Panasonic HVX200 workflow questions (used to DVCAM).

[Nigel]

Sorry, forgot to also mention this little fact. There is a little glitch to this, there is going to be another camera used for this shoot as a second camera. It is a JVC GY-HD101E, which shoots HDV progressive or standard dv interlace only. No HD.

[Stephen Downes]

[iMan] “It is a JVC GY-HD101E, which shoots HDV progressive or standard dv interlace only. No HD.”

iMan,

The JVC GY-HD101E will in fact be shooting pictures that are more ‘HD’ when shooting in HDV mode than the HVX 200 you’ll be using – it records with CCD’s that are 1280×720 pixels, whereas the Panasonic has CCD’s that are only 960×540 (not that much greater than Standard Definition!!!). I’ve seen side-by-side tests with both cameras and the JVC definitely produces sharper looking pictures than the Panasonic. Also with such small sized CCD’s in the Panasonic, the degree of up-sampling of the chroma information means there is probably not that much difference between the colour held in the JVC’s 4:2:0 recorded image compared to the Panasonic’s 4:2:2 recorded image. I’m not saying there is anything wrong with either camera, it’s just that I get tired of people saying HDV is not HD – I’m sure the HDV consortium did a disservice to the format by putting a DV in it’s name. HDV is definitely HD because it produces pictures that are larger in dimension than standard def. Standard def is 720×576 for PAL and 720×480 for NTSC. The 1080i HDV spec is 1440×1080 (which incidentally is the same as HDCam) and the 720p HDV spec is 1280×720 (compare this to Panasonic’s high-end VariCam that has a native resolution of only 960×720!!). Both JVC’s HDV camera and Panasonic’s DVCProHD P2 camera are most definitely HD. The difference lies merely in the algorithms used to compress the material – and by the way, there is no uncompressed HD tape format on the planet!

Stephen Downes

[Jan Crittenden livingston]

[Stephen Downes] “The JVC GY-HD101E will in fact be shooting pictures that are more ‘HD’ when shooting in HDV mode than the HVX 200 you’ll be using – it records with CCD’s that are 1280×720 pixels, whereas the Panasonic has CCD’s that are only 960×540 (not that much greater than Standard Definition!!!). I’ve seen side-by-side tests with both cameras and the JVC definitely produces sharper looking pictures than the Panasonic. Also with such small sized CCD’s in the Panasonic, the degree of up-sampling of the chroma information means there is probably not that much difference between the colour held in the JVC’s 4:2:0 recorded image compared to the Panasonic’s 4:2:2 recorded image.”

Stephen, you are very mistaken in the fact that if you just look at the chip size, you think you know how it all works out. Frankly you have to look at each of these cameras as a complete system and thus when you record the JVC to tape and then start to make the comparisons to the HVX’s recordings, the JVC definitely starts to fall down. You cannot say that the color is not hampered by 4:2:0 vs 4:2:2, 19mbs vs 100mbs, frame independent capture vs long GOP, and yes I know the GOP on the JVC is only 6, but trust me when I say you can cause it some heartburn, just like you can when you try to chromakey the stuff. No you can say you have looked, but there are just as many tests out there that say more of the resolution in the JVC is not resolution but more sharpening.

The fact they you say that there is upsampling in the Panasonic camera says that you are not very clear on how CCDs work. They are an analog device. The signal that comes off of them is an analog signal. What comes off of the HVX chipset is offset spatially to gain additional resolution and only then does it get captured and moved into the digital domain. There is not pixel for pixel relationship to the output on any manufacturer’s camera. Doesn’t work that way. To see read how the HVX CCDs are handled, take a read on http://www.defperception.com. There is an article there that discusses how the signals are sampled off of the HVX CCD set. And it really doesn’t matter with manufacturer, the ccd is an analog device.

Frankly presenting HDV with too much information can come back to hurt you as well, once the camera starts to move. You cannot look at live camera output and sy I know these cameras. You have to look at them in total with the recording system working.

[Stephen Downes] “I’m not saying there is anything wrong with either camera, it’s just that I get tired of people saying HDV is not HD – I’m sure the HDV consortium did a disservice to the format by putting a DV in it’s name. “

I don’t think that anyone really means that it it really isn’t HD but I do feel that what they are really talking about is that it is a very difficult codec to work with in production and maintain image quality. It virtualy demands that you work in an intermediate codec during the edit process. This is not true of the DVCPRO HD codec, which was designed for professional use. The difference between it and HDV is night and day. Even the 25Mbs HDV has a struggle against the DVCPRO HD at 100Mbs.

