.WMV9 encoded quality and filesize vs MPEG2

[Achilles_wf]

OK I encoded a 3 minute clip last night in both TMPEnc to .m2v and WMV9 encoder to .WMV. I used VBR 6000Kbps average bit rate for video in .WMV and VBR 8000Kbps in TMPEnc. Otherwise the default settings for the most part, except that I kep the video size locked to my source for .WMV (instead of letting it go to the bigger size it seemed to be heading based on one of the dialogs).

Resulting file sizes are roughly the same (source was 650Meg, .WMV was 150Meg and .m2v was 170Meg). I'd say the .m2v file quality visually was superior. The .WMV file had some odd horizontal artifacts (like misaligned scanlines), for instance when paused every other line was slightly out of sync with where it should be given its neighbors. Maybe I could fix this with some hardcore investigation of the interlacing settings, but as my source material is pretty standard I wouldn't think I'd have to do that.

I'm just encoding a DV_AVI file captured straight from my DV camcorder. I am wanting to archive my home movies, and was planning to go with .WMV9 probably except that with file sizes not much less than standard .MPEG2 and quality that isn't as good (at least not without getting crazy with low level manipulations) now I don't know.

I was really expecting the .WMV file to come out significantly smaller and still be almost as good quality. I used what seemed to be a pretty high quality setting (at least, it took 30 minutes to encode my 3 minute clip!) and I also kept the image size locked to that of my source material (rather than lettin WMV9 make it bigger as it seemed to want to do).

Can anybody explain how/why .WMV9 can make a relatively similar quality playable video file to MPEG2(4?) at a savings in file size that makes it worth using as a codec? Ideally I'd like to be able to get 2 hours of high quality video on a single 4.7G DVD.

Also, one other question. After the first pass in WMV9 the file size was like 60Meg or so. Since it was on the second pass I was assuming the file size wouldn't change much from that (presumably it was just improving some things here and there). When finished with the second pass however the file was 150Meg! Why would the second pass add so much to the file size?

Thanks!

[shulthise]

Achilles_wf,

The first pass only gathers data on the entire encoded section/video. It does so for optimization reasons - that's why you do 2nd pass.

The actual encoding takes place during the 2nd pass, taking into consideration the data collected in the first pass, and the process parameters (avg. bitrate, max bitrate, etc).

As I recall, most encoding applications and plugins explain that you should typically do a second pass when the video involves fast motion and/or high lighting, contrast or patterns.

If you're encoding family videos you probably don't benefit from doing a 2nd pass since typically family videos lack the characteristics i specified above.

I know you're aiming for highest quality, but you gain nothing with 2nd pass, and you lose a lot of time.

Cheers,

Shulthise.

[RNCSerge]

I don't think there's any way MPEG2 could beat WMV9 at lower bitrates. Try de-interlacing the footage. Takes more time, but it should make your encoded WMV look better.

[shulthise]

DO NOT DEINTERLACE!!

DO NOT DEINTERLACE!!
At least not with WM technology.

A deinterlaced video has a fuzzy-blurry look to it. You'll notice that mostly on border areas, like the edge of a table or hairline.
This is a serious quality reduction.

Also - if your objective is to play those videos on TV, then you should definitely1 not de-interlace the video, since TV is designed and built to play Interlaced pictures.

[Achilles_wf]

Thank you for your replies.

Current TVs like interlaced pictures...but I'm preserving these videos for 20 years from now when I'm old and my kids have moved out of the house and I want to re-live my glory days! By then, certainly TVs will all be progressive, right? I say this because my current plan is to use de-interlacing when encoding with WME9, which I found produced the best looking output on my PC monitor anyway. I didn't find it fuzzy at all.

Also, I didn't use the allow non-square pixels option. Tried it once and it didn't seem to do anything but add some black bars on the sides (or top/bottom I can't remember) of the resulting encoded video. Should I learn more about this before blasting through 150 hours of video to DVD?

Thanks!

[Achilles_wf]

Sorry to resurrect this post! But now I see where I was told NOT to deinterlace in WME9, and I want some other thoughts on the subject if possible.

Thanks!

[shulthise]

Warning - I'm writing this reply after a night with no sleep

Sorry to be the one bringing bad news, again, but I do think you should be aware of that today instead of 20 years from now…

My objective – like yours – was to create a video file that would (a) be significantly smaller than the DV file (backup to DVD media?) and (b) maintain the highest possible quality.
This would allow me to re-use the expensive video tapes (yes, where I live it’s almost twice as much expensive than in the USA, while the economy is ****).

WMEncoder suited my objectives perfectly – it was available, free, and easy to use. (The documentation target audience is the lamer – not the video pro – that suited me too.)

After doing a considerable number of tests on various DV files – each 2 minute in length, to save time – I determined the right configuration for me and encoded a number of my digital video files with WMEncoder.

My test results showed that de-interlacing a video while encoding with WME will result in a light-blur of the video – just as if you applied a very light blur filter on the video.
As I gather, de-interlacing changes the video data by merging two frames that were created to be displayed one after the other. It’s like taking two consecutive cartoon animation frames and displaying them at the same time, one over the other (hence the blur effect?).
I can’t explain otherwise. Like it or not - you can deny it or ignore it – it’s there, and you can’t help it.

