PDA

View Full Version : What's the current state of x265?


Chengbin
17th January 2016, 22:48
I've been out of doom9 for a while. Last time I remembered x265 was still struggling to compete with x264 and it had a lot of trouble retaining fine details. What's the current state of x265? I hope it is better than x264 now. If x264's file size is 100%, what would x265's file size roughly be for equivalent quality? What about the speed of the encoder?

Thanks

movmasty
18th January 2016, 01:45
Can do less than half of the size at decent quality but washed, i guess that at 70% would be ok for now
x265 faster gives 45-50% of the speed of x264 medium-slow-

BTW, if High Efficiency here means the same than in AAC,
x265 cannot ever reach the quality of 264, just be more good at lower bitrates,
as to get this goal the frames get a kind of pre-processing, blurring in video case.

birdie
18th January 2016, 05:34
@Chengbin

Not yet there. Come back in a year.

huhn
18th January 2016, 08:49
BTW, if High Efficiency here means the same than in AAC,
x265 cannot ever reach the quality of 264, just be more good at lower bitrates,
as to get this goal the frames get a kind of pre-processing, blurring in video case.

both mean high efficiency but that's it.

i have no clue why you would compare them like this...

LoRd_MuldeR
18th January 2016, 20:16
Can do less than half of the size at decent quality but washed, i guess that at 70% would be ok for now
x265 faster gives 45-50% of the speed of x264 medium-slow-

BTW, if High Efficiency here means the same than in AAC,
x265 cannot ever reach the quality of 264, just be more good at lower bitrates,
as to get this goal the frames get a kind of pre-processing, blurring in video case.

both mean high efficiency but that's it.

i have no clue why you would compare them like this...

HE-AAC is simply the "plain old" AAC format combined with SBR (Spectral Band Replication). The idea of SBR is to reduce the sampling rate to 1/2 before the actual encoding, which - according to Nyqust's Theorem - will reduce the highest frequency that can be retained to 1/2. It's effectively a strong low-pass filter. Furthermore, an SBR-enabeld decoder will do some "magic" to re-create the upper half of the frequency spectrum. The SBR technology is not limited to AAC at all. For example, SBR combined with MP3 is called "mp3PRO". Clearly, you want HE-AAC only at the ultra-low bitrates, where "plain" AAC fails to retain acceptable quality. At medium to high bitrates, you don't want to use HE-AAC, for obvious reasons.

HEVC/H.265 is nothing like that! It's not just "AVC/H.264 with some additional pre-processing". It is a completely new, enhanced compression standard. An there is no reason why HEVC/H.265 shouldn't perform better (or at least equivalent) than AVC/H.264 at any bitrate. But it's clear that HEVC/H.265 will shine most at the very low bitrates, where AVC/H.264 starts looking bad. At the higher bitrates, where AVC/H.264 already looks great, there's just not much room for improvement ;)

Sagittaire
19th January 2016, 19:30
HE-AAC is simply the "plain old" AAC format combined with SBR (Spectral Band Replication). The idea of SBR is to reduce the sampling rate to 1/2 before the actual encoding, which - according to Nyqust's Theorem - will reduce the highest frequency that can be retained to 1/2. It's effectively a strong low-pass filter. Furthermore, an SBR-enabeld decoder will do some "magic" to re-create the upper half of the frequency spectrum. The SBR technology is not limited to AAC at all. For example, SBR combined with MP3 is called "mp3PRO". Clearly, you want HE-AAC only at the ultra-low bitrates, where "plain" AAC fails to retain acceptable quality. At medium to high bitrates, you don't want to use HE-AAC, for obvious reasons.


Well I think that with same source HE-AAC at 96 KHz, 128 Kbps should be better than AAC at 48 KHz 128 Kbps.

And certainely that HE-AAC 5.1 192 Khz, 24 bit at 448 Kbps will be always better than AC3 5.1 48 Khz, 16 bits at 448 Kbps ... and by far.

And certainely that HE-AAC 5.1 192 Khz, 24 bit at 6144 Kbps will be always better than PCM 5.1 48 Khz, 16 bits at 6144 Kbps ... ;-)

Sagittaire
19th January 2016, 19:42
Can do less than half of the size at decent quality but washed, i guess that at 70% would be ok for now
x265 faster gives 45-50% of the speed of x264 medium-slow-

BTW, if High Efficiency here means the same than in AAC,
x265 cannot ever reach the quality of 264, just be more good at lower bitrates,
as to get this goal the frames get a kind of pre-processing, blurring in video case.