[Stephen Downes] “HDV is definitely HD because it produces pictures that are larger in dimension than standard def. Standard def is 720×576 for PAL and 720×480 for NTSC. The 1080i HDV spec is 1440×1080 (which incidentally is the same as HDCam) and the 720p HDV spec is 1280×720 (compare this to Panasonic’s high-end VariCam that has a native resolution of only 960×720!!). Both JVC’s HDV camera and Panasonic’s DVCProHD P2 camera are most definitely HD. The difference lies merely in the algorithms used to compress the material – and by the way, there is no uncompressed HD tape format on the planet!”

And on another forum I mentioned to Iman that the problem he will have here in bringing these two formats together on the same project and having them work well will require that he work virtually from the get go in uncompressed on the the edit domain. The fact is that the two algorithms are so different you will need to move out of them completely to acheive anything of value.

Best,

Jan

Jan Crittenden Livingston
Product Manager, DVCPRO, DVCPRO50, AG-DVX100
Panasonic Broadcast & TV Systems

[Donatello]

bottom line is a real image at 24fps/30fps not a resolution chart at 1 frame per 3 min.
moving away form the HVX .. take a look at the higher end camera’s varicam 720p against the other 1080p camera’s ..
varicam has lowest res , and lets just say ( don’t know if it does) it comes in last on all technical test – does that mean it has the worst image ?
well you will see that on just about every test/comparison between hi end HD camera’s the Varicam is always mentioned as the most film like , the preferred image, the best image when viewing clips at running speeds ..

so lets just say the HVX has the lowest res of the hand size camera’s. and lets just say it has the most noise and comes it last on all test – what you don’t hear test/comparision say is it has worst image – infact you hear just the opposite .. very good to excellent image and ALWAYS it is said most film like !!!

so you want to show a audience ressolution/technical test charts or real images?

now some will prefer one camera image over another image – IMO look at the test but decide by real image clips at running speed !!

[Stephen Downes]

Jan,

I think you misinterpret the purpose of my post. It certainly was not to dish the HVX 200, it was merely out of the frustration of others unfairly stating that HDV is not HD, when it patently is (it produces images that are larger than standard def so it has to be hi def!). I think the HVX 200 is another fantastic tool to add to our burgeoning HD filming toolkit, as I think are some of the HDV cameras – each has it’s place. If you need variable frame-rate in a small form factor then the HVX is the one for you, however the P2 workflow has serious limitations for long-form documentary and maybe an HDV camera is the way to go here (note: Discovery Channel International are not encouraging their co-producers to shoot with the HVX 200 due to archiving and storage implications but encourage the use of the Canon XL H1 and the Sony Z1 and are saying DV is no longer acceptable for SD production).

I would never advocate choosing a camera on it’s CCD dimensions alone and totally agree with you that one must judge the camera as a whole and look at the quality of results. This is in fact the challenge I’d put to many who suggest that HDV is not HD – just look at properly shot material on the appropriate high-end HD monitor and then judge for yourself, don’t listen to here-say. I’ve done exactly that and the HDV material looked stunning as did the HVX material. My dissertation on camera pixel dimensions was merely to make my point that HDV cameras are in fact true HD. And by-the-way, the comparison tests were performed with both cameras set to their minimum detail settings – the JVC could resolve around 700 TV lines in both the vertical and horizontal, whereas the HVX was resolving about 550 TV lines in both dimensions.

I do understand CCD’s are analogue devices and that there isn’t a 1:1 pixel relationship. I also happen to know that pixel shift technology acts to up the luminance resolution but does little to the chrominance information. Therefore take the analogy of another analogue source. If you were to record a VHS signal to DVCProHD would it improve the picture quality – no. But would it improve the image’s robustness through the multi-generation edit process – yes.

This brings me to my last point. To get the most out of HDV it must be thought of as an acquisition format only. To post in HDV beyond mere cuts is mad. Everyone I know who is shooting HDV, is either cutting as uncompressed if they’ve got fast enough disk arrays, or as another intra-frame codec, if not. Many are cutting HDV acquired footage as DVCProHD due to the codec’s high image quality and low data footprint – the quality of the end results look stunning. Indeed, HDV shooters owe a debt of gratitude to Panasonic for their DVCProHD codec, for it allows them to achieve fantastic image quality at very low cost of entry.

And just for the record, when I’m shooting high-end HD productions I usually shoot VariCam.

