AHH!! DV lossy!!! whyyy you never warm me!

[CCEncoder]

According to this thread:
http://forum.doom9.org/showthread.php?t=12683
DV is loosy vs Huffyuv that is lossless...!!!!
Should I change to Huffyuv then? Does this mean DV has less quality then the other? I also heard DV has some pixel aspect ratio not being square issuee! what does this mean... And what about MJPEG codec is it good? And if I want Huffyuv I know I can't use DV Dazzle Bridge anymore can't it? what are my chances then? :o

[stars]

Hi....

Well DV has a data rate of approx 25 Mbit/s....

I dont think you will see any loss of quality at this data rate....

And about mpeg....well if you use a data rate of 7.5-8 Mbit/s (mpeg2) you wont spot any differens...

Many DVD players have trouble with data rates higher than 8.5 Mbit

And the new tv format HDTV uses mpeg4 compression and everybody thinks it looks great....


So...stay cool and encode......


stars....

[JohnnyMalaria]

According to this thread:
http://forum.doom9.org/showthread.php?t=12683
DV is loosy vs Huffyuv that is lossless...!!!!
Should I change to Huffyuv then? Does this mean DV has less quality then the other? I also heard DV has some pixel aspect ratio not being square issuee! what does this mean... And what about MJPEG codec is it good? And if I want Huffyuv I know I can't use DV Dazzle Bridge anymore can't it? what are my chances then?

Without any idea what you are wanting to do, who can say?

Sure, Huffyuv is lossless. And, yes, of course DV is compressed.

What is your source? If it's DV already, you will gain NOTHING by using Huffyuv.

Huffyuv isn't better or less quality. It depends on the history of the YUV data beforehand.

Pixel aspect: very simple. Your everyday, standard TV has an aspect ratio of 4:3. That is, the ratio of the width of the video to the height is 4:3. This is in the analog world.

NTSC video has 480 lines of video information per frame. To get pixels with a 1:1 aspect ratio, you need to sample each line 480 * 4 / 3 = 640 times.

For DV, it uses the D1 sampling rate - 720 times for each line across the video. So they are more pixels and they are narrower than the 1:1 version.

For PAL, you need to sample each line 576 * 4 / 3 = 768 times. For DV (PAL), it still uses 720 across the frame. So the pixels are wider than the 1:1 version.

But it doesn't matter.....

[jagabo]

Should I change to Huffyuv then (instead of DV)?

It depends on your video source, how you capture it, how you process it, where your ultimate output is going, and how much you care.

[CCEncoder]

Hi....

Well DV has a data rate of approx 25 Mbit/s....

I dont think you will see any loss of quality at this data rate....

And about mpeg....well if you use a data rate of 7.5-8 Mbit/s (mpeg2) you wont spot any differens...

Many DVD players have trouble with data rates higher than 8.5 Mbit

And the new tv format HDTV uses mpeg4 compression and everybody thinks it looks great....
So...stay cool and encode......
stars....

He even says he will prefer Huffyuv over DV to obtain highest quality for DiVx! I'm a quality guy so I'll give Huffyuv a try... By the way how do I use it? Any good advice? Oh he says all DV codecs are loosy and Huffyuv is the only code lossless is this true in all cases?

[edDV]

CCEncoder will you tell us what you are trying to do?

Quickcam?

VHS transfer?

$60K broadcast studio camera?

What do you want to do with it? DVD? HD/BD DVD? Stream?

[jagabo]

he says all DV codecs are loosy and Huffyuv is the only code lossless is this true in all cases?

It depends on your source and how you capture it.

If your source is RGB, say from a computer rendered animation, then both HuffYUV and DV are lossy, HuffYUV less so. The conversion of RGB to YUV will lose a little accuracy though this is generally imperceptable unless you do a lot of downstream processing. The further conversion to YUY2 for HuffYUV will lose some color resolution. DV will lose even more color resolution and potentially generate DCT artifacts.

If your source is a video capture card that captures YUY2 video from a high quality source, HuffYUV is lossless. DV will have less color resolution and and, potentially, DCT artifacts.

If you source is a DV camcorder and you are capturing via the IEEE 1394 port the DV is lossless -- the DV data that's on the tape is simply being transferred to your computer. There is no point in converting to HuffYUV except in some specific circumstances (like several generations of processing and editing).

[edDV]

he says all DV codecs are loosy and Huffyuv is the only code lossless is this true in all cases?

It depends on your source and how you capture it.

...
DV will have less color resolution and and, potentially, DCT artifacts.

That would be true only it the source color bandwidth was wider than 1.75MHz for U and V. There is nothing in the typical home other than RGB VGA that qualifies at SD resolutions, that is unless you buy some used broadcast gear. Even so RGB capture isn't an option with consumer products either.

Affordable capture devices can sample 4:2:2. Doom9 folk tend (historically) to convert most things back to RGB for filtering. If you do a lot of that, then there may be some benefit to 4:2:2. They should be capturing 4:4:4 for pixel space purity.

There are many disadvantages to Conexant chips as well. Very very consumer quality.

The process design depends on your project source, destination and filtering needs.