BD/720p.MKV -> XviD conversion: desired avg. DRF 3.0?

[Randm]

i don't really know how DRF works, so sorry if this sounds silly...

-i mostly try to get around 2.8-3.5 average DRF with DVDRip.XviD
-i want to make the same DVDRip.XviD quality, but with my BluRays and HD.MKV rips as the source
-do i need to change my desired avg, or will 2.8-3.5 yield the same results regardless of source (using same output settings)

[manono]

So you're doing 1-pass CQ encodes? If so, Q is Q and the results will be the same, no matter the source, although the final filesize might be larger than one using a DVD as a source, all other factors being the same (final DRF, XviD Settings, resolution and other filtering).

And you make encodes for specific DRFs, but have no idea what that means? Hehe. Then how are you discovering the final DRF? By using DRF Analyzer? Anyway, DRF stands for Detail Removal Factor, and it's pretty much another name for quantizer. However, doing an encode for a quant of 3 will wind up with a higher final DRF, how much higher depending on how you set up your B-Frames.

[Randm]

k, thanx for the info

i do size-based rips...
-2pass
-512-640+ width
-128 mp3 (mostly vbr)

i use DRF analyser to check my results...my general guideline before ripping is FPS x Runtime x .28
ie. 23.976 x 144 x .28 = 966.71, so i make a 1400mb or 2CD rip...

btw, i duno if that formula makes sense, but my rips seem fine using it as a guideline?

[manono]

I see - 2 pass encodes, adjusting the resolution so that the result will fall within the DRFs you like. No, it doesn't make much sense, since you're using the standard 2 CD B/P*F .28 figure, which doesn't take into account the compressibility of the source. Have you considered just running a quick compression test to help determine the resolution? That way you could get a pretty good idea of the final DRF before running the encode.

But that's OK, if your method works for you. Better than no method at all. Even your highest 3.5 DRF is pretty low, so, unless you wind up using a low resolution and everything looks overly smooth, I expect your results are pretty good.

[Randm]

well i mostly already know which res. i want for each aspect ratio...sometimes there is trial and error with rip size regarding compressibility, but i can mostly guess which movies are worth 3cd beforehand like action/cgi-heavy video...problems don't happen much for me anyways since im usually too broke to buy movies often lol...regarding compressibility test, only way i know how to do that is using autogk...which i can't use with BD/.MKV input

[manono]

Oh yeah, Blu-Ray and MKV. I forgot about that. I'll assume you're using an AviSynth script and, if so, you can probably adapt the original compression test to your needs, the one developed for use in Gordian Knot. Either that or just add:

SelectRangeEvery(280,14)

to the bottom of the script. That says to encode 14 frames in a row out of every 280 frames, or 5% of the total. The GKnot test then drops the first of the 14 frames when figuring the result because it'll be a keyframe (I-Frame), and will skew the results. You could just lower the final file size by a little bit. I do this kind of thing myself quite a bit to get an idea of the final file size for a given resolution, and filters, and quality. I'll run a quick 1-pass at quant 3, multiply the final size by 20, subtract a very little bit from the final size (10% maybe), add in the audio and a little bit for the overhead and get a good idea of the final size. But as I said earlier, the DRF will be higher than the Q, because of B-Frame settings. B-Frames are of lower quality (higher quant) usually than are the I and P-Frames. Depending on your B-Frame settings. johnny developed a good compression test as well and has some more information about the process:

http://forum.doom9.org/showthread.php?s=&threadid=44414

In any event, your method works OK. You might consider switching to the use of DVDRs instead of CDs on which to burn. That will allow you to run the much faster 1-pass CQ encodes, and you won't be hampered by having to make them all for a certain number of CDs.

[Randm]

thanx for the info...btw, i don't burn them to CD, i'm just used to CD sizes over the years ...plus, if it needs to be over 3CD xvid then i just stick to BD or 720p.MKV...soon as i'm done figuring out basics in BD/MKV->XviD rip, i'm gonna start trying out the CQ method...peace

PS. btw, how do i data burn video over 2GB file size for standalone? i'm used to basic data DVD method for XviD/etc., but that doesn't work with over 2GB filesizes...do i pick UDF or UDF/ISO hybrid, or what...