Best,

Stephen Downes

[Jan Crittenden livingston]

[Stephen Downes] “I think you misinterpret the purpose of my post. It certainly was not to dish the HVX 200, it was merely out of the frustration of others unfairly stating that HDV is not HD, when it patently is (it produces images that are larger than standard def so it has to be hi def!).

I understood that part, but you did go on to try and diss the HVX by saying the the 4:20 of the HDV in the JVC would be equal to the 4:2:2 of the HVX, which frankly is patently untrue.

>I think the HVX 200 is another fantastic tool to add to our burgeoning HD filming toolkit, as I think are some of the HDV cameras – each has it’s place. If you need variable frame-rate in a small form factor then the HVX is the one for you, however the P2 workflow has serious limitations for long-form documentary and maybe an HDV camera is the way to go here

Well again you may not realize how many long form docus are currently in the making using the HVX, but there are a good number with one of the leading contenders being done by Marc Singer, Award winning filmmaker, (Dark Days Sundance 2000). In fact he has taken the camera where no other camera has succeeded into a field of explosions and bullets, and even with a richocet bullet wound it is still recording. And yest he is strictly using P2 cards and the P2 Store. It is an 18 month Documentary that will move to Afghanistan in about a month or so.

>(note: Discovery Channel International are not encouraging their co-producers to shoot with the HVX 200 due to archiving and storage implications but encourage the use of the Canon XL H1 and the Sony Z1 and are saying DV is no longer acceptable for SD production).

Actually that was not the reason for their stipulation at all. Their rationale was that the image quality was not up to their standards of 2/3″ cameras. Here is where the guidelines for the HVX actually match those for the HDV product at Discovery. They do not prefer HDV over the HVX at all. It has the same 15% clause.

>This is in fact the challenge I’d put to many who suggest that HDV is not HD – just look at properly shot material on the appropriate high-end HD monitor and then judge for yourself, don’t listen to here-say. I’ve done exactly that and the HDV material looked stunning as did the HVX material.

Well your experience is different than others that I have seen, but it is definitaely something that can be attributed to content dependent recordings. Shoot your test on what you want to record and if it works for you, fine. If not, not amount of resolution chart recording will make it so.

>My dissertation on camera pixel dimensions was merely to make my point that HDV cameras are in fact true HD. And by-the-way, the comparison tests were performed with both cameras set to their minimum detail settings – the JVC could resolve around 700 TV lines in both the vertical and horizontal, whereas the HVX was resolving about 550 TV lines in both dimensions.

This is because there is excessive sharpening in the camera that is not present in the HVX. There are numerous catfights about the resolution on these cameras and frankly no one gets paid for resolution, they get paid for pictures and this is where the HDV products get the bad rap. That is, when you least expect it, there is a problem in the algorithm and boom, you have footage that is unusable and you don’t know it unless you review the footage.

>I do understand CCD’s are analogue devices and that there isn’t a 1:1 pixel relationship. I also happen to know that pixel shift technology acts to up the luminance resolution but does little to the chrominance information. Therefore take the analogy of another analogue source. If you were to record a VHS signal to DVCProHD would it improve the picture quality – no. But would it improve the image’s robustness through the multi-generation edit process – yes.

This is ridiculous but in essence if the camera has a low resolution it would be enhanced greatly in its stability and newly found color depth, all coming out of the 4:2:2 and frame based algorithm. Given that same VHS level of resolution to HDV, it would fare as well as any HDV signal does.

>This brings me to my last point. To get the most out of HDV it must be thought of as an acquisition format only. To post in HDV beyond mere cuts is mad. Everyone I know who is shooting HDV, is either cutting as uncompressed if they’ve got fast enough disk arrays, or as another intra-frame codec, if not. Many are cutting HDV acquired footage as DVCProHD due to the codec’s high image quality and low data footprint – the quality of the end results look stunning.

And that is the whole point of the HVX, shoot in DVCPRO HD to begin with.

>Indeed, HDV shooters owe a debt of gratitude to Panasonic for their DVCProHD codec, for it allows them to achieve fantastic image quality at very low cost of entry.”

However, I think that one needs to be concerned with the fact that if starting in HDV and making a cross concatenation to another compressed format is not the best way to go and many engineers would argue that this sets you up for problems that may crop up in further generations. Be assured though 4:2:0 does not turn into richer color when trancoded to DVCPRO HD. The color that is present in that original signal becomes 4:2:2. This is very different than being there in the first place as with the HVX.

