Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Hacked together a replacement of the ellitic filters with the tabuli #73

Draft
wants to merge 35 commits into
base: main
Choose a base branch
from
Draft

Hacked together a replacement of the ellitic filters with the tabuli #73

wants to merge 35 commits into from

Conversation

zond
Copy link
Collaborator

@zond zond commented May 22, 2024

rotator filters.

zond
zond commented Jun 14, 2024
View reviewed changes
cpp/zimt/masking.cc Outdated
@@ -64,7 +64,7 @@ void HwyComputeEnergy(const hwy::AlignedNDArray<float, 2>& sample_channels,
samples =
Load(d, sample_channels[{sample_index + downscale_index}].data() +
channel_index);
accumulator = MulAdd(samples, samples, accumulator);
accumulator = Add(samples, accumulator);
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why did you remove the square here?

zond
zond commented Jun 14, 2024
View reviewed changes
go/data/study.go
@@ -203,14 +203,14 @@ func (c CorrelationTable) String() string {
for _, scores := range c {
row := Row{string(scores[0].ScoreTypeA)}
for _, score := range scores {
row = append(row, fmt.Sprintf("%.2f", score.Score))
row = append(row, fmt.Sprintf("%.15f", score.Score))
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If you only use -leaderboard this won't have any effect.

@jyrkialakuijala
less memory traffic, 2.5x speedup for tabuli
5518df7 
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.101162755033022 |0.563709113431054 |0.773872237318597 |0.691086927997167 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

real	3m37.704s
user	184m0.443s
sys	42m52.675s
@jyrkialakuijala
Merge pull request #104 from jyrkialakuijala/tabuli
4ff27be 
less memory traffic, 3x speedup for tabuli
@jyrkialakuijala
less memory access by streaming the log10s for dbs
a700eeb 
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.101162746942526 |0.563709113431054 |0.773872237318597 |0.691086968125479 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

real	3m39.800s
user	186m13.002s
sys	42m43.169s
@jyrkialakuijala
Merge pull request #106 from jyrkialakuijala/tabuli
1cebe3c 
less memory access by streaming the log10s for dbs
@jyrkialakuijala
tiny fix
a6af47d
@jyrkialakuijala
Merge pull request #107 from jyrkialakuijala/tabuli
a3d6cb6 
tiny fix
@jyrkialakuijala
more rotators (128->150) improves MSE by a few %
9794994 
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.098637219950560 |0.567187736693103 |0.777010861372363 |0.694493861132456 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

real	4m19.323s
user	222m49.575s
sys	43m16.317s
@jyrkialakuijala
Merge pull request #108 from jyrkialakuijala/tabuli
4411d99 
more rotators (128->150) improves MSE by a few %
@jyrkialakuijala
more optimization of heuristics, ~1 % improvement
86b05dd 
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.097175453861064 |0.581050442393818 |0.776654974298645 |0.696445877155239 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |
@jyrkialakuijala
Merge pull request #109 from jyrkialakuijala/tabuli
7d3e6b0 
more optimization of heuristics, ~1 % improvement
@jyrkialakuijala
~8 % improvement
50178f3 
not quite sure why this improved so much

changes are too complicated to trace it exactly

much of the improvement was related how I refactored the log10
calculations to be simpler and just log based

|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.090239608814795 |0.534520229950808 |0.772224479921073 |0.707988288758192 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

real	4m18.248s
user	221m18.093s
sys	46m19.418s
@jyrkialakuijala
Merge pull request #114 from jyrkialakuijala/tabuli
37c51d7 
~8 % improvement
@jyrkialakuijala
simple loudness in fourier_bank
90d422a 
this incrementalizes/simplifies the tabuli path

|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.089327353427112 |0.559850956043716 |0.780744060923064 |0.708638004179837 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

real	4m19.928s
user	223m47.057s
sys	46m23.955s
@jyrkialakuijala
Merge pull request #115 from jyrkialakuijala/tabuli
91bb0f4 
simple loudness in fourier_bank
@jyrkialakuijala
temporal masking
6542578 
previous tabuli zimtohrli MSE score is 14 % higher than this

after this PR VisQOL will have 47 % higher MSE score than Zimtohrli

|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.078112668596219 |0.607368005163725 |0.828363954766286 |0.727804879676665 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

