I am a novice at manipulating video files. I tried my first Ripbot 264 conversion to get a movie I own that was ripped to my hard drive and is contained in one large file from VC-1 to h264.
I used the default profile settings for [High 4.0]:
AVC 4.0
Buffer size 25000 Max bitrate 25000
The original m2ts file was 33GB; the MKV file generated was 5GB. I understand there are other tracks involved but the video stream is by far the largest. I am not interested in saving space. I want the video to be the same quality just converted from VC-1 to h264. Are these the three settings that need adjusted? Do they need to match the source file? If so, how can I tell what the values are for the source file? Or are there other settings involved?
Try StreamFab Downloader and download from Netflix, Amazon, Youtube! Or Try DVDFab and copy Blu-rays! or rip iTunes movies!
+ Reply to Thread
Results 1 to 23 of 23
Thread
-
-
I use MKV, two pass and set the output size to 8150MBs to fit on a DVD-9 disc. No complaints about quality.
-
Do you keep AVC 4.0? I read the wiki on it but am not sure if increasing this value will create a better picture on a 16x9 DLP. Where do you set output size? Do you leave the max bitrate and buffer size at 25000?
-
This is what I see when I select AVCHD. I don't see 'lock size'. I believe the 'save as' just changes the container type for the output file.
-
OK, I see that I have to change mode from CQ to 2-pass. I can now 'lock size' to 23GB. I have not changed AVC Level, Buffer size, or max bitrate. Should these be changed? Doesn't max bitrate affect quality? Should buffer size always be set to the same size as max bitrate? What exactly does AVC Level > 4.0 do for you if you are viewing on a 16X9 display?
-
Personally I use the default encoder settings. 4.0 only selects 1920x1080 resolution. HD audio will use 3-4GBs,but AC3 only 300-500Mbs. If video qaulity is most important use AC3 audio,640AC3 is standard for BlueRay audio.
-
Can anybody explain what AVC Level, Buffer size, and max bitrate should be set at if you do not care about file size but want the best quality. Closest you can get to the original 25+GB VC-1 video clip.
-
If you want a high level of video quality that essentially looks the same as the source, use CRF 17.5-19 (for HD encodes).
For additional compression efficiency (the same quality in a smaller size), use subpixel refinement 9, adaptive b-frames mode 2, b-frames 5, reference frames 4. You can do deblock -2:-2 or -3:-3 to retain more sharpness during playback.
The VBV buffer size and VBV max bitrate settings are only useful for playback on specific devices that have a limit on max bitrate/buffersize. They do not increase the quality of the video; they only ensure it is playable on a device with limits. The quality of the video is determined by the entirety of the settings, most importantly the CRF or average bitrate setting.
If you really don't care about filesize and only care about quality, just keep the original VC-1 clip. Encoding to a lossy codec like H.264 will always result in losses of some kind (visible or not). Encoding a 25GB VC-1 clip to a 23GB x264 clip is just a huge waste of time and bitrate. Let x264 decide how much bitrate to allocate to the clip by using CQ/CRF. -
Ok so I change the CQ/CRF to 18. If I don't care about size should I leave all the profile settings to the default for [High 4.0]? Or should I tweak those settings as you describe or some other way. My image above shows the defaults for this profile. Changing the AVC level is also not needed for retaining the best picture quality? Thanks.
-
The actual H.264 Level is determined by the settings. Using higher/more demanding settings results in a higher level. Setting the Level in that box only labels the file as being a certain Level; it doesn't actually modify the encoding settings to comply with that level. I would suggest selecting "autoguess" if it's an option there, to let x264 try to guess the level based on your settings.
Using my suggested settings will slow down your encoding significantly and improve its compression efficiency somewhat. Since you don't care about filesize, you can just stick with the defaults. I do recommend reducing deblocking to -3 and -3, which weakens the deblocking filter (i.e. results in a bit more original detail being retained at risk of very mild H.264 blocking).
If you find CRF 18 gives you too-large encodes after all, you can try CRF 19-20 and use the higher-quality settings I suggested.
If you really, really don't care about file sizes, then why not stick with the VC-1 original? -
The reason for my questions is that I stream movies to my PS3. The media server software will convert video on the fly to a format recognized by the PS3. If not recognized natively it converts to mpeg. VC-1 video requires too much processing power to convert on the fly in realtime. I have movies on Blu-Ray that I want to stream from my PC. I do not want to sacrifice any quality in doing so.
I have already tested 2 conversions at CRF 22 the first came in at 5 GB. The second one I locked the size to the largest setting and like you said I could hardly tell the difference. Now I understand these settings a little better. I am currently cooking a conversion with CRF 18 and deblocking -3,-3. I will check it out when done.