Best regards,

Jan

Jan Crittenden Livingston
Product Manager, DVCPRO, DVCPRO50, AG-DVX100
Panasonic Broadcast & TV Systems

[Stephen Downes]

Jan,

I refuse to continue this dialogue because with words such as ‘patently untrue’ and ‘ridiculous’ to some of my statements, you obviously already have your mind made up and no amount of discussion will change your view. I’m one of the people who uses the equipment you sell, to make films. I have no brand loyalty (as I stated earlier, I use VariCam much of the time) and all I can relate are mine and my colleagues findings using the equipment in the field and how we manage to make it work for us. I also know there are many people on these web threads who’ve commented on and found the same things as I.

The one thing I will comment on is Discovery policy since it effects the company I work for greatly. The following is an internal memo at Discovery – you make up your own mind whether they have problems with the P2 workflow at present:

The document was written by Josh Derby who is Discovery’s Manager of Technical Standards and Operations.

The memo is as follows:

Panasonic P2 technology represents the next generation of acquisition formats, cameras that record files to re-usable media and enable truly tapeless production and post production. The day for that technology is certainly coming but there are questions as to whether its time has tryly arrived. P2 cards are expensive and are in no way disposable, nor are they intended to be stored as camera masters. As DCI requires producers to deliver original camera masters for all original and comissioned projects, the concept of fixed media poses a challenge. There may be a time in the future where DCI will be prepared to handle digital file-based delivery from its producers but until that time P2 will be governed by the following restrictions.

1. Material captured using P2 for commissions and original productions must be cloned to an approved tape format for final delivery to DCI. Tapes should be cloned in such a way as to ensure that they are properly referenced in the program’s final EDL or NLE project with proper timecode and tape name.

2. The program producer must demonstrate that workflow cost savings achieved by using the P2 format will cancel the expense of the transfer to tape.

3.Material for standard definition must be captured on P2 in the DVCPRO 50 codec. Material for high definition programs must be captured in the DVCPRO 100 codec.

4. The Panasonic HVX200 camera, which can record video in HD standards onto P2 media, has not been approved for unrestricted HD acquisition. While this camera records HD video onto the recording media the optical performance of the camera is closer to the performance of an HDV than to a oprofessional 2/3″ camera system. Consequently, HD media captured using this camera will be governed by the HDV guidelines. The HVX200 may be used in the DVCPRO format for unrestricted Standard Definition capture.

I also have a Discovery document that recommends the use of the Canon XL H1 and Sony Z1 in Standard Def production and how HDV should be handled in post – it’s exactly what I outlined in my earlier post (cutting preferably uncompressed).

Also, the company I work for has just purchased a number of HVX 200’s. The intention of my original post was not to flame the HVX but was merely to state how sick and tired I was of people incorrectly implying that HDV wasn’t really HD. Enough said.

Stephen Downes

[Nigel]

Thanks everyone for the great input. One more question I’ve thought of. Regarding shooting in SD mode on both cameras. If I shoot using the HVX200’s best quality SD mode, think that is DVCPRO50? and use the JVC in SD Mini-DV mode, am I right in assuming that I can mix these two lots of footage in the same Final Cut Pro sequence as they are both 720×576 right? or will the DVCPRO50 standard def codec not work as it is different from the JVC standard def mini-dv codec?

Thanks again people.

[Stephen Downes]

iMan,

Final Cut can’t mix codecs within the same timeline, so material brought in digitally over firewire or directly from P2 would leave you in the situation of choosing a codec (probably the DVCPro50 codec for its higher 4:2:2 quality and lower compression) and then having to forever render the other material. The only way around this is to have either a third party video card or an AJA IO and capture all material through one of these via component (both cameras have component output) to a common high quality codec – the best would be 10 bit uncompressed.

Hope this helps,

Stephen Downes

[Nigel]

That helps a lot. I assume I bring in the SD DVCPRO50 footage from my P2 card, then bring in the SD footage from JVC in Mini-DV mode, then with a sequence set up to higher quality one i.e. the Panny Standard def DVCPRO50, I bring the DVCPRO50 clips to the timeline, then as I drop the JVC standard def mini-dv footage onto the timeline, it will be red-lined indicating that I will need to render it?
Then I simply render it and away I go?

If I chose to use an AJA IO card in 10 bit uncompressed and brining in via component, won’t their be a loss of quality as component is not as good as FireWire?
Also, with 10 bit uncompressed, won’t I need to upgrade my hardware i.e. raid aray etc or will my exsisting internal 7200 hard drive do the job.