Sometimes the best way to find out how little we know about a subject is trying to teach it to others. This happened to me while trying to explain video compression.

For example, the new Avid XDV Pro claims to handle “uncompressed” or 1:1 video. But all digital video is compressed. At least 4:1:1 – correct? So what does “uncompressed” mean?

My ill fated lecture emanated from trying to explain why it would not make much difference if we transferred some VHS material to DigiBeta or to DVCam. The plan was to transfer it to both and use the DVCam version to “off-line” the edit in the Avid XDV and then “conform” with a linear edit off the DigiBeta.

I’ll stop my show of ignorance and confusion here. Can someone explain or at least point me to a place where all this compression stuff can be laid out in a way I can understand?

Thank you in advance.

Well, I don't have any good websites off hand, but in my "easy-to-understand" words, it just means that the video signal is reduced to keep the file size down. Of course, the more compressed something is, the worse the quality.

If your talking about transfering VHS to MiniDV/DVCAM....theres no comparison, between VHS and DV...VHS is a old format thats been around from the 70's, DV is newer technology, been around since 1995. VHS has only 250 lines of resolution, vs DV's 525...VHS is analog, DV is digital.

i don't kow the new avid cut system, but i suppose they say "uncompressed" but mean the normal compressen you mentioned.
i found something about compression here: http://www.adamwilt.com/
take a look at it, perhaps you find your answer there

Thank you Elmar and Alex. But I know the differences between all the formats, digital and otherwise. My question (and the Wilt site does come close to explaining although it never hits it on the nail) is simply - what does Avid mean by "uncompressed" video when digital video is compressed even as it being recorded. When digitizing from DVCam or MiniDV, as far as I have understood, there has been no further compression - not like the old AVR3 or 4 which we used for reducing file size in Media Composer.

From the Wilt site I can see why and how DigiBeta has the edge over DVCam or MiniDV (although minor to the naked eye) but not what this "uncompressed" nomenclature is all about.

hmmm... the video is stored compressed on the dv tape but it is transfered uncompressed to the hdd and back and could "uncompressed" be edited (without any more compression, cause it is compressed, it doesn't make any sense blowing it up to uncompressed). i think avid uses the same kind of material than any other program, and thats the material that comes from xl-1 and co, and that is stored on tape in compressed form. i don't know any consumer, prosumer or something like that that stores dv uncompressed, it is always compressed but in a way that doesn't really affect the image

I've found a faster method to identify my ignorance. Instead of waiting for someone to "explain" to I explain it to myself, out loud of course, so my fellow commuters think I'm over the hill. Then, when I can't, like you I go looking. I've just spent 15 minutes trying to Google my way to an answer and this is as close as I've got: Uncompressed NTSC video is 27MBps. Uncompressed DV would be 18MBps, based on the fact that what you capture comes out at 3.6MBps after a 5:1 compression in-camera. Since as you point out DV is sampled 4:1:1 I'm guessing that one gets to 27MBps by sampling at 4:2:2, which higher end acquisition equipment does (digibeta?). Clearly, capturing a stream that weighty requires more than a 7200 rpm HDD, so most of us will never play there. Now perhaps someone who actually knows can explain it to all of us.

David Hurdon

David,
The new Avid DNA system will allow you to edit "Uncompressed" meaning 601 video. SD video as well as DV.

www.avid.com will help answer your questions.

Try this, maybe:
http://www.avid.com/community/forums/forum39/html/000003.html

I don't think you will find the (exact) answer you're looking for, but reading the thread might be helpful anyway...

Thanks, Stylianos. Now, when I talk to myself about this I'll make that much more sense. ;~)

David Hurdon

Richard, Stylianos, and David - thank you all for your replies. I'll check out the thread Stylianos suggests. Meanwhile, it is what Richard says that rings a very loud bell - "SD video as well as DV". If SD stands for serial digital then the "uncompressed" term means it can handle D2 in its native state. That's quite a feature for less than US $2k. I'm hot on the track now. And David, for me it's not my fellow commuters - it's my wife who thinks I'm over the hill, but then, she's been talking to herself for years - I just haven't noticed. ;-)

Vegas is calling DV uncompressed as well, although it really
ain't (twice compressed). Now the 4:1:1 you are talking about
isn't like the digital compression where we say DV is 5 times
compressed. 4:1:1 is the videospace where we get 4 samples
of Y (luma) and 1 sample of U & V (color offsets). Ofcourse
this is a form off compression as well because you will "loose"
color resolution.

Generally when we talk about compression we are not talking
about color spaces but actually reducing the whole information
stream.

Rob,

I'm getting closer to understanding. I'm aware 4:1:1 or 4:2:2 doesn't relate in any way to the old sense of compression. In the old, ancient, Avids we would off-line at AVR-3 or 4 to save on drive space but the image quality was strictly for off-air viewing. Then we would either re-digitize at, say, AVR-77 or go straight linear and edit from the camera originals. That made sense - we knew AVR-77 was compressed but air quality, and linear, of course, was not

Then digital came and the AVR standards were cast away. Now we have ratios - 1:1, 5:1, 4:1:1, and so on. What's confusing is the insertion of the idea that AVID XDV Pro can edit "uncompressed" video. My search is trying to find where that uncompressed video is coming from and how much drive space it takes up. The only clue as to where it is coming from is D2 which, as far as I can tell, in uncompressed digital and used mainy in post. Although nice to know, what other digital format is not compressed. None really. And does it make any practical difference?

