Comparison Videos: HV20 and HDC-SD1

[Guy Bruner]

I have posted two comparison videos shot simultaneously with the Canon HV20 and the Panasonic HDC-SD1. The videos were shot in auto with normal shutter. Both videos are low light. The flowers video was shot at dusk after sundown and the night video was shot well after sundown.

The Panasonic's AVCHD was converted to MPEG2 1920x1080i @ 25 Mbps using Nero Vision. Then, the files from both camcorders were brought into Sony Vegas for titling, light editing (just trimming), and merging. The final video was rendered out to HDV 1080i. The conversion of the AVCHD to MPEG2 degrades the source slightly, but the softness shown by the SD1 is present in the original source video as well.

Flowers 140 MB M2t
Night Town 84 MB M2t

[Paulo Teixeira]

It’s a good comparison but this is only comparing the NTSC models. Since the PAL version of the SD1 allegedly have a better decoder inside it would have fared better against a PAL version of the HV20.

A comparison between the AG-HSC1U and the HV20 will be interesting.

Guy, do you think you would do a comparison between the HV20 and the HC7?

[Guy Bruner]

As far as I know, the PAL and NTSC encoders are essentially the same...basic difference being framerate. There is some speculation on a German site that the encoder was improved post CES, but that has never been officially confirmed.

Sure, I'd throw an HD7 into the mix...if someone would get one to me ;-).

[Paulo Teixeira]

Originally Posted by Guy Bruner

As far as I know, the PAL and NTSC encoders are essentially the same...basic difference being framerate. There is some speculation on a German site that the encoder was improved post CES, but that has never been officially confirmed.

Sure, I'd throw an HD7 into the mix...if someone would get one to me ;-).

I do wish Panasonic would at least say something about it.

I was talking about the HC7 not HD7 but I could see how easily someone can mix that up. There are too many camcorders with similar model names.
The HD7 would still be a good comparison against the HV20 and the HC7.

[Ken Ross]

Originally Posted by Guy Bruner

I have posted two comparison videos shot simultaneously with the Canon HV20 and the Panasonic HDC-SD1. The videos were shot in auto with normal shutter. Both videos are low light. The flowers video was shot at dusk after sundown and the night video was shot well after sundown.

The Panasonic's AVCHD was converted to MPEG2 1920x1080i @ 25 Mbps using Nero Vision. Then, the files from both camcorders were brought into Sony Vegas for titling, light editing (just trimming), and merging. The final video was rendered out to HDV 1080i. The conversion of the AVCHD to MPEG2 degrades the source slightly, but the softness shown by the SD1 is present in the original source video as well.

Flowers 140 MB M2t
Night Town 84 MB M2t

Guy, I'm in the process of downloading your clips, but what's your impression of the comparable video quality of these two?

[Ken Ross]

Well frankly I'm surprised. I just loaded the 2 clips on to tape and played them on my 50" Fujitsu plasma. I was surprised by the softness of the Panasonic. I had just read a glowing review in this month's Perfect Vision. The reviewer gave the image quality 4 1/2 stars out of 5, the same rating he gave the Canon, so naturally I was curious.

At first I thought the SD1 just wasn't focusing on the subject, but no matter what clip was played, the softness remained. I also thought the colors on the Canon were a bit better although not seeing the original subject can make that a tough call. I was mainly focusing on the greens which looked more life-like on the Canon.

The only thing in favor of the SD1 was the low noise (as mentioned in the Perfect Vision review), but anything that soft will generally have extremely low noise. You also mentioned that the softness was in the original footage even though some sharpness was lost in the transfer.

[Emmanuel Bertin]

The pixels of the Pana SD1 and JVC HD7 are much larger than those of the HV20; both cameras use the "pixel shift" trick to generate a well-sampled HD image from 3 overlapping lowres sensors. Larger pixels integrate more light and provide higher S/N at a given aperture/exposure-time, but inevitably smooth out the smallest details. I assume that sharpening is applied within the camera to compensate partly for this effect. Nevetherless this won't get rid of the "boxy" aspect of small details caused by the larger pixels.

[Guy Bruner]

Sure, I'd throw an HC7, HD7, FX7... into the mix. Just box them up and send to...

The two cams are close in image quality, IMO. The Canon is sharper and the colors are more accurate compared to the actual scene. However, I wouldn't throw either out. If you weren't comparing the videos side-by-side, you probably wouldn't discern a difference. I think that is why the reviewers have given a comparible score to the two cams. Either way, the buyer is a winner. IMO, the only differentiators are the other features on the camcorders...and the ability to edit the footage. Ugh! AVCHD is a bear right now.

[Paulo Teixeira]

Pixel shifting is not as bad as it sounds.
From all the reports I have read, the SD1 and the DX1 at 13mbps produced a better picture than the Sony SR1 at around 15-1/2mbps.
The quality of the decoder also determines the picture quality.




Just wait until you see the picture quality of the Panasonic AG-HSC1U compared to the SD1 or even the HV20.

[Ken Ross]

Well, as I've felt in the past, tape-based camcorders are still providing the best images. It may not always be the recording methodology that makes this true, but nonetheless the non-tape based units just don't seem to have quite as good a picture as the tape-based units.

Add to that the nightmare of editing some of these new formats and my feeling is "I'll pass on these right now, thanks".

[Paulo Teixeira]

Originally Posted by Ken Ross

Well, as I've felt in the past, tape-based camcorders are still providing the best images. It may not always be the recording methodology that makes this true, but nonetheless the non-tape based units just don't seem to have quite as good a picture as the tape-based units.

Have you seen footage from XDCAM-HD or DVCPRO-HD?

It doesn’t matter where the footage gets recorded to.

[Chris Hurd]

Originally Posted by Paulo Teixeira

Pixel shifting is not as bad as it sounds.

Not only is it "not as bad as it sounds," it's also a very good thing to have.

You want Pixel Shift in a 3-chip camcorder. Most all 3-chip camcorders use it, with rare exceptions.

[Emmanuel Bertin]

You want Pixel Shift in a 3-chip camcorder

Sure, but these 3-chip "hi-def" camcorders only have 500kpixels per chip.

[Ken Ross]

Originally Posted by Paulo Teixeira

Have you seen footage from XDCAM-HD or DVCPRO-HD?

It doesn’t matter where the footage gets recorded to.

Paulo, you used as an example a professional and very expensive system used by major networks that's not exactly within the reach of consumers or semi-pros. Not exactly fair to compare that. Do you know how much the equipment or recording media costs in that system?? Here we are talking about consumer and semi-pro equipment.

But to your point, in a sense it DOES matter where the footage gets recorded to. Why? Because manufacturers of consumer equipment, when using storage such as hard drives and memory media, know that storage space is limited. So what do they do? They compress the signal to maximize the amount of recording time. So yes, you are correct that technically the media doesn't matter, but when manufacturers are faced with a storage size limitation imposed by the media itself, decisions are made and quality is often compromised. That's simply a fact.

Tape of course imposes much less of a limitation. Compare the storage capacity of a 1 hour mini-DV tape with other recording media and you'll see why manufacturers don't need to tackle the issue of compression in the same way. At $5 a pop, popping another tape in the camera is simply no big deal.