What bothers me about 24F

[Raymond Toussaint]

Originally Posted by Bill Pryor

I think the reason HVX200 users are recording to Firestore is because of the tremendous cost and low capacity of P2 cards.

Yes,... but that is not the issue here. What do you want to say in relation to 24, 25, 30F? I don't get it. I am talking about the Canon HDV implementation and how to work with it without a tapedeck.

[Bill Pryor]

Without a deck you use the camera as a deck, or you buy the HV10 for about 1200 bucks to use as a deck. It will play all the different modes, so no problem there. For me, using the camera as a deck is working OK for personal stuff. But for the company, we couldn't live without regular decks.

[Thomas Smet]

I actually prefer the way Canon does it. Yes it may not work in any decks but how many of us really go out and buy a deck after we bought a new camera?

1. progressive encoding is just cleaner per bitrate compared to interlaced at the same bitrate.
2. progressive encoding uses a cleaner form of 4:2:0 color that gives much better results and is closer to how jpeg chroma compression works on still images.
3. Most NLE's now support 24F editing while the SONY flavor may take a few months to get supported and who knows how many bugs there will be at first.
4. 24F has more bits and less artifacts per frame due to less number of frames compared to 30F or 60i. SONY 24P will have the same quality level as 60i shot with the same camera.

There are some things I really like about the SONY cameras but the best image in the world doesn't mean much if it is encoded badly. Look at PBS HD. They use much higher end HD cameras then we are using but yet the channel suffers in quality due to bad compression. A lot of people notice 720p channels to be cleaner then 1080i channels while having almost the same level of detail. The 720p channels do not use better cameras but 720p is easier to encode. Not because it has less pixels which isn't really true when you talk about 60p but mainly because it is progressive in nature.

[Matthew Nayman]

Dude, I was wondering why PBS looks so bad! Especially compared to discovery!

[Michael Y Wong]

Codecs and compression techniques are everything when it comes to digital delivery!

Just look @ Blu-Ray, their first generation of Blu-Ray discs we're GARBAGE (despite their brilliant 1080p advertising ploy).

The quality of House of flying daggers was a joke (the SD dvd looked bette) which imo is unacceptable for a $1000 'top of the line state of the art' video player.

Sony blames the first generation Blu-Ray discs to look like junk cuz they were encoded via Mpeg2 & not VC-1 but that is just BS since Sony uses Mpeg2 for HDV, they well know that Mpeg2 can be effectly used to encode HD material VERY VERY well.

[Thomas Smet]

Originally Posted by Michael Y Wong

Codecs and compression techniques are everything when it comes to digital delivery!

Just look @ Blu-Ray, their first generation of Blu-Ray discs we're GARBAGE (despite their brilliant 1080p advertising ploy).

The quality of House of flying daggers was a joke (the SD dvd looked bette) which imo is unacceptable for a $1000 'top of the line state of the art' video player.

Sony blames the first generation Blu-Ray discs to look like junk cuz they were encoded via Mpeg2 & not VC-1 but that is just BS since Sony uses Mpeg2 for HDV, they well know that Mpeg2 can be effectly used to encode HD material VERY VERY well.

I agree. HD cable broadcasts use less then 25 mbits/s and they can look very good. mpeg-2 can look very very good with film based material. I'm not sure what SONY did to make theose movies look so bad. It was almost as if they took an SD master and tried to upscale it to then encode as HD because they didn't want to have to deal with creating a new HD master to source from the film. By saying the mpeg2 version at 25 mbits or even 35 mbits looked bad because it was mpeg2 based would be saying that all HDV is garbage because it is mpeg2 based.


As for PBS I think they broadcast at around only 12 mbits/s which is why it looks so bad. at that rate I think they would have been better off broadcasting as 720p or even 854x480x60p.

Anyways getting back on topic here, the 24F method is about the cleanest form of mpeg2 encoding you can find in any of these cameras and can even come close in terms of raw encoding quality to the 35 mbits mode in XDCAM HD.

The JVC method of HDV is also very clean and I'm sure it is even better with the new super encoder in the 200 series of the cameras.

[David Ziegelheim]

Originally Posted by Barry Green

Well, now you come to the crux of the problem. What is "HDV"? What does it support? There are currently three manufacturers offering "HDV" gear, and all three are, to some degree, incompatible with each other. So you have JVC HDV, which won't play any Canon or Sony footage, and Canon won't play JVC footage, and Sony will display JVC 30P footage but not its 24P or 60P footage...

Then in the 1080 realm you have the original 1080 spec, which is what Sony is compliant with, which is 1080/60i and 1080/50i and that's it. Then Canon came along and invented their own recording format for 30F, 25F and 24F. So now the HDV specification has been extended to include those modes, but Sony doesn't support any of them. You can't play 24F, 25F, or 30F footage on any Sony equipment.

Yet it's all "HDV". Very confusing.

So, therein lies the crux of the matter. Sony gear doesn't support progressive recording, so when they made a progressive camcorder they chose to implement the progressive footage within an interlaced data transport stream. That gave them backwards compatibility with all their other HDV equipment; any Sony (or Canon) camera or deck can read any Sony HDV footage. JVC cameras and decks can't, but Sony cameras & decks, and Canon cameras, can.

