After two months with the A1

[Chan Ee Jien]

Originally Posted by Ian G. Thompson

Question: Why call HDV a fake HD? Is it bandwidth that makes HD HD...or the pixels in the picture or resolution? I mean...compression is one thing...but to say it's fake HD because of compression artifacts...I don't get it.

Agreed. Sounds like an uneducated comment to me. I have encountered many people (not from here) who do the say the same things over and over again though they can't really explain why its fake except they heard it from someone someone.

[Chris Hurd]

It is a horribly uneducated and misinformed comment, which is never welcomed here. On this site, HDV is considered as real an HD format as any other HD format... including AVCHD up to HDCAM SR. While it has its limitations just like every other format, there is nothing "fake" about HDV.

Originally Posted by Will Griffith

Funny because I just edited a speedboat video for a big manufacturer that was shot with Sony HDV Cams and it turned out great. It was all plane jane 60i hdv with some uv filters slapped on the front. looked great to me.

Would love to host a sample clip of one of those shots, Will!

[Joe Batt]

The way that I've been able to make dark shots look better is to keep the gain at -3 and slightly overlight the shot, then darken it in post. It works for me. If I try to accomedate a dark scene by turning up the gain I always get grain.

[Jerome Marot]

I am a bit surprised that some people here do not see the compression artifacts, because they are clearly visible on some subjects. And all digital cameras have them. For example, I remember posting similar remarks when I first got my Sony hc1, a few years ago.

Sorry, but you'll have to live with artifacts. All compressed codecs have some. Even DV is not exempt, try filming pebbles.

This being said not all cameras are created equal. When I compare the Sony HC1 with the Canon A1, the A1 is noticeably better. Not that the Sony is bad, but the difference is indeed visible on difficult subjects.


So what are the artifacts and on which common subjects do they appear?

1: mpeg high frequency artefacts appear on detailled subjects with lots of repetitive details. DV also exhibits them. Typical subjects are:
-plants: trees from the distance, grass from a few meters away, etc. (not leaves close-up). This is the most annoying problems: plants are common.
-human artefacts: brick walls, oriental carpets, some fabrics, printed material
-pebbles
2: mpeg motion artefacts will appear when the moving subject is tiny compared to the picture, and there is lots of details in the picture. You can film someone in the distance with a uniform background, not in front of plants...
3: chroma subsampling artifacts will appear when you have saturated colors, especially reds, in front of a white or light grey background. Try filming a red car to see it, for example.

I'm sorry, but you'll have to live with it. And that may be more a good thing than you think because whatever you film will end up distributed in mepg2 or mpeg4 format starved for bits anyway, be it DVD, cable TV blue-ray, or anything else. Only film avoids them (or rather replace them with other artifacts), and if you are shooting for film, you should consider Kodak and not Canon...

Besides even film ends up on DVD. And I suspect that some filmmakers know about the problem. For example, if you watch Stanley Kubrick's "Eyes wide shut", there are some scenes that look as if they were especially designed to crash mpeg-2 compression (and they do). Knowing that Stanley Kubrick was generally very interested in technology, I would imagine that he did that on purpose.

[Chan Ee Jien]

I don't think anyone here is denying not seeing any compression artifacts. Who gave you that idea?

[Jerome Marot]

Originally Posted by Chan Ee Jien

Who gave you that idea?

The first page of this thread.

[Ian G. Thompson]

Originally Posted by Joe Batt

The way that I've been able to make dark shots look better is to keep the gain at -3 and slightly overlight the shot, then darken it in post. It works for me. If I try to accomedate a dark scene by turning up the gain I always get grain.

This is exactly what I do and it is perfect everytime. I'm telling you this cam can take "perfect" no grain shots in dark scenery. If one were to leave the cam in auto with these types of shots then they would be disappointed. But thank God for the workaround.

[Jonathan Gentry]

Overlight the shot? Some people don't light shots. They have to work with what light is there. Not everyone here are producing movies.

Originally Posted by Joe Batt

The way that I've been able to make dark shots look better is to keep the gain at -3 and slightly overlight the shot, then darken it in post. It works for me. If I try to accomedate a dark scene by turning up the gain I always get grain.

[Jon McGuffin]

Originally Posted by Ian G. Thompson

Here's a thought...get an HV20 and an Intensity card...film those complex scenes solely with the Intensity card's HDMI input...totally bypassing the HDV codec and your problems are history. Still a heck of lot cheaper than an HVX with all of its accessories.

By the way...I never had those types of problem with my HV20..but maybe I need to start analyzing my pictures more closely.

Question: Why call HDV a fake HD? Is it bandwidth that makes HD HD...or the pixels in the picture or resolution? I mean...compression is one thing...but to say it's fake HD because of compression artifacts...I don't get it.

If I captured my "live" HDMI footage with Cineforms codec (at any of its rates) would you still say its not real HD?

I agree with this post. First of all, going with HD-SDI or HDMI with the HV20 would be your solution in those "special" complex scenes.

To call HDV a fraud or a fake HD is absurd because the only definition of HD is resolution. And as the original post mentioned, the cameras can and do produce many fantastic and beautifull HD content. I use two Sony FX1's and I've shot some footage that I thought was better than what I see many broadcast companies sending in HD to my television. In the end, whenever any compression is used, there is going to be a "breakdown" point where the lack of performance of a codec will be exploited. HDV was built and targeted at the prosumer market and frankly, I'm still blown away this codec can deliver this image quality and resolution inside the same 25Mb/sec as standard DV.

If you want more resolution and a better recording Codec, the HVX200 might be the ticket or maybe you should start considering RED, Silicon Imaging, or Sony's new XDCAM EX. Ahhh.. but then again, those camera's cost 3x - 6x more than a solid HDV cam like the A1. Something's got to give.. Calling out HDV as a fake almost feels disrespectfull...

