output, uncompressed, then is there not a piece of hardware that can compress it to a diffrent format using a better codec like h.264.

For example, www.real-steam.com

I have seen several clips from the fx1, and the mpeg compression looks good in my opinon. I used vlc to view it and told it to deinterlace. It looked great. 7 sata drives could capture the HD output with know problem uncompressed, but unless I am wrong that will only give you about 3 hours of recording. I was using 400 GB drives to come up with that number. Seams like a waste of 2.4 terabytes of storage.

Is there any small portable device that can be used to convert the raw HD video into a less compressed format besides an Xraid and Xserve G5.

Ricardo Renaldi · November 9th, 2004, 01:40 PM

<<<-- Originally posted by Joshua Herrell :

Seams like a waste of 2.4 terabytes of storage.
-->>>

It depends. If the fotage you adquire is blured because the Mpeg compresion can not cope with the task ( heavy-fast motion), I think that the blured footage is a waste.

Anyway i do not have the camera yet to test this situation.Time will tell.

Ric

Alex Raskin · November 9th, 2004, 01:46 PM

Ricardo, I'm intrigued what application do you have for the camera.

Blur will occur anyway, whether it's a film, or a digitl camera, if there's a movement and you don't track it with the camera.

(With standard 24 to 30 frames per second, anyway. Except when you use special high-speed cameras that are designed to take hundreds and thousands of pictures per second.)

Please elaborate.

Kaspar Stromme · November 9th, 2004, 02:13 PM

Blurring is not the issue, COMPRESSION ARTIFACTING is!. If you pan real fast the image may/may not break up into chunky blocks when there's too much compression. This is the key issue this thread is addressing; whether or not the Component output had 25mbps MPEG2 compression applied to it.

If compression has been applied already then there's no POINT grabbing from the Components IMHO, just grab from tape/Firewire (once the appropriate software comes out)

BUT

If the magic genie spits the stuff out earlier you can capture it raw and choose your own level of compression. Not many people will capture and work with uncompressed video for the reasons stated earlier (disk,cpu)

But a lot of us might be interested in capturing somewhere in between, say 50mbps.

Ricardo Renaldi · November 9th, 2004, 02:13 PM

Alex,

We are begining to shot a documental piece about south american ethnic music. Most of the musician's shots include people playing all kind of drums and percussion instruments in very fast tempos. Also the native dancers complement of choice, all shot in situations without any reheasal and no second chance for another shot.

I understand that you get some bluring as a byproduct of the capturing process in film and video. But I'd like to avoid putting more Mpeg blur on top of that. (If I can avoid it).

Also I'm very computer savy. We currently shot wiith several PC laptops for video and one powerbook for multichannel surround audio (proTools), so all this "scaring about the computer and the costs involved" gets no impact in me. An smile, may be.

cheers

Ric

Darren Kelly · November 9th, 2004, 02:27 PM

<<<-- Originally posted by Ricardo Renaldi : Alex,

Also I'm very computer savy. We currently shot wiith several PC laptops for video and one powerbook for multichannel surround audio (proTools), so all this "scaring about the computer and the costs involved" gets no impact in me. A smile, may be.

cheers

Ric -->>>

Ric,

I guess that comment is aimed in my direction. I apologize if you took it as me scaring anyone. The basic fact is I've done this. I've purchased the computer, the SCSI card and the RAID. I did it so I could edit uncompressed HD.

I've also seen lots of people fail trying to cut corners by using IDE raids, SATA raids, etc.

What seems to be forgotten here is the camera is a $3700 camera. It has it's limitations.

Since I am capable of doing this test, I'll add it to the Comparison DVD that's shipping this week so that the people who buy the DVD can see the difference without spending a lot of cash.

DBK

Jeff Kilgroe · November 9th, 2004, 02:37 PM

All the people who are scared of building a computer to capture and deal with uncompressed HD are kinda comical... It's not a big deal and there are plenty of off-the-shelf components to build such a rig. Cost is the real issue and even that isn't that big of deal - the most expensive part would be the HD capture card at about $8500 for the Bluefish444 HD|Fury which can capture uncompressed 4:2:2 component or RGB and it has a host of other nifty options. I could probably assemble the whole system with 3TB of storage capable of 300+MB/sec sustained for about $15K give or take and that would be a dual 3.4GHz Xeon system. I may consider dual Opteron as well...