This debate looks exactly like when x264 appeared face to XviD ... and like XviD vs x264 your conclusion is completely false.

XviD can't fight with x264 simply because for the same bitrate you can use 720x400 for x264 for same source and only 640x360 for XviD if you want acceptable quality.

It's the same thing with HEVC: you can encode with upsampling for the same quality/pixel: with same bitrate, encoding at 720p or/and 10 bits in HEVC will be always better (and by far) than the same encoding with AVC at 480p or/and 8 bits in AVC.

In conclusion, use more powerfull codec change always the coding technical: actually I make test encoding with Star Wars III, in 1080p, with HEVC at 1500 Kbps, and never x264 can obtain the same quality or same detail retention whatever the resolution.

sneaker_ger
19th January 2016, 19:58
What took a long time wasn't just x264 beating XviD, it was beating XviD consistently. x265 is not yet beating x264 consistently, but depending on the usage case it might already be a lot better. The main consensus at the moment seems to be: low-bitrate (like your 1500 kbps 1080p example) or clean cartoons are x265's strengths. x264 still has the upper hand at higher bitrates with grainy/non-blurry content, especially if you factor in encoding speed. Your mileage may vary.

Sagittaire
19th January 2016, 20:10
What took a long time wasn't just x264 beating XviD, it was beating XviD consistently. x265 is not yet beating x264 consistently, but depending on the usage case it might already be a lot better. The main consensus at the moment seems to be: low-bitrate (like your 1500 kbps 1080p example) or clean cartoons are x265's strengths. x264 still has the upper hand at higher bitrates with grainy/non-blurry content, especially if you factor in encoding speed. Your mileage may vary.

Well bitrate problem is false problem for me: Encoding in 4K, 10 bits at 10 Mbps with HEVC will be always better than the same encoding in 2K, 8 Bits at 10 Mbps with AVC ...

Asmodian
20th January 2016, 01:15
Well bitrate problem is false problem for me: Encoding in 4K, 10 bits at 10 Mbps with HEVC will be always better than the same encoding in 2K, 8 Bits at 10 Mbps with AVC ...

Why 10-bit HEVC but only 8-bit AVC?

At 4K it is pretty easy to pick HEVC. With a 1080p source, high bitrates, and 10-bit, x264 still often, but not always, beats x265. If a x264 encode looks great (~transparent to the source) it often looks better (more detailed) than a x265 encode at the same bitrate, at least in my experience.

Upscaling a 1080p source to 4K and encoding that at 10 Mbps with x265 does not look better than encoding the 1080p source at 10 Mbps with x264. Actually, I haven't done this exact test but that would be a very inefficient workflow. :p

foxyshadis
20th January 2016, 01:52
Why 10-bit HEVC but only 8-bit AVC?

Hardware support. 10-bit HEVC will be playable on almost all hardware decoders (except for the first-gen 8-bit only ones), including Bluray UHD.

Upscaling a 1080p source to 4K and encoding that at 10 Mbps with x265 does not look better than encoding the 1080p source at 10 Mbps with x264. Actually, I haven't done this exact test but that would be a very inefficient workflow. :p

It actually does, because x264 loses significant efficiency due to its tiny blocks at that size. There's no fine detail for it to shine on, so flat blocks are everywhere.

Asmodian
20th January 2016, 03:19
Hardware support. 10-bit HEVC will be playable on almost all hardware decoders (except for the first-gen 8-bit only ones), including Bluray UHD.

will be ;)

If you can decode 10-bit HEVC in hardware then you can probably decode 10-bit AVC in software too.

It actually does, because x264 loses significant efficiency due to its tiny blocks at that size. There's no fine detail for it to shine on, so flat blocks are everywhere.

I was talking about x264 encoding a 1080p source and x265 encoding the same source upscaled to 4K. With a 4K source x265 is the preferred encoder in my opinion, as pretty much everyone seems to agree.

The situation where x264 still often beats x265 is 1080p or below when using bit rates high enough x264 is transparent. For other use cases I wouldn't argue that x264 is better than x265.