I guess the rule of thumb would be when you don't care about size use 18, -3, -3. When wanting to get the output file size down to say 8GB (dual layer DVD) then maybe CRF one setting lower than 22 and the compression efficiency settings that you stated previously or that AND locking the file size?
I am still hazy on bitrate. I understand that devices may not function with too high a bitrate but isn't there a point where the bitrate is so low that the video is affected? What if it is set to 1000,1000? -
You're using PS3 Media Server I presume? I don't have experience with PS3 H.264 streaming, but PS3s have level limitations which your encode has to adhere to. I suggest you google for these limits/requirements.
I have already tested 2 conversions at CRF 22 the first came in at 5 GB. The second one I locked the size to the largest setting and like you said I could hardly tell the difference. Now I understand these settings a little better. I am currently cooking a conversion with CRF 18 and deblocking -3,-3. I will check it out when done.
I guess the rule of thumb would be when you don't care about size use 18, -3, -3. When wanting to get the output file size down to say 8GB (dual layer DVD) then maybe CRF one setting lower than 22 and the compression efficiency settings that you stated previously or that AND locking the file size?
Those deblock settings (-3:-3) don't affect compression at all; they're merely a flag that tells the playback device/decoder how much to deblock the encode during playback.
I am still hazy on bitrate. I understand that devices may not function with too high a bitrate but isn't there a point where the bitrate is so low that the video is affected? What if it is set to 1000,1000?
Yes, setting these peak bitrate settings too low will hurt your video quality. You should set certain bitrate limits since you're encoding for the PS3, a device. (Google for the specific limits.) IMO, a peak bitrate of 12,000kbps is more than enough for HD, even for 1080p material. Only the most complex and noisy scenes will get to that bitrate range, assuming you're using a sane CRF in the range of 18-22.
(Do note that limiting peak AND average bitrates may also be important if your streaming connection isn't fast enough. My own powerline connection at home only supports bitrates of up to 12mbps, including audio.) -
I can sustain connections over 100mbps (gigabit switch & nic and cat-6 cables).
The raw de-muxed VC-1 stream is 26,760,792 KB.
The raw h264 converted stream is 11,116,513 KB.
I have been told that h264 can give you the same picture quality as mpeg at half the file size (bitrate). How do VC-1 algorithms compare to h64? The converted stream is less than half the size. I was expecting larger. Is h264 that much more efficient than VC-1? Or are there other settings that can be tweaked to produce a better image? -
You're not thinking about this correctly. Being surprised that a VC-1 stream was shrunk to half its size when compressed by H.264 makes no sense at all. Comparing two encoders requires a comparison of the same source being compressed by the two encoders respectively. The half-MPEG2-bitrate claim means that a source stream appears the same quality whether compressed to MPEG-2 at x bitrate or H.264 at x/2 bitrate.
x264 is a very efficient encoder, and I believe it's slightly-to-somewhat more efficient than VC-1 (although I don't have evidence on-hand). A 11GB 1080p x264 encode is quite normal. What matters is that you're using a generous quality level (CRF 18 ) so you generally won't need to worry about encodes being "too low bitrate", because x264 takes care of that for you. -
Hello All
As many on this forum, I'm new to the world of ripping HD content and would appreciate the thoughts and views of those who have been working in this arena for some time. I'll try to provide as much information as I can about what I'm trying to accomplish but if I've left anything out, please let me know.
Goal: Take existing media collection (combination of DVDs, HD-DVD and Blu-Ray) and rip them to my HTPC (specs in the my sig)
Concerns: While not the most discerning audiophile, I do want good quality sound should I choose to output the content through a high-end receiver or directly to the TV/Soundbar solution. I would also like to ensure the image/sound quality is as high as possible while reducing the storage requirements typically required of HD source rips (particularly blu-ray). In short, I'd like to keep the rips around 6-10GB in size.
Setup: I plan to use XBMC in favor of Windows Media Center to manage my content. I've found that it handles MKV files very nicely.
Tests: I've done some preliminary testing with the latest version of RipBot and found the following (source Harry Potter - Half Blood Prince Blu-Ray rip using AnyDVD HD):
- Using the 4.0 profile, CQ/22, MKV and AC3 640kbps: Source File: 26GB M2TS, Converted File: 3.5GB MKV
- Using a custom 4.0 profile, CQ/18, MKV, AC3 640kbps, Auto Crop, de-block: -2/-2: Converted File: 6.5GB MKV
From my eye, it almost appeared as if the first test yielded better visual quality. I also noticed that with Auto Crop enabled, everything seemed shifted to the left. Sound appeared fine and didn't lag/skip but the image in the second pass seemed a bit more grainy.