Jon

[Richard Hunter]

Maybe fake is not a good word to use, but HDV is definitely a compromised form of HD. It has limitations that we need to be aware of, rather than deny, but for me it's an acceptable way to get high definition at a reasonable price.

Saying that HD is only about resolution is an oversimplification that is normally best left to marketing people. (e.g. "CD Quality" or "DVD Quality" to sell dubious recordings that happen to share the same sampling rate or video codec as something else). Footage that has HD resolution but with codec artifacts on moving fine details is obviously not the same as clean HD footage. Also, video with a data rate of 100Mbps is obviously going to be better quality (all other things being equal) than video at 20Mbps. HDV is impressive at being able to squeeze so much into the available bandwidth, but in the end it does have limitations.

And on a related issue, I also think it is fair that footage that obtains its HD resolution by upsampling from lower-resolution sensor blocks is considered by some to be not true HD. You can call it pixel shift or whatever you like, but it's not the same as having discrete sensor points for each pixel.

Richard

[Mikko Lopponen]

Strange. I've seen a lot of motion breaking up with the hc1, but the canon hv20 has held up extremely well even with high motion stuff (except for the rolling shutter). I would presume that the 24f mode in the xh a1 would be so good that the artifacts would be minimal. Atleast 25p on the hv20 is awesome in regards to motion artifacting.

[Carl Middleton]

I've worked with the Sony HDR3, A1 and Z1 (my baby) filming adventure videography. I took the A1 underwater, down ziplines, filmed from the roof of boats 30 miles offshore (with that sort of rocking motion of 3 foot waves, LOTS of attempts to break the codec).

I've generally been disappointed with the A1's abilities to properly capture in varying light scenarios.... and I really didn't like the HDR3 (no external control of my iris?!) But all-in-all, after filming a season of fast paced, fast motion video for international broadcast using the HDV codec, I've been happy. I've never had to give up a shot because the codec failed me, and that includes filming from a zipline. Those trees fly by pretty fast, and it still looked good! Most of the episodes got a kudos from the art department AND the tech department... and I personally think that speaks well for the flexibility of shooting with HDV cameras, as long as you tweak your settings dependent on the environment you are in.

Carl

[Chris Hurd]

Originally Posted by Richard Hunter

HDV is definitely a compromised form of HD.

All HD formats are compromised in one way or another. For example, the best one of all, HDCAM SR, is compromised in terms of its cost: it's incredibly expensive; in fact, prohibitively expensive for most folks.

footage that obtains its HD resolution by upsampling from lower-resolution sensor blocks is considered by some to be not true HD.

And yet HDCAM, the single most widely accepted format in the world for broadcast HD masters, is 1440 x 1080. So-called "true HD" is an effective marketing term for single-chip consumer camcorders which can't reap the benefits of Pixel Shift which are afforded to professional three-chip designs, but it's still just that: a marketing term.

Once the pro industry follows the lead established by most Digital Cinema UHD camera manufacturers out of three-chip systems into single-chip designs, all of this will be a non-issue anyway. As far as I'm concerned, it's a non-issue right now. The only question is, "is it good enough?" and the obvious answer is yes, 1440 x 1080 certainly is good enough, judging by the broad acceptance of HDCAM.

You can call it pixel shift or whatever you like, but it's not the same as having discrete sensor points for each pixel.

Pixel Shift is actually *better* than discrete sensor points for each pixel. Certainly not any worse.

[Richard Hunter]

Originally Posted by Chris Hurd

Pixel Shift is actually *better* than discrete sensor points for each pixel. Certainly not any worse.

Hi Chris. I'd be interested to know why you think so.

Richard

[Chris Hurd]

Sure. Here's the easy part first:

All things being equal -- say we have two 3-chip systems, pixel counts on the chips are identical in each, chip size is identical in each, one system uses Pixel Shift and the other doesn't. Obviously the advantage goes to the system using Pixel Shift because of the inherent resolution boost it provides by delivering more sampling points per pixel to the camera's A/D converter. But that's academic.

Now let's complicate things a bit:

All things *not* being equal -- say we we have two 3-chip systems, pixel counts on the chips are *not* identical in each. One system has chips with a lower pixel count than the other. Chip size is identical in each, but the system with fewer pixels on the chips also uses Pixel Shift. Let's say the resolution boost provided by Pixel Shift equals the native resolution of the other system. The advantage still goes to the system using Pixel Shift, this time because of greater sensitivity and better low light performance. Fewer pixels on the chip means the pixels are physically larger and therefore gather more light than a chip of equal size containing more pixels.

And yet another advantage for the system using Pixel Shift: fewer pixels means not only greater sensitivity, but reduced cost as well. For any two sensors of equal size, the one with more pixels is more expensive. Pixel Shift negates any loss of resolution, provides for greater light sensitivity and allows for a lower cost to the end user.

However, the main advantage of Pixel Shift as a resolution booster is the primary reason why it's incorporated in one form (H-axis only) or another (H-axis and V-axis) into almost every single three-chip camcorder currently available; especially within the affordable HD acquisition arena which the Canon XH series camcorders fall into. In fact there is only one popular three-chip HD camera system that does *not* use Pixel Shift, and that's the JVC Pro HD series; their reasoning for this is that Pixel Shift is not needed for 720p nearly as much as it is for 1080i. That said, every three-chip HD 1080i camera on the market today uses some form of Pixel Shift (to include Sony's ClearVid sensors in the HVR-V1U; these chips aren't using a traditional form of Pixel Shift but the resolution boost they get from diagonally offset pixels is practically the same concept for discussion purposes) -- just as most all three-chip standard definition camcorders always have, and with very good reason. Hope this helps,