It is not that x264 is ever more efficient than x265 it is that x265 cannot get to the same quality as x264 when unrestricted.

Right now I can see the comparison to HE-AAC v.s. AAC (assuming the same source and sample rates :p ) but unlike HE-AAC this is not because of the way it works and it will probably not be true forever.

foxyshadis
20th January 2016, 04:42
will be ;)

If you can decode 10-bit HEVC in hardware then you can probably decode 10-bit AVC in software too.

Not on a TV, Bluray player, etc. That's what I mean by hardware support.

Chengbin
20th January 2016, 20:28
What is the reason/limitation that x265 is not as good as x264 at high bitrates? Can this be fixed?

Asmodian
20th January 2016, 22:45
Yes it can be fixed. As to why, that is a bit completed but it probably has something to do with the larger transforms and matrix decimation.

foxyshadis
21st January 2016, 02:30
The situation has become significantly better since one of the 64x64 prediction modes was disabled for the upcoming 1.9, that seemed to be causing a lot of the trouble, plus there's been some psy-rd updates. Some people just disable 32x32 and up entirely, which barely impacts efficiency at lower resolutions and definitely looks a little sharper. There's still more that can be done with mode decision, I'm sure, so keep watching.

movmasty
21st January 2016, 11:35
HE-AAC is simply the "plain old" AAC format combined with SBR (Spectral Band Replication). The idea of SBR is to reduce the sampling rate to 1/2 before the actual encoding, which - according to Nyqust's Theorem - will reduce the highest frequency that can be retained to 1/2. It's effectively a strong low-pass filter. Furthermore, an SBR-enabeld decoder will do some "magic" to re-create the upper half of the frequency spectrum. The SBR technology is not limited to AAC at all. For example, SBR combined with MP3 is called "mp3PRO". Clearly, you want HE-AAC only at the ultra-low bitrates, where "plain" AAC fails to retain acceptable quality. At medium to high bitrates, you don't want to use HE-AAC, for obvious reasons.

HEVC/H.265 is nothing like that! It's not just "AVC/H.264 with some additional pre-processing". It is a completely new, enhanced compression standard. An there is no reason why HEVC/H.265 shouldn't perform better (or at least equivalent) than AVC/H.264 at any bitrate. But it's clear that HEVC/H.265 will shine most at the very low bitrates, where AVC/H.264 starts looking bad. At the higher bitrates, where AVC/H.264 already looks great, there's just not much room for improvement ;)
OK, Then means not the same, Thanks.
And The "magic" with SBR is that it doesnt completely destroy the higher sampling info, keeps some that the SBR-enabeld decoder can use.
Good Idea for video enc. too.

However, h265, actually, has a blur problem,
and it looks so similar at the HEaac audio preprocessing :D

movmasty
21st January 2016, 11:42
Well I think that with same source HE-AAC at 96 KHz, 128 Kbps should be better than AAC at 48 KHz 128 Kbps.

And certainely that HE-AAC 5.1 192 Khz, 24 bit at 448 Kbps will be always better than AC3 5.1 48 Khz, 16 bits at 448 Kbps ... and by far.

And certainely that HE-AAC 5.1 192 Khz, 24 bit at 6144 Kbps will be always better than PCM 5.1 48 Khz, 16 bits at 6144 Kbps ... ;-)Magic! if they halves...we double it!

So h265 4k, at xKbps should look better than h264 2k at xKbps?

...except for the coding speed maybe :o

movmasty
21st January 2016, 11:57
This debate looks exactly like when x264 appeared face to XviD ... and like XviD vs x264 your conclusion is completely false.As i remember the debate was only on the initial x264 low speed, not on quality/bitrate

XviD can't fight with x264 simply because for the same bitrate you can use 720x400 for x264 for same source and only 640x360 for XviD if you want acceptable quality.At the same bitrate of xvid 640x360, x264 can do easily 576p.

It's the same thing with HEVC: you can encode with upsampling for the same quality/pixel: with same bitrate, encoding at 720p or/and 10 bits in HEVC will be always better (and by far) than the same encoding with AVC at 480p or/and 8 bits in AVC.

In conclusion, use more powerfull codec change always the coding technical: actually I make test encoding with Star Wars III, in 1080p, with HEVC at 1500 Kbps, and never x264 can obtain the same quality or same detail retention whatever the resolution.In my few tests I've noticed a blur problem for x265,
How did you get rid of this to obtain "the same quality/pixel" ?