Had they used progressive recording, they'd be in the same boat Canon is in, which is that no HDV deck can read Canon 24F, 25F, or 30F footage. But, Canon records progressively rather than embedded in an interlaced data stream, which should yield technically superior results. But at the expense of having made their own proprietary format which no deck can read (because Canon doesn't make decks).

To further confuse the issue, when the Canon outputs 24F over HD-SDI or analog component, it does add 2:3 pulldown into the signal. It's only the HDV tape itself where the footage is encoded as raw 24-frame progressive; on the analog or HD-SDI outputs it's treated as a 60i data stream.

Barry, Wow, excellent answer, now I understand...


Since all the tape drive are 25Mb/sec, the JVC is encoding 24 1280x720 frames, the Canon 24 1440x1080 frames, and the Sony 60 1440x540 frames, does that mean that the JVC has less compression than the Canon which in turn has less than the Sony? Or can the Sony set a flag to that leaves the duplicate frame blank in the stream? Does this impact image quality in the actual implementations?

The HD-SDI output of the H1 and G1, and HDMI output of the V1 would bypass both these issues. Has anyone posted HD-SDI and HDMI outputs respectively?

Thanks,

David

[Steve Wolla]

We have had significant incompatibilites between brands in the SD realm as well. Almost nothing made on a Canon recorder wants to play back cleanly and consistently on a Sony or JVC playback deck. Yet it will always play back wonderfully in the cam that it was made in.
Our solution was to dub down from the cam that the master is shot in, to our playback deck of choice.

Can you process HDV masters in a similar way? Could you do the same thing with HDV, dubbing for example, a Canon A1-made master to a Sony HVR-M15U or other playback deck, and thereby achieve reliable playback of your Canon made tapes? There should not be any appreciable signal loss in the process.

[Pete Bauer]

Originally Posted by David Ziegelheim

Since all the tape drive are 25Mb/sec

JVC folks correct me if I'm mistaken, but I think "HDV1" (1280x720) is recorded at a 19Mb/sec bit rate?

For SDI output examples, look through the Clips subfourm in the XL H1 area. There are a quite a few links there. A lot of it is down-sampled at capture to DVCProHD (1280x1080 anamorphic), though. Not sure off the top of my head if there are any Cineform Prospect 1920x1080 clips in the subforum.

[Chris Hurd]

Yes that's right, HDV1 (1280x720 from JVC) is 19Mb/s.

[Chris Korrow]

Whoa, There is some great info to be found on this site, Thanks to you all (especially to you Chris)

Okay so I think ya'll have brought me back to Canon (originally my first choice) but I'd like to get an FX7 for the slow motion capabilities, but then I'd need an HV10 for a deck, What a pain.

So can a Canon HV10 play back Sony footage?

[Philip Williams]

Originally Posted by Chris Korrow

<snip>
So can a Canon HV10 play back Sony footage?

-------------

Yes.

[David Ziegelheim]

JVC says:

Compressing progressive signals is more efficient than compressing interlace signals. At any given compression ratio, progressive images can be reconstructed more faithfully and with fewer artifacts. That's one reason all high end digital cinematography systems utilize progressive scan.

The following table compares the relative amount of compression in JVC's GY-HD100U with that used by competing systems. These numbers represent data rates, not actual picture quality. Typically, in systems that are less efficient (i.e. frame bound or interlace) manufacturers have reduced the resolution in the image so that the DCT compression works more efficiently.

So, even at 24 frames a second for both, 1440x1080 would have 69% more pixels than 1280x720. Using 25Mb/sec and 19.7Mb/sec respectively, that is 33% more compression. If using 30frames/sec with 3:2 pulldown, it becomes 66% more compression. Then JVC says that progressive frames compress better.

So, in practice...since we are only comparing 3 cameras (HD110, A1, and V1) and their immediate relatives (HD200/250, G1/H1, FX7) how does this affect the image? And of course, this begs the image, how much does capturing HD-SDI or HDMI output help? If the HDV compression hurts Sony the most, presumably capturing the HDMI would help it the most.

David

[Chris Hurd]

What you're failing to realize is that the compression is not affecting the image nearly as much as a wide variety of other factors including the quality of optics, the efficiency of the DSP, etc.

Why are you obsessing over numbers? If you want to know how compression affects the image, simply look at the image on an HDTV display. There are enough downloadable sample clips available on this site and others which prove that it's pointless to get hung up on this stuff.

The chart you've included from JVC's marketing material is excellent fodder for measurebators, but carries no siginificant impact for the majority of folks whose prime motivation is to actually *use* this gear in a productive way.

[David Ziegelheim]

No, not numbers, but the quality impact.

I believe all of these cameras both allow recoding of input that bypasses the camera section and allow output that bypasses the tape recording section. So the question, since these cameras seem to have much more significant differences in their recording than in their capture or processing, how does the recording impact quality.

MPEG is pretty well known, and I believe there are some people around here who are relative experts on it, I thought we could identify its impact on the result.