Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > (HD) DVD, Blu-ray & (S)VCD > DVD & BD Rebuilder

Reply
 
Thread Tools Search this Thread Display Modes
Old 11th March 2020, 19:42   #29321  |  Link
Mike-uk
Registered User
 
Join Date: Jun 2018
Location: Dorset
Posts: 94
30xx series nvidia cards are rumoured to be coming out in a few months, mayby h265 has been improved again over the current turing cards
Mike-uk is offline   Reply With Quote
Old 11th March 2020, 22:40   #29322  |  Link
gonca
Registered User
 
Join Date: Jul 2012
Location: Somewhere over the rainbow
Posts: 976
Quote:
Originally Posted by jdobbs View Post
Alan Parsons Project, 1980
You realize that was 40 years ago?
Aging yourself a bit.

Disclosure
Used to have the LP (vinyl), now have the digital version
gonca is offline   Reply With Quote
Old 12th March 2020, 02:28   #29323  |  Link
MrVideo
Registered User
 
MrVideo's Avatar
 
Join Date: May 2007
Location: Wisconsin
Posts: 1,665
Quote:
Originally Posted by gonca View Post
YUsed to have the LP (vinyl),
I still have that LP and ALL of the singles/LPs that I've ever obtained.
MrVideo is offline   Reply With Quote
Old 12th March 2020, 13:51   #29324  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,489
Quote:
Originally Posted by MrVideo View Post
I still have that LP and ALL of the singles/LPs that I've ever obtained.
I do to... and digital versions for almost all of them (mostly FLAC that I ripped from CDs).
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net
jdobbs is offline   Reply With Quote
Old 12th March 2020, 13:58   #29325  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,489
Quote:
Originally Posted by gonca View Post
You realize that was 40 years ago?
Aging yourself a bit.

Disclosure
Used to have the LP (vinyl), now have the digital version
Scary. I was working in assembler language back then on a computer assisted radar system that had actual core memory (donut shaped magnets threaded with conductors) and an 18 bit address bus. Debugging consisted of pushing the STOP button and reading a group LEDs to find the current address and the contents of the accumulator, followed by scanning the 4 foot stack of code listings to find out how we got there.

"Old" hardly covers it.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 12th March 2020 at 14:02.
jdobbs is offline   Reply With Quote
Old 12th March 2020, 22:18   #29326  |  Link
gonca
Registered User
 
Join Date: Jul 2012
Location: Somewhere over the rainbow
Posts: 976
But it is still a good song, along with most of what Alan Parsons Project released
gonca is offline   Reply With Quote
Old 13th March 2020, 06:14   #29327  |  Link
Tyrell63
Registered User
 
Join Date: Feb 2020
Location: Singapore
Posts: 5
Quote:
Originally Posted by cartman0208 View Post
try using DirectShowSource as Frameserver
Direct Show is already set (Default setting)
Tyrell63 is offline   Reply With Quote
Old 13th March 2020, 16:53   #29328  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,489
Quote:
Originally Posted by gonca View Post
But it is still a good song, along with most of what Alan Parsons Project released
That it is. I think I may have every album Alan Parsons released.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net
jdobbs is offline   Reply With Quote
Old 14th March 2020, 07:10   #29329  |  Link
AmigaFuture
Registered User
 
AmigaFuture's Avatar
 
Join Date: Apr 2003
Location: Within the main Source.
Posts: 843
Quote:
Originally Posted by gonca View Post
You realize that was 40 years ago?
Aging yourself a bit.

Disclosure
Used to have the LP (vinyl), now have the digital version
Quote:
Originally Posted by MrVideo View Post
I still have that LP and ALL of the singles/LPs that I've ever obtained.
The only LPs I kept were of Rush. The only band progressive enough worth keeping..okay, okay, also AC/DC. But Rush..that's golden, _The Best_ when it comes to music. Neil and Geddy made sure those lyrics where unique and honest. Zero fluff.

I left vinyl when CD's started becoming really well made, when "Remasters" were being created. I don't miss the Pops and Hisses of, Hisses of, Hisses of, Hisses of, LPs..and I was also deeply into how to care for them...and much, much more than I'll post here. One thing I do not support is Apple Digital or any other kind of Digital that does not support Lossless. Without a CD you're missing soooo much. That's what I miss about LPs...Artwork, and correct grammar and punctuation with information contained inside.

I've obtained all my favorites in Digital as well. I've encoded all my (sooo many) CDs with FLAC or APE. I've made sure Artwork is part of it.