I remember that also wmv9 used the blur technique to lower its bitrate with acceptable quality.

movmasty
21st January 2016, 12:09
The situation has become significantly better since one of the 64x64 prediction modes was disabled for the upcoming 1.9, that seemed to be causing a lot of the trouble, plus there's been some psy-rd updates. Some people just disable 32x32 and up entirely, which barely impacts efficiency at lower resolutions and definitely looks a little sharper. There's still more that can be done with mode decision, I'm sure, so keep watching.
I'm disabling 64x64, glad to know that i can do with 32x32 barely impacting efficiency at lower resolutions,

Believed that bigger bloks were the core of x265. :goodpost:

Sagittaire
21st January 2016, 19:51
Magic! if they halves...we double it!

So h265 4k, at xKbps should look better than h264 2k at xKbps?

...except for the coding speed maybe :o

Not Magic ... :devil:

Professional audio recording can be in 192 KHz and 24 Bits. I claim that HE-AAC encoding at 192 KHz, 24 bits, 5.1 at 6 Mbps with good Psy will be better than same downsample lossless PCM 48 Khz, 16 bits, 5.1 at 6 Mbps.

Same thing with Movie Source. All masters are at least in 4K. In my test HEVC in 4K, 10 bits at 10 Mbps are always better than the downsizing AVC encoding in 2K, 8 Bits at 10 Mbps ... :devil:

Why 10-bit HEVC but only 8-bit AVC?


Simply because AVC hardware decoding never support 10 Bits, all the HEVC decoder will support 10 bits.


Upscaling a 1080p source to 4K and encoding that at 10 Mbps with x265 does not look better than encoding the 1080p source at 10 Mbps with x264. Actually, I haven't done this exact test but that would be a very inefficient workflow. :p

I never speak about "upscaling 1080p source at 4K". All movie master are, at least in 4K.

burfadel
21st January 2016, 22:33
Professional audio recording can be in 192 KHz and 24 Bits. I claim that HE-AAC encoding at 192 KHz, 24 bits, 5.1 at 6 Mbps with good Psy will be better than same downsample lossless PCM 48 Khz, 16 bits, 5.1 at 6 Mbps.

Don't you mean LC-AAC? HE-AAC and LC-AAC were stupidly named as it creates confusion, HE-AAC (High Efficiency AAC) sounds better than LC-AAC (Low Complexity), however HE-AAC is meant for low bandwidth situations and reduces sound quality significantly. LC-AAC is the one meant for high quality audio, music etc. Probably a bit of an oversimplification but is suitable for the purpose of stating that HE-AAC should be avoided unless a strictly low bitrate scenario, for example below maybe 96 Kbits/s.

Sagittaire
22nd January 2016, 21:05
Don't you mean LC-AAC? HE-AAC and LC-AAC were stupidly named as it creates confusion, HE-AAC (High Efficiency AAC) sounds better than LC-AAC (Low Complexity), however HE-AAC is meant for low bandwidth situations and reduces sound quality significantly. LC-AAC is the one meant for high quality audio, music etc. Probably a bit of an oversimplification but is suitable for the purpose of stating that HE-AAC should be avoided unless a strictly low bitrate scenario, for example below maybe 96 Kbits/s.

No HE-AAC is simply LC-AAC + SBR (spectral band replication). Make encoding in HE-AAC for 192 KHz is like use LC-AAC for 96 Khz with spectral recontruction at 192 KHz. 192 Khz, 24 bits encoding with SBR at high bitrate can produce extremely high quality enconding with good psychoaccoustic profil.

Asmodian
23rd January 2016, 01:11
If you are encoding 192 KHz audio wouldn't it be better to use LC-AAC at 192 KHz instead of HE-AAC? I don't understand why you would compare HE-AAC to LC-AAC at half the sample rate. As an aside: Is SBR for a >48 KHz source even useful?

There is the hardware decoder issue with 10-bit H.264 so I suppose there is a reason to compare 8-bit H.264 to 10-bit H.265. I never use hardware decoding so personally it is not useful to limit the quality of H.264 when doing these comparisons. However, right now 8-bit H.264 can still beat 10-bit H.265, assuming high enough bitrates. This is also assuming only using presets and tunings; I haven't tried tweaks like disabling 32x32 and larger transforms for very high bitrate encodes, that is a quite interesting option I will have to play with. :)

LoRd_MuldeR
23rd January 2016, 02:17
As an aside: Is SBR for a >48 KHz source even useful?