Anyhoo... Considering such a system for the FX1 is, well, uh, stupid. I haven't spent much time with the FX1 - actually I just took a gander at one last night at the local shop. We did plug its component out into a 34" Toshiba HD set and it looks to me like the MPEG2 artifacting is indeed there. So, I'm pretty sure that the output over the component interface is showing the video stream after it has been MPEG2 encoded. So capturing from the FX1 would appear to be pointless. Even if the output is a raw component stream, this would still be introducing a digital to analog back to digital transformation in the capture workflow.

Anyway, this is all just my opinion, but buying a $15K PC to complement ia $3600 camera is somewhat dumb. If you need a camera that can shoot higher quality video than the FX1 for a project, why not just save your sanity and go rent a nice camera with a decent lens mount, 2/3" CCDs that can shoot the format you need and save to HDCAM-50/100??? You don't have to go all the way and rent a Varicam or CineAlta... Sony has HD camcorders in the broadcast line that can do 4:2:2 1080i on HDCAM-50 that retail for less than $50K. I'm sure it wouldn't cost more than about $2500 to rent them for a few days/week.

Ricardo Renaldi · November 9th, 2004, 02:51 PM

<<<-- Originally posted by Darren Kelly : <<<

I guess that comment is aimed in my direction. I apologize if you took it as me scaring anyone.
DBK -->>>

Darren,

Not at all. In fact, I really apreciate a good dose of reality. More if it comes with the experience to back it.And I think that's your case.

What I'm refearing to is that all this talking about computer tech is a moving target. Developement and price drops are going so fast that a situation resolved with last month technology can be solved differently today or tomorrow.
Take the new maxtor Sata drives with the new instruction set for example. As soon as a motherboard that support that instruction is available (besides the new flawed intel boards) it will open a new performance frontier for Sata, and a new price/performance drop and this all just for implementing an instruction.
It goes fast and we need to benefit from that.

About the argument based on the cameras's price (4K), well, I would pay ten grands for something like the FX1(Z1) wiithout the mpeg video AND audio thight limitations (I mean less compression, using DVCAM speed?. I do not expect a 50K product for ten times less, of course, just a bit less compression for 100% price increase, for example) .

But as you know, there's not such a product YET. Not even the new JVC 20K product arrived, and I begin shooting in two months.

So the name of this game for me, right now is: Buy what I can now (can't wait for the Z1) and try to overcome the technical limitations imposed in a product like this.

cheers

Ric

Ricardo Renaldi · November 9th, 2004, 03:05 PM

<<<-- Originally posted by Jeff Kilgroe :

We did plug its component out into a 34" Toshiba HD set and it looks to me like the MPEG2 artifacting is indeed there. So, I'm pretty sure that the output over the component interface is showing the video stream after it has been MPEG2 encoded. -->>>

If the component out is already compressed I'd say there's no point at all to capture uncompressed, of course.

Ric

Anhar Miah · November 9th, 2004, 05:24 PM

I was just wondering that sounds a bit odd that the componet output is mpeg2 and not just plain anolouge beacuse if you think about it this would mean that the DSP would not only have too do RT mpeg2 coding in the first place but also simultaenoulsy decode that mpeg2 and convert to Anolouge! thats a lot of work..... just my thought....

is there any conclusive way we can fid out?

Kaspar Stromme · November 9th, 2004, 07:36 PM

imho I think Sony would go the extra step to make sure component out was encoded to HDV specs. Uncompressed HD at $3700 would put a lot of pressure on their $50k pro cameras (even with the cheaper CCDs, fewer manual controls, etc)

btw: if the above turns out to be false and component is uncompressed, using a $3700 camera with a $10-15k capture/edit station isn't all that crazy when the alternative is a 50k lens. Sure, rent the 50k for your biggest shots, but a cheapo FX1 can handle a lot of the other stuff (and you can own it)