To have the highest quality you need to tamper with the video data as little as possible. De-interlacing tampers with the video data – so don’t de-interlace.
You can always do that in the future – 20 years from now. But if you de-interlace it now, and delete the original – you can’t have the same quality as the original, and you can’t re-interlace it to the original quality.
Remember that in the future you’ll have powerful machines, storage and software (or house of stone and radio-active water.. hehe) that would be able to do with the original interlaced files things far greater than you can today.
Also – with the adaptation of non-interlacing display devices – multimedia companies will find a huge market for interlace-to-noninterlaced applications – be sure that they’ll think of all those billions of interlaced videos people have.

I hope this was not too confusing.. Didn’t get much sleep last night…

Cheers,

Shulthise

[Achilles_wf]

...house of stone and radio-active water...LOL

OK, I understand I think. You're saying that I can de-interlace later and not introduce bluriness now. The only thing is I was hoping not to reencode again, but rather just have a set top box that would play raw .wmv files (which seems likely), and a non-interlaced based display device that was happy with the data as is.

Mainly, the video looked great on my PC even with de-interlacing on. I suppose I should burn a DVD with and without and test on my TV, but like I said I figure in the future TVs won't be made to handle interlaced signals natively so who cares which looks better right now. I'll try to do this anyway and if the quality is really noticeably worse when de-interlacing I'll archive without it.

So you yourself use .WME9 for archiving your DV home video? Please confirm if this is the case...I've seen a number of your posts here in these forums and you really seem to know what you are talking about. So it would be nice to hear that my codec of choice is being used by someone that's looked into this as much as yourself.

Thanks!

[tigerman8u]

read here for a good guide on interlaced vs. deinterlaced

https://www.digitalfaq.com

[shulthise]

Well...

Most of my videos are backed up in WMV9 format. I reduced the bitrate to 8000bps, which is a lot.. maybe even too much. But like you – I'm a quality freak.

The reason I used the WMV9 format is the same reason why most successful things in the world are used – it’s available, accessible, free, easy to use, and spread just about everywhere in the internet. So the question is Why shouldn’t I have used it ?!

Think about it this way – most chances are that you wear jeans when you’re off work, hanging out, on a date, even at work... why ? same reason you would use WMV!

I'm not against WMV – I'm going to keep using it.

BUT!

I'm not going to backup my tapes with this format. I found most MPEG4, DIVX, and WMV codecs have these drawbacks I don’t like:
1. They feel as if they were created for low bitrate streaming rather than hard high bitrate to the screen. Sure you can use high bitrate for them, but the encoding and decoding algorithms are still optimized for streaming. I'm not sure about that, but this is the feeling I always get when I play these codecs.

2. Editing a video or producing a DVD from a WMV/MPEG4/DIVX source video is SLLLLLLLLLLLLLOOOOOOOOOOOOWWWWWWWW. I want to use the videos I backup to create DVD’s of Junior (he’s 2yo) and send it to dad and mom and those-guys-in-law... when I tried to do that I had to re-encode the entire WMV file back to DV, and only then it would work OK.

I found the answer in MPEG2. It may not be as economic as WMV/MPEG4/DIVX, and it may not be editable as DV, but it’s not too slow for editing, and it renders great quality!
The second half of the answer was – no matter how expensive are the DV tapes – Junior is worth it!

There! Hope my answer satisfied your curiosity
If you do find the perfect answer for the Quality vs Size BE SURE TO LET ME KNOW!

Cheers,

Shulthise

[Achilles_wf]

ArghhhhH!

OK, now I'm back wondering again. I certainly do hope to move my .WMV files to playable DVD at some point. I was hoping I wouldn't have to re-encode (if I didn't want to edit them), since at some point there will be set top boxes to play .WMV content directly (right?).

But if you are telling me that I will have to re-encode to DV (or whatever) then to .mp2 anyway, perhaps I should just archive in .mp2 format anyway?!? I really did do some major comparison runs that showed .WME achieving better quality for given file sizes over some of the .MP2/4 encoders, but it wasn't THAT much.

I really just need to decide on a format for archiving...I certainly don't want to use a format that will penalize me with a double conversion quality degradation in order to go from my archive file to a playable DVD format.

One other small item...when I capture my DV source to PC, do I go ahead and let it make a giant 30G file that I then encode, or is it better to in some way break up the resulting .AVI file into chunks, convert those independently and then in some way join them to get a single .mp2/v file ultimately. Thanks!

Achilles_wf

[shulthise]

If you play the WMV as is, no editing, than there's no need for an additional re-encoding to DV or MPEG2.
If you find a DVD able to play WMV files.. it's even better.

My point was that i backup the captured file unedited.
Only when i create a DVD movie out of the backup file i need decode it to DV because of slow processing and rendering.

At 8000KB i don't think you'll notice the quality differences between WMV and MPEG2 with your eyes... But if you really want push it - Encode your backups at 12000/16000KB and back it up to a double layer DVD. The hardware is there... all you need is the time and money for it.