As an aside, this compression issue is a very old one. Years ago (in the early 90s) we began to get complaints from the linkup people about our compressed video. They claimed it was causing havoc with the satellite signals. I think they are still complaining but resigned to the fact that comression is the status quo.

If they really support uncompressed video at normal TV resolution
it will be around 30 MB/s (megabytes per second). For HDTV it
does up ofcourse and if it is lower then it will be compressed.

<<<-- Originally posted by Rob Lohman : Vegas is calling DV uncompressed as well [...] -->>>

No, I don't think we do. If I'm wrong, please tell me where. We all know DV is compressed, not just in colorspace (4:1:1 NTSC or 4:2:0 PAL) but also using DCT lossy compression.

///d@

When I click a file in the media explorer it says:

Video: 720x576x24; 25,000 fps interlaced; 00:07:00:03; DV
Audio: 48.000 Hz; 16 Bit; Stereo; 00:07:00:01; Uncompressed

I see where I went wrong with this. Uncompressed is for the
audio. Video is DV (compressed). My bad. Is audio really
uncompressed in DV? I thought it was PCM (ofcourse this is
not a lossy compression but it still is a compression, right?)

While we are at the topic of Vegas. I noticed Vegas is using
some really odd Pixel Aspects that no-one else seems to use
(I checked this with Premiere, LightWave & Digital Fusion):

PAL:
1.067 Premiere 6.0
1.0667 Lightwave 7.5
1.0926 Vegas 4
1.0667 DFusion

NTSC:
0.9 Premiere 6.0
0.9 LightWave 7.5
0.9091 Vegas 4
0.9 DFusion

Can you tell me what is going on with those? I'm working on
a Pixel Aspect calculator and this struck me as odd. Also no-one
I have talked to in the business or otherwise could explain
these numbers....

Thanks!

PCM (Pulse Code Modulation) is uncompressed. WAV and AIFF are both PCM.

Pixel aspect ratio doubts come up from time to time. There are a bunch of threads on our forum about it. Trust us, it's right. We've offered a free copy of Vegas to anyone who can prove it wrong.

///d@

I thought PCM was a lossless compression. Oh well, we learn
every day.

I want to trust you that the PA is right. BUT, every other piece
of software uses a different PA. Both LightWave and DFusion
are being used in a very serious working environment (I'm not
saying that Vegas cannot or is not used there, but I have not
heard of it yet) where it MUST be correct.

If I am not mistaken the calculation for PA is:

Pixel Aspect Ratio = (Overscan Display Height/Overscan Display Width)*Display Aspect Ratio

If I do this for NTSC or PAL I get:

(480/720) * (4/3) = 0,88888888888888888888888888888889
(576/720) * (4/3) = 1,0666666666666666666666666666667

Which are different from Vegas' 0.9091 and 1.0926. Again, my
primary concern is why everbody else (including the high end
packages) are using different PA's then you guys are.

Does this count as PROOF? :)

The below is a quote from a different thread and might shead some light on the differences between Vegas and Premiere in their reporting of the PAR. I think that there is no other difference than what comes from rounding:


"I have done some additional reading and I think that I'm beginning to understand. It seems as Vegas and Premiere uses different basis for the calculation. Vegas seems to use the a method including also the fraction of non-square pixels (702 + 54/59 non square 59/54 pixels per line) that fits in 768 square pixels per line whilst Premiere excludes the fraction. Although the figures differ the visable result might be the same?

Include links to what I have read in case anyone is interested.


http://geocities.com/wunder01au/widescreen.html

and

http://www.mir.com/DMG/aspect.html

(From the thread: http://www.dvinfo.net/conf/showthread.php?s=&threadid=6482)

Can you elaborate some more on the 54/59 and 59/54 comment?
I'm not following it. Can someone show me a calculation that
gets us the numbers Vegas is using?

I'll try by using the figures in the "mir" link in my previous post.

The PAR for PAL video is 59/54 (i.e. they are 59 wide by 54 high). The AR on PAL widescreen will then be ((59/54)/(4:3)*(16/9)) = 1,4568 (the AR Vegas reports for widescreen).

Premiere reports 1,422 as the PAL AR for widescreen. It seems (I'm not sure) that they have arrived at this number using the formula ((1024/720)/576), in which 720 is the 4/3 ratio and 1024 is the 16:9 ratio if the hight is 576.

Hans

<<<-- Originally posted by Rob Lohman : Does this count as PROOF? :) -->>>

Hardly. :)

Vegas uses the correct PAR, but if you need to set it to something else in order to interface with programs that got it wrong, we let you set it in the project properties and many of the render dialogs.

///d@

Gee, and I thought I was asking a simple question. Judging from the lengthy discussion, I see I am not alone in my confusion.

Does ANYONE REALLY KNOW WHY Avid refers to XDV Pro as capable of handling UNCOMPRESSED video? If they mean digital video that is not compressed any more than it already is in its native state, then I can understand, but I thought that is what XDV did from the very beginning. IF Avid means it can handle, say, analog Beta SP, without compression, I also understand; but nowhere has it been described in clear, solid terms.

I’ll keep trying to find a simple answer. By the way, I’ve also asked a few “old time” engineer types who work with Avid MCs, they haven’t been able to come up with clear, positive answers, only guesstimates. So none of us should feel too bad about falling short of understanding.

Perhaps you should send an e-mail to their technical department
or ring them? (AVID that is).

From the Avid website...

UNCOMPRESSED STANDARD DEFINITION VIDEO

Looks like SD 601 to me.