[manono]

do i pick UDF or UDF/ISO hybrid, or what...

I don't know anything about that stuff. Sorry. Maybe someone else can help. Better might be to make sure they don't go over 2 GB. Or encode in two parts. Or split it afterwards.

[Randm]

kool, i'll figure it out, or i'll just split them, etc...i'm all out of questions for now peace

[Randm]

i also want these XviDs to play on an old philips DVP642 for my parents...just out of curiosity, do any of these settings affect compatibility? i'm mostly guessing from what i've seen around, so feel free to correct me and/or suggest which options to pick for best quality and compatibility...how long the encode takes does not really matter...

-vhq mode: 4-wide search ok? is this just analysis or does it affect compatibility?
-b-frames: max, 0-4? does this affect compatibility?

-closed gop?
-vhq for b-frames?

[manono]

The players I've had have used a different company's chipset (the MTK by Sigma Designs) from that of the Philips player (ESS chipset), so I'm not completely familiar with the DVP-642's specific requirements. The wide search value doesn't make any difference, I don't think. Not sure about the number of B-Frames, although I don't think you should go over 2. You should use a Home Theater Profile, though.

One thing I might suggest is to use AutoGK, setting it up to use the ESS chipset option. Then you can be sure the AVI output will be completely compatible with the DVP-642:

ESS-based standalones. MPEG4 players with ESS chipsets don't work with the matrices that AutoGK uses with the XviD Codec (you don't have to understand what a matrix is, just follow the instructions). Please choose this option if you have such player. Perhaps the problem will be fixed with a firmware upgrade, and perhaps not. But you won't be losing anything, or getting a movie inferior in any way by choosing that option. If you have such a standalone player, and your XviD videos play with corruption and/or smearing, try this setting. For both XviD and DivX codecs this option also enables the Home Theatre profile which is a part of DivX certification for hardware devices and which enables control over VBV buffer. Most standalones have issues with high bitrate spikes that cause the internal memory of the player to be full and not to accept more data for a short period of time. This causes pauses, skips and and shuttering. Both DivX and XviD support intelligent control of output buffer overflows so that this problem can be eliminated. Make sure to turn this option on if you experiencing such symptoms during playback on your standalone. Note that its not the only possible reason for having pauses and skips - users reported that by burning movies onto DVDRs instead of CDRs playback can be dramatically improved.

Under Software Requirements And Installation:

http://www.autogk.me.uk/modules.php?name=TutorialEN

But since it doesn't accept blu-ray or MKV as input, I guess you won't be able to use it.

[Randm]

do i even need b-frames though, i'm told it just helps with compression & seeking/rewind/fastforward or something like that...i heard that disabling b-frames is better for compatibility?

UPDATE: nvm, found the info i think i will choose to disable B-Frames...

https://www.videohelp.com/forum/archive/b-frames-and-trellis-quantization-t323862.html

"Yes, _generally speaking_ it's better to use B-Frames and Trellis Quantization to achieve the best possible quality when encoding to XVID/DIVX. But keep in mind that some hardware DIVX player are unable to read B-Frames."

[Randm]

k i switched to gordian knot so compressibility test is no longer a problem...anyways, so if i use the same settings for size based rips (not cq mode) in filesize/resolution/etc. with a dvd as a source & then a bd as a source...the final DRF average on each will be generally the same? despite the image looking better on the bdrip of course...i'd test it out myself, but i don't have both a BD & DVD of the same movie hehe...

also, its said that standard deviation should be less than 1, but why? what exactly is it...

[manono]

I'd say that if all the encoding settings are the same, the Blu-Ray encode will have a (slightly) higher average quant (higher DRF or lower quality) because the source is sharper. Just a guess, though, as I've never encoded from a Blu-Ray movie either.

As for the Standard Deviation, did you lose the other thread you made asking about it:

https://forum.videohelp.com/topic371657.html

If you ever lose a post or thread you made and want to find it again, click on any thread and at the top hit "Advanced Search". Fill in your own nickname in the "Search For Author" box. Up will come all threads in which you participated.

[Randm]

sorry, i didn't mean to cross post, i had little sleep that day or somn hehe...thanx for the info, think i learned nearly everything i need to know...i might have 1 or 2 last settings to figure out, if i need help, i'll post here instaed of a new topic