Progressive music is much better than 95% of the stuff that's played these days. Auto-Tune = You're unable to sing = Stop!
__________________
Life is not a journey to the grave; but rather to skid out broadside, thoroughly used, torn and warn and loudly proclaim; WOW; What a ride!!! Soon, I'm going to do it AGAiN in different skin!!
AmigaFuture is offline   Reply With Quote
Old 14th March 2020, 19:03   #29330  |  Link
MrVideo
Registered User
 
MrVideo's Avatar
 
Join Date: May 2007
Location: Wisconsin
Posts: 1,665
What sucks about CDs is the 44.1 kHz sample rate. At a minimum it should have been 48 kHz. You can find 48/96 kHz material out there, but they are few and far between.
MrVideo is offline   Reply With Quote
Old 14th March 2020, 20:09   #29331  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 3,624
Quote:
Originally Posted by MrVideo View Post
What sucks about CDs is the 44.1 kHz sample rate. At a minimum it should have been 48 kHz. You can find 48/96 kHz material out there, but they are few and far between.
44.1 kHz sampling rate sucks? Hmmm, 44.1 kHz sampling rate accommodates audio frequencies up to about 20....22 kHz. I have some doubts whether people beyond their 20ies can even hear 20 kHz. Lot of placebos in the audio industry .....

Last edited by Sharc; 14th March 2020 at 20:12.
Sharc is offline   Reply With Quote
Old 14th March 2020, 20:30   #29332  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,489
Quote:
Originally Posted by Sharc View Post
44.1 kHz sampling rate sucks? Hmmm, 44.1 kHz sampling rate accommodates audio frequencies up to about 20....22 kHz. I have some doubts whether people beyond their 20ies can even hear 20 kHz. Lot of placebos in the audio industry .....
Yep. Anytime your hear the term "HD Audio" you can bet you are being sold a bill of nonsense.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 14th March 2020 at 20:33.
jdobbs is offline   Reply With Quote
Old 15th March 2020, 02:19   #29333  |  Link
MrVideo
Registered User
 
MrVideo's Avatar
 
Join Date: May 2007
Location: Wisconsin
Posts: 1,665
The point is that the upper frequencies harmonics are lost.
MrVideo is offline   Reply With Quote
Old 15th March 2020, 14:00   #29334  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 3,624
Quote:
Originally Posted by MrVideo View Post
The point is that the upper frequencies harmonics are lost.
Harmonics contribute to an instrument's 'timbre', but only with those frequencies which are within a human's hearing range which is normally below about 20kHz for youngsters. (Your avatar may chime in at this point though ).
Now take the highest C of a piano which is about 4186 Hz. Its audible overtones would be 8372, 12558, 16400, 20930Hz which are still covered by the 44.1kHz sampling rate. An 48kHz sampling rate will not even include the next harmonic of 25116Hz for which the minimum sampling rate would have to be about 51kHz.
So I don't see a significant improvement between 44.1kHz sampling rate or 48kHz sampling rate. More important than extending the frequency range is the resolution (quantization) of the samples (true 16bit is fine) and how the samples are processed in the sequel.
Oh well, audio quality can be discussed to death by audiophiles .

Edit:
Now this makes we wonder why the sampling rate has been increased from 44.1 to 48kHz at all. Technology convenience? Marketing gimmick? Or perhaps a better or simpler design of the anti-aliasing filter with a smoother roll-of and a better transient response? Just guessing.

Last edited by Sharc; 15th March 2020 at 17:40.
Sharc is offline   Reply With Quote
Old 16th March 2020, 20:41   #29335  |  Link
MrVideo
Registered User
 
MrVideo's Avatar
 
Join Date: May 2007
Location: Wisconsin
Posts: 1,665
Quote:
Originally Posted by Sharc View Post
Its audible overtones would be 8372, 12558, 16400, 20930Hz which are still covered by the 44.1kHz sampling rate.
20930 Hz has better sampling at 48 kHz than it does as 44.1 kHz (more data points). I suspect that 44.1 kHz was used on a CD in order to be able to supply more music. That doesn't hold water when existing LPs are redone for CDs, considering that a 48 kHz rate would have let them fit just fine.

In any event, back to our normal programming.
MrVideo is offline   Reply With Quote
Old 17th March 2020, 14:44   #29336  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,489
I have a question in the hope that one of the X265 experts out there can help me understand something. It's regarding how CRF mode works.

In X264, if I select a CRF value, I can be confident that it actually results in "constant quality". In other words, if I choose "medium" as my encoding preset and set the CRF to 20, I will get approximately the same quality as if I do the same using the "ultrafast" preset. Of course there will be a sacrifice in file size in order to make up the decreased efficiency of the faster mode. Great. Makes perfect sense.