I would say no. If you want to compress audio with a sampling rate above 48 KHz, the first thing to do is downsampling to 48 KHz, using your favorite resampler(e.g. SoX). SBR is not needed for that ;)

Reasoning: According to Nyquists Theorem, the highest frequency that can be retained (without aliasing) is 1/2 the sampling rate. So, with a sampling frequency of 48 KHz, you can already retrain frequencies up to 24 KHz. And, considering that the human ear can hear frequencies in the range from ~15 Hz to ~20 KHz (for most adults the upper limit is actually even lower), 48 KHz is the highest sampling frequency you are ever going to need, realistically. With 96 KHz or even 192 KHz sampling you keep frequencies that no human can hear anyway. It means you can safely downsample to 48 KHz without loosing anything relevant. Again: No SBR needed for that! That's also the reason why new audio formats, such as Opus, are fixed to 48 KHz.

I think using SBR does make more sense at ultra-low bitrates, where you can't encode "full" 48 KHz at decent quality. In this case, using SBR is probably the lesser of two evils. The other one would be downsampling to 24 KHz in the traditional way. Note that SBR also is going to downsample an 48 KHz source to 24 KHz before the actual encoding. But, at least, SBR is going to re-upsample to 48 KHz at the decoder side - and then it uses some "magic" to make up the missing upper frequencies. At the same time, applying the SBR upsampling "magic" on a 96 KHz source, which has been encoded at 48 KHz (after SBR downsampling), is a rather pointless effort. The SBR decoder would be making up frequencies that no human being can hear...

(BTW: Driving your "analog to digital" converter at 96 or 192 KHz can make sense, because of the characteristics of the analog low-pass filter. It's called "oversampling". But, as soon as you are in the digital domain, you can downsample to 48 KHz)

See also:
http://xiph.org/video/vid1.shtml (cf. chapter about "Sampling Rate")

Sagittaire
23rd January 2016, 15:10
1) The professional audio formats use 192 kHz and 32-bit.

2) SBR is designed for high frequencies reconstruction. You can use it for 24/48 Khz or 96/192 KHz reconstruction ... there are no theoritical limit. HE-AAC at 48 Khz is in fact LC-AAC at 24 KHz with SBR reconstruction at 48 Khz: that mean HE-AAC is particulary good for low bitrate (half bitrate LC-AAC + SBR reconstruction). Anyway you can use HE-AAC at 96 Khz if you want SBR reconstruction with LC-AAC at 48 Khz.

Asmodian
23rd January 2016, 21:24
1) The professional audio formats use 192 kHz and 32-bit.

This is to allow editing and format conversions without impacting audible quality, not because it sounds better itself. No one uses >48 KHz for lossy compression.

2) SBR is designed for high frequencies reconstruction. You can use it for 24/48 Khz or 96/192 KHz reconstruction ... there are no theoritical limit. HE-AAC at 48 Khz is in fact LC-AAC at 24 KHz with SBR reconstruction at 48 Khz: that mean HE-AAC is particulary good for low bitrate (half bitrate LC-AAC + SBR reconstruction). Anyway you can use HE-AAC at 96 Khz if you want SBR reconstruction with LC-AAC at 48 Khz.

LC-AAC at 48 KHz would sound better than HE-AAC at 96 KHz because you don't want to waste bits on SBR for frequencies you cannot hear.

Anyway, this is a HEVC thread and I think the "HEVC is to AVC what HE-AAC is to LC-AAC" analogy has run its course.

Sagittaire
24th January 2016, 16:20
This is to allow editing and format conversions without impacting audible quality, not because it sounds better itself. No one uses >48 KHz for lossy compression.


Completely false. You can find realy easely HDaudio with 24 bits and 192 Khz in FLAC and not only for professionnal:
https://www.hdtracks.com/

All modern audio chip can use 192 Khz and 24 Bits (audio chip on PC Motherboard, smartphone, Blu Ray audio, SACD ... ect etc) and not for audio conversion.


LC-AAC at 48 KHz would sound better than HE-AAC at 96 KHz because you don't want to waste bits on SBR for frequencies you cannot hear.