November 9th, 2004, 07:28 AM	#47
Michael Garrett New Boot Join Date: Nov 2004 Location: Australia Posts: 5	Re: Re: Data stream? <<<-- Originally posted by Darren Kelly : I personally think reording to the mini dv tape will be fine for 99.999999% of all shooting, but if you are convinced you need to record from the component out. Frankly, there is no one who bought this camera that needs the extremely small percentage of improvement this will give you. Remember, HDCAM records to MPEG2 compression, so does Varicam. Seems the Networks think it's OK. If you want the best in quality, buy a Viper System. If you don't know what a Viper System is.....chances are you don't need one and can't afford one. The Viper is what Lucas uses to do Star Wars. DBK -->>> The topic of this thread is of interest to me, and to others here, who do serious compositing work and are interested in the possibilities of HDV as a low-cost HD acquisition medium. The fact that it is sampled at 4:2:0 obviously makes it much harder to pull a nice key, but if you factor in MPEG 2 compression, this is a not trivial addition. I gather that MPEG compression is funkier than the more DCT-based DV or HD CAM compression because it is interframe, thus decimating the footage further. Isn't it generally considered to be a delivery codec? I have seen Varicam footage and the compression is much milder (even if it is an MPEG variant?), as well as the image being sampled at 4:2:2. This is probably the practical entry point for shooting HD against a green screen, but is much more expensive. So enquiring minds want to know if uncompressed HDV is possible and provides a big enough boost in raw quality. Also, I believe that Lucas used the the Sony HD CAM SR cameras for the final SW movie (the 4:4:4 RGB variant of HD CAM), not the Viper cams. Rgds Michael

November 9th, 2004, 10:19 AM	#49
Carlos E. Martinez Trustee Join Date: Oct 2003 Location: Rio de Janeiro, Brasil Posts: 1,138	Re: Re: Data stream? <<<-- Originally posted by Darren Kelly : If you want the best in quality, buy a Viper System. If you don't know what a Viper System is.....chances are you don't need one and can't afford one. The Viper is what Lucas uses to do Star Wars. -->>> A much more impressive Viper application can be seen in Collateral. Being the film all shot at night, it's a remarkable achievement. There's a scene in total darkness, with actual night-lit LA at the background, through glass windows, where you can see the shadows of the characters moving. I remember thinking at that moment that nothing like that could ever be achieved with video, so it was a surprise when I read of Michael Mann telling about the Viper they used. Carlos

November 9th, 2004, 02:13 PM	#54
Ricardo Renaldi Regular Crew Join Date: Sep 2004 Location: Buenos Aires, Argentina Posts: 40	Easy Alex, We are begining to shot a documental piece about south american ethnic music. Most of the musician's shots include people playing all kind of drums and percussion instruments in very fast tempos. Also the native dancers complement of choice, all shot in situations without any reheasal and no second chance for another shot. I understand that you get some bluring as a byproduct of the capturing process in film and video. But I'd like to avoid putting more Mpeg blur on top of that. (If I can avoid it). Also I'm very computer savy. We currently shot wiith several PC laptops for video and one powerbook for multichannel surround audio (proTools), so all this "scaring about the computer and the costs involved" gets no impact in me. An smile, may be. cheers Ric

November 9th, 2004, 02:27 PM	#55
Darren Kelly Major Player Join Date: Aug 2003 Location: Vancouver Posts: 351	Re: Easy <<<-- Originally posted by Ricardo Renaldi : Alex, Also I'm very computer savy. We currently shot wiith several PC laptops for video and one powerbook for multichannel surround audio (proTools), so all this "scaring about the computer and the costs involved" gets no impact in me. A smile, may be. cheers Ric -->>> Ric, I guess that comment is aimed in my direction. I apologize if you took it as me scaring anyone. The basic fact is I've done this. I've purchased the computer, the SCSI card and the RAID. I did it so I could edit uncompressed HD. I've also seen lots of people fail trying to cut corners by using IDE raids, SATA raids, etc. What seems to be forgotten here is the camera is a $3700 camera. It has it's limitations. Since I am capable of doing this test, I'll add it to the Comparison DVD that's shipping this week so that the people who buy the DVD can see the difference without spending a lot of cash. DBK __________________ Darren Kelly