Any thoughts on optimizing my rips for playback from my HTPC? I don't plan to stream them.
I also tried the MP4 route but it failed with a note saying that it couldn't open the executable. Also, I have the latest Shark 007 codecs installed as it seemed to satisfy most of the codec requirements for RipBot. AviSynth is the latest version.
Thanks in advance for all your help/suggestions.
Joe -
With RipBot, I get the best results using MKV two pass with a target size of 8150MBs. That allows me to backup the conversion to DL DVD discs in case of hard drive failure on my storage servers. If you don't need DVD backup, then a target size of about 10GB might be the best size/quality. Either way, you should be able to fit 50 - 60 BD backups on a 500GB HDD.
I don't have any problems with playback using VLC, TMT3, Zoom player, or MPC-HC. I do have a plugin for TMT3 that works for Media Center, but I find MC too restrictive and never use it for MKV or BD disc playback.
And welcome to our forums. -
-
If you don't need DVD backup, then a target size of about 10GB might be the best size/quality.
-
I agree. But some of us have limited hard drive storage space and it helps to know the finished size of a file so we know how many videos we can backup to a particular drive. I'm quite happy with the quality of my 8GB backups, but I suspect a slightly larger size might help the quality on some encodes.
In my case, I don't trust that my backup hard drives won't fail at some point, so I do an additional backup to DVD DL media and that determines my filesize. I guess I should try a couple of CQ backups as they would likely be quite a bit faster. And since the OP's original question was about optimum quality, CQ may be a better choice in his case.
Otherwise, I do use the default RipBot settings, as I see no reason to modify them, except for MKV two pass and filesize. -
Sorry to dredge this thread up but I've been reading it with keen interest, thanks so far.
From everything I understand about riptbot, I'm more suited to using CQ17-18 as I have masses of storage and size isn't an issue. I'm purely like the other guys - all about as little visual quality loss as possible. It got me thinking, I recoded one last night at CQ18- the average bitrate of the source was 30,000kbps and a meaty 35gb in size, the post reocde mkv was about 6gb with average video bitrate of about 5,800kbps. Seemed a massive drop to me, could that be right?? Is it just down to x264 being more efficient than the original VC-1? I always though that minimal standards for a 720p file should be around 5,000kbps - but for a 1080p file? I dunno I guess I thought it should be higher really. Same with a lot of 1080p torrents floating around, bitrates from 8,000kbps through to about 15,000kbps - all massively less than their sources upon reference checking.
Then I read this:
What matters is that you're using a generous quality level (CRF 18 ) so you generally won't need to worry about encodes being "too low bitrate", because x264 takes care of that for you.
From what I understand it wouldn't be any use to change the VBV buffer / max values to influence the quality as this is only used for the device the file is played from, correct? As when I first started out I was changing this to 35,000kbps.
Generally 20-22 is recommended as a good range for HD, and 18-20 for SD
Cheers. -
t got me thinking, I recoded one last night at CQ18- the average bitrate of the source was 30,000kbps and a meaty 35gb in size, the post reocde mkv was about 6gb with average video bitrate of about 5,800kbps. Seemed a massive drop to me, could that be right?? Is it just down to x264 being more efficient than the original VC-1?
Same with a lot of 1080p torrents floating around, bitrates from 8,000kbps through to about 15,000kbps - all massively less than their sources upon reference checking.
Does this mean to say the 5,800kbps average bitrate of my recode is no big deal and that the visual quality loss will be miniscule to un-noticeable?
From what I understand it wouldn't be any use to change the VBV buffer / max values to influence the quality as this is only used for the device the file is played from, correct? As when I first started out I was changing this to 35,000kbps.
For clarity, do you mean if recoding an SD source use CQ18, and for HD you could get away with 20-22?
Only I initially read that as for best quality HD recode use 20-22, which didn't seem right to me.
Similar Threads
-
'Test card' video for optimum settings?
By CompVid in forum Software PlayingReplies: 6Last Post: 22nd Dec 2011, 13:04 -
Capturing Video 8 with optimum quality
By bwt in forum CapturingReplies: 7Last Post: 1st Sep 2011, 21:52 -
TS to MKV using RipBot video quality issue
By dracusorr in forum Video ConversionReplies: 5Last Post: 14th Sep 2010, 06:28 -
TS to MKV using RipBot video quality issue
By dracusorr in forum Newbie / General discussionsReplies: 1Last Post: 11th Sep 2010, 23:34 -
HELP! RipBot Settings for Highest quality HD-DVD to Blu-Ray Project
By clarkkent57 in forum Blu-ray RippingReplies: 6Last Post: 9th Sep 2010, 12:58