Anyway, this is a HEVC thread and I think the "HEVC is to AVC what HE-AAC is to LC-AAC" analogy has run its course.

completely wrong ... another one. HE-AAC is simply LC-AAC with SBR mode. You can have LC-AAC at 48 KHz byte to byte identical to LC-AAC part in HE-AAC encoding at 96 KHz (you can desactive SBR in decoder if you want)


And for initial subject. Like for XviD vs x264, x265 is always able to encode at higher resolution than x264. At high bitrate if x264 produce high quality for 2K, HEVC will encode at 4K the same source with higher quality and more detail. x265 can encode 1080p movie source at 1.5 Mbps with really good quality and good detail preservation, x264 can't make that. Like AVC in the past, HEVC will simply change the encoding technical (Particulary for Blu Ray Ripp for exemple).

foxyshadis
24th January 2016, 18:48
This is still the HEVC forum, guys. There's a forum for audio if you want to discuss HD audio and its encoding.

Asmodian
25th January 2016, 02:47
And for initial subject. Like for XviD vs x264, x265 is always able to encode at higher resolution than x264. At high bitrate if x264 produce high quality for 2K, HEVC will encode at 4K the same source with higher quality and more detail. x265 can encode 1080p movie source at 1.5 Mbps with really good quality and good detail preservation, x264 can't make that. Like AVC in the past, HEVC will simply change the encoding technical (Particulary for Blu Ray Ripp for exemple).

My use case is re-encoding blurays to put them on my RAID array for watching. 3Mbps is fine with me, or even 6Mbps, I only don't want to store the original 20-40 Mbps.

Are you suggesting that upscaling 1080p to 4K and encoding with x265 offers better quality at the same bitrate compared to x264 encoding 1080p? I haven't actually tried this, I just assumed it couldn't possibly be even equivalent. :o

I have found that for 1080p at 3-6 Mbps I get better quality using x264 than I do using x265. x265 is very close, and looks great, but x264 is still better, with the minor side benefit that it is also faster to encode and easier to decode. However, some early tests of mine with --ctu 32 --max-tu-size 16 show very promising results so I will have to keep testing x265.

This does not mean x265 is worse for any use case, simply that there is room for improvement when encoding relatively high bitrate 1080p or lower resolution. Maybe it will be as simple as a way to bias towards smaller transforms as the quality ratio improves.

I feel like we are probably arguing different points. If I was to encode a 4K source I would definitely use x265.

Also, FLAC isn't lossy and SBR does take a few bits to encode. :p

Sagittaire
25th January 2016, 19:27
My use case is re-encoding blurays to put them on my RAID array for watching. 3Mbps is fine with me, or even 6Mbps, I only don't want to store the original 20-40 Mbps.

Well x265 can produce good quality in midle case at 1.5 Mbps for 1080p. x264 can't.

Are you suggesting that upscaling 1080p to 4K and encoding with x265 offers better quality at the same bitrate compared to x264 encoding 1080p? I haven't actually tried this, I just assumed it couldn't possibly be even equivalent. :o

I always speak about 4K source and never about upscaling 2K to 4K.

This does not mean x265 is worse for any use case, simply that there is room for improvement when encoding relatively high bitrate 1080p or lower resolution. Maybe it will be as simple as a way to bias towards smaller transforms as the quality ratio improves.

X265 is simply already unbeatable at low bit / pixel. If you find that x264 produce good result for 720p at 1.5 Mbps, x265 will able to make better job at 1.5 Mbps but for 1080p

Also, FLAC isn't lossy and SBR does take a few bits to encode. :p

And? you were wrong , right?
- HE-AAC is just LC-AAC+SBR (bit to bit identical for LC-AAC part if the encoder want) with higher sampling rate
- HDaudio is not just for professional, and you can make lossy + SBR at 32 bits and 192 Khz with extreme quality if you want

Asmodian
25th January 2016, 20:28
I don't have any 4K sources and I don't want to do 1.5 Mbps encodes... I agree with you and would use x265 if I did. :rolleyes:

blublub
14th February 2016, 19:42
Hi

Has anyone with very good eyesight already tested the new version 1.9 with grain setting and higher nitrates (equivalent x264 19-17)?

I am asking because it supposed not to use lagerger partition from what I have understood should result in better details.