November 9th, 2004, 02:37 PM	#56
Jeff Kilgroe Major Player Join Date: Jun 2003 Location: Golden, CO Posts: 681	All the people who are scared of building a computer to capture and deal with uncompressed HD are kinda comical... It's not a big deal and there are plenty of off-the-shelf components to build such a rig. Cost is the real issue and even that isn't that big of deal - the most expensive part would be the HD capture card at about $8500 for the Bluefish444 HD\|Fury which can capture uncompressed 4:2:2 component or RGB and it has a host of other nifty options. I could probably assemble the whole system with 3TB of storage capable of 300+MB/sec sustained for about $15K give or take and that would be a dual 3.4GHz Xeon system. I may consider dual Opteron as well... Anyhoo... Considering such a system for the FX1 is, well, uh, stupid. I haven't spent much time with the FX1 - actually I just took a gander at one last night at the local shop. We did plug its component out into a 34" Toshiba HD set and it looks to me like the MPEG2 artifacting is indeed there. So, I'm pretty sure that the output over the component interface is showing the video stream after it has been MPEG2 encoded. So capturing from the FX1 would appear to be pointless. Even if the output is a raw component stream, this would still be introducing a digital to analog back to digital transformation in the capture workflow. Anyway, this is all just my opinion, but buying a $15K PC to complement ia $3600 camera is somewhat dumb. If you need a camera that can shoot higher quality video than the FX1 for a project, why not just save your sanity and go rent a nice camera with a decent lens mount, 2/3" CCDs that can shoot the format you need and save to HDCAM-50/100??? You don't have to go all the way and rent a Varicam or CineAlta... Sony has HD camcorders in the broadcast line that can do 4:2:2 1080i on HDCAM-50 that retail for less than $50K. I'm sure it wouldn't cost more than about $2500 to rent them for a few days/week. __________________ - Jeff Kilgroe - Applied Visual Technologies \| DarkScience - www.darkscience.com

November 8th, 2004, 07:43 PM	#46
Anhar Miah Major Player Join Date: Dec 2003 Location: United Kindom, England Posts: 290	Sorry to butt in, But i know this is not conclusive but: Why not see if you can still view a "live" video when there is no tape inside the FX1, if it does it MAY indicate that no MPEG2 compression is goining on i.e. video feed live from the CCD's.....

November 9th, 2004, 10:04 AM	#48
Joel Corral Major Player Join Date: Nov 2003 Location: Brea, CA Posts: 356	i always thought that as soon as images hit the CCD's the video is compressed. isn't that true? if so how can any footage be played out as uncompressed at a better picture quality? please some enlight me on this, i am very interested to understand the capabilities of the HDR-FX1, i have one on order and should be here this week. thanks joel

November 9th, 2004, 01:12 PM	#50
Joshua Herrell Tourist Join Date: Nov 2004 Posts: 4	Okay, I am not to sure how component out works to deleiver HD, to my knowledge there are two types of component, the old yellow, red, white, and the wierd one, Yb, Yr, whatever. If you can take the HD output, uncompressed, then is there not a piece of hardware that can compress it to a diffrent format using a better codec like h.264. For example, www.real-steam.com I have seen several clips from the fx1, and the mpeg compression looks good in my opinon. I used vlc to view it and told it to deinterlace. It looked great. 7 sata drives could capture the HD output with know problem uncompressed, but unless I am wrong that will only give you about 3 hours of recording. I was using 400 GB drives to come up with that number. Seams like a waste of 2.4 terabytes of storage. Is there any small portable device that can be used to convert the raw HD video into a less compressed format besides an Xraid and Xserve G5.

November 9th, 2004, 01:40 PM	#51
Ricardo Renaldi Regular Crew Join Date: Sep 2004 Location: Buenos Aires, Argentina Posts: 40	<<<-- Originally posted by Joshua Herrell : Seams like a waste of 2.4 terabytes of storage. -->>> It depends. If the fotage you adquire is blured because the Mpeg compresion can not cope with the task ( heavy-fast motion), I think that the blured footage is a waste. Anyway i do not have the camera yet to test this situation.Time will tell. Ric

November 9th, 2004, 01:46 PM	#52
Alex Raskin Trustee Join Date: Sep 2003 Posts: 1,435	Ricardo, I'm intrigued what application do you have for the camera. Blur will occur anyway, whether it's a film, or a digitl camera, if there's a movement and you don't track it with the camera. (With standard 24 to 30 frames per second, anyway. Except when you use special high-speed cameras that are designed to take hundreds and thousands of pictures per second.) Please elaborate.