But when I run X265 it seems to be all over the place. For example, I ran a job last night in which I used the exact same command line (with CRF=20) but changed only the preset from "ultrafast" to "medium". The "medium" job resulted in an HEVC file that is 69% larger than the "ultrafast" encode??? Huh? That is the exact opposite of what common sense would dictate. I haven't tested it yet, but I'm certain the ultrafast encode won't look as good... but how is that "constant quality"? I'm confused.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 17th March 2020 at 16:24.
jdobbs is offline   Reply With Quote
Old 17th March 2020, 17:46   #29337  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 3,624
@jdobbs
You may find this discussion interesting:
https://forum.doom9.org/showthread.php?t=180438
At the end CRF means just constant rate factor and provides a 'constant quality' for a particular encode. It is not an absolute quality measure and hence even for the same CRF the quality differs from source to source and between different encoder settings.
But yes, the behavior for x264 was more logical than it seems to be for x265.

Last edited by Sharc; 17th March 2020 at 17:52.
Sharc is offline   Reply With Quote
Old 17th March 2020, 21:51   #29338  |  Link
jdobbs
Moderator
 
Join Date: Oct 2001
Posts: 20,489
Quote:
Originally Posted by Sharc View Post
@jdobbs
You may find this discussion interesting:
https://forum.doom9.org/showthread.php?t=180438
At the end CRF means just constant rate factor and provides a 'constant quality' for a particular encode. It is not an absolute quality measure and hence even for the same CRF the quality differs from source to source and between different encoder settings.
But yes, the behavior for x264 was more logical than it seems to be for x265.
You would think, though, that "constant quality" for the exact same source would have some root in logic. It very definitely doesn't. All I get from all this is that in X265 CRF=20 doesn't equal CRF=20. Now that is crazy.
__________________
Help with development of new apps: Donations.
Website: www.jdobbs.net

Last edited by jdobbs; 17th March 2020 at 22:02.
jdobbs is offline   Reply With Quote
Old 18th March 2020, 11:10   #29339  |  Link
Sharc
Registered User
 
Join Date: May 2006
Posts: 3,624
Quote:
Originally Posted by jdobbs View Post
You would think, though, that "constant quality" for the exact same source would have some root in logic. It very definitely doesn't. All I get from all this is that in X265 CRF=20 doesn't equal CRF=20. Now that is crazy.
Only when one links CRF with 'constant quality', or (mis)interprets it as a measure for absolute quality. I understand that it has been 'sold' like this to make it popular or easier to understand, but in fact it just means Constant Rate Factor for the encoding process, and nothing else. It aims at providing a 'constant quality' visual experience for that particular encode, but when one changes encoding parameters for the same CRF and same source the quality as well as the files size will change. Apparently this seems to be much more pronounced for x265 than for x264.
When x265 encoder preset is changed from fast to slow and the file size increases (unexpectedly) for the same CRF one would expect in return that the quality has become better, like less blurring, better noise and details retention in low contrast areas etc. But seen from the encoding process it is still the same CRF, but not the same quality.
Whether this could be aligned more logical in the quality/encoder settings/file size context for x265, I really don't know. Perhaps an expert could shed some light on this.

What remains at the end is that for a given set of encoding parameters (preset) and for a given same source the quality becomes worse (@lower file size) for a higher CRF and becomes better (@higher file size) for a lower CRF. So I'd assume that the 1-pass CRF size prediction algo in BD-RB will still be working.

Last edited by Sharc; 19th March 2020 at 09:41.
Sharc is offline   Reply With Quote
Old 28th March 2020, 13:09   #29340  |  Link
cartman0208
Registered User
 
Join Date: Jun 2010
Posts: 38
Quote:
Originally Posted by Sharc View Post
Out of curiosity I made a quality comparison between cartman's QP24 HW encode and a BD-RB x265 'Very Good (Very Fast)' CPU encode at same bitrate (about 7740 kbps).

- cartman's HEVC QP24 GPU encode: VMAF = 95.7191, SSIM = 0.997884, PSNR = 49.1751
- BD-RB x265 'Very Good (Very Fast)' CPU encode: VMAF = 96.3585, SSIM = 0.997893, PSNR = 49.4838 (CPU encoding speed was about 9 fps with my i5-8400 CPU)

No big difference in the quality metrics, both encodes are softening and denoising the picture somewhat.
Visually they look much the same IMO, by far good enough for casual viewing (a very subjective statement of course ).
So the GPU HEVC encode has about the same quality as the CPU x265 'Very Good / Very Fast' encode in this test.
How can I compare the results myself? Is there some kind of How-to?
I'd like to tweak my settings to see if I can get even more quality
cartman0208 is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 14:41.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, vBulletin Solutions Inc.