November 9th, 2004, 02:13 PM	#53
Kaspar Stromme Regular Crew Join Date: Oct 2004 Posts: 36	Blurring is not the issue, COMPRESSION ARTIFACTING is!. If you pan real fast the image may/may not break up into chunky blocks when there's too much compression. This is the key issue this thread is addressing; whether or not the Component output had 25mbps MPEG2 compression applied to it. If compression has been applied already then there's no POINT grabbing from the Components IMHO, just grab from tape/Firewire (once the appropriate software comes out) BUT If the magic genie spits the stuff out earlier you can capture it raw and choose your own level of compression. Not many people will capture and work with uncompressed video for the reasons stated earlier (disk,cpu) But a lot of us might be interested in capturing somewhere in between, say 50mbps.

November 9th, 2004, 02:51 PM	#57
Ricardo Renaldi Regular Crew Join Date: Sep 2004 Location: Buenos Aires, Argentina Posts: 40	Re: Re: Easy <<<-- Originally posted by Darren Kelly : <<< I guess that comment is aimed in my direction. I apologize if you took it as me scaring anyone. DBK -->>> Darren, Not at all. In fact, I really apreciate a good dose of reality. More if it comes with the experience to back it.And I think that's your case. What I'm refearing to is that all this talking about computer tech is a moving target. Developement and price drops are going so fast that a situation resolved with last month technology can be solved differently today or tomorrow. Take the new maxtor Sata drives with the new instruction set for example. As soon as a motherboard that support that instruction is available (besides the new flawed intel boards) it will open a new performance frontier for Sata, and a new price/performance drop and this all just for implementing an instruction. It goes fast and we need to benefit from that. About the argument based on the cameras's price (4K), well, I would pay ten grands for something like the FX1(Z1) wiithout the mpeg video AND audio thight limitations (I mean less compression, using DVCAM speed?. I do not expect a 50K product for ten times less, of course, just a bit less compression for 100% price increase, for example) . But as you know, there's not such a product YET. Not even the new JVC 20K product arrived, and I begin shooting in two months. So the name of this game for me, right now is: Buy what I can now (can't wait for the Z1) and try to overcome the technical limitations imposed in a product like this. cheers Ric

November 9th, 2004, 03:05 PM	#58
Ricardo Renaldi Regular Crew Join Date: Sep 2004 Location: Buenos Aires, Argentina Posts: 40	<<<-- Originally posted by Jeff Kilgroe : We did plug its component out into a 34" Toshiba HD set and it looks to me like the MPEG2 artifacting is indeed there. So, I'm pretty sure that the output over the component interface is showing the video stream after it has been MPEG2 encoded. -->>> If the component out is already compressed I'd say there's no point at all to capture uncompressed, of course. Ric

November 9th, 2004, 05:24 PM	#59
Anhar Miah Major Player Join Date: Dec 2003 Location: United Kindom, England Posts: 290	I was just wondering that sounds a bit odd that the componet output is mpeg2 and not just plain anolouge beacuse if you think about it this would mean that the DSP would not only have too do RT mpeg2 coding in the first place but also simultaenoulsy decode that mpeg2 and convert to Anolouge! thats a lot of work..... just my thought.... is there any conclusive way we can fid out?

November 9th, 2004, 07:36 PM	#60
Kaspar Stromme Regular Crew Join Date: Oct 2004 Posts: 36	imho I think Sony would go the extra step to make sure component out was encoded to HDV specs. Uncompressed HD at $3700 would put a lot of pressure on their $50k pro cameras (even with the cheaper CCDs, fewer manual controls, etc) btw: if the above turns out to be false and component is uncompressed, using a $3700 camera with a $10-15k capture/edit station isn't all that crazy when the alternative is a 50k lens. Sure, rent the 50k for your biggest shots, but a cheapo FX1 can handle a lot of the other stuff (and you can own it)