real	4m15.939s
user	220m7.740s
sys	46m34.613s
@jyrkialakuijala
Merge pull request #116 from jyrkialakuijala/tabuli
766ee9f 
temporal masking
@jyrkialakuijala
switching loudness and masking
ca6b7dc 
loudness happens first after this commit

this will help to reduce computation slightly later

|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.078443138275192 |0.602595657239603 |0.824102489042876 |0.727318570435013 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

it is a 0.5 % degradation for now
@jyrkialakuijala
Merge pull request #117 from jyrkialakuijala/tabuli
f471890 
switching loudness and masking
@jyrkialakuijala
~0.5 % improvement
780adbe 
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.077647743787295 |0.608127552926780 |0.829014091535645 |0.728679787505096 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |
@jyrkialakuijala
Merge pull request #119 from jyrkialakuijala/tabuli
b9fbd50 
~0.5 % improvement
@jyrkialakuijala
7-8 % improvement
d67c2cc 
further solidifies Zimtohrli as the best correlating metric

|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.072593165859890 |0.619257037510440 |0.814230081078870 |0.737751119792009 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |
@jyrkialakuijala
Merge pull request #120 from jyrkialakuijala/tabuli
a3dbc16 
7-8 % improvement
@jyrkialakuijala
quality fix
a933744 
after something likely went wrong in previous comment

git clumsiness

|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.072654586073989 |0.619257037510440 |0.814288552764735 |0.737622587952872 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |
@jyrkialakuijala
Merge pull request #121 from jyrkialakuijala/tabuli
59cc535 
quality fix
@jyrkialakuijala
~5% worse, but faster and simpler
6000340 
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.076254461155707 |0.589172424723762 |0.806090520228513 |0.731427610098147 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |

reducing the gammatone filters from 150 to 128

reducing NSIM size from 32x16 to 16x8 -- 32 bands would be more than 2
octaves of blurring!?!

revamping masking into something simpler, but also not axis separating
temporal and frequency masking effects
@jyrkialakuijala
Merge pull request #122 from jyrkialakuijala/tabuli
b112037 
~5% worse, but faster and simpler
@jyrkialakuijala
further simplifications
577f2fb 
taking correct frequency buckets from ISO loudness

reducing structural similarity support area

'''
Score type |MSE               |Min score         |Max score         |Mean score        |
-----------|------------------|------------------|------------------|------------------|
Zimtohrli  |0.074958337059633 |0.609170013051075 |0.804501676659602 |0.732603836275400 |
ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |
'''
@jyrkialakuijala
Merge pull request #123 from jyrkialakuijala/tabuli
fdcef96 
further simplifications
@jyrkialakuijala
minor improvement
bf7b003 
adjusting the nsim stabilization parameters

'''
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.073622255247300 |0.611416768760542 |0.797709696340402 |0.734967593303992 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |
'''
@jyrkialakuijala
Merge pull request #124 from jyrkialakuijala/tabuli
09e3309 
minor improvement
@jyrkialakuijala
simplifications and more elegance
604f19c 
|Score type |MSE               |Min score         |Max score         |Mean score        |
|-----------|------------------|------------------|------------------|------------------|
|Zimtohrli  |0.072503005846665 |0.613305088893694 |0.801783318140190 |0.737440109899061 |
|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |
|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |
|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |
|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |
|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |
|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |
|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |
|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |
|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |
|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |
@jyrkialakuijala
Merge pull request #125 from jyrkialakuijala/tabuli
d6493c5 
simplifications and more elegance
@jyrkialakuijala
several quality improvements
915309f 
tuning the existing structures

'|Score type |MSE               |Min score         |Max score         |Mean score        |'
'|-----------|------------------|------------------|------------------|------------------|'
'|Zimtohrli  |0.069100905174817 |0.628756659185876 |0.804728638916679 |0.744043538017276 |'
'|ViSQOL     |0.115330916105424 |0.520833375452983 |0.801480831107469 |0.675101633981268 |'
'|2f         |0.129541391104905 |0.484687555319526 |0.797475783883375 |0.661870345773127 |'
'|PESQ       |0.147425552045669 |0.342342966279351 |0.841271127756762 |0.647128996775172 |'
'|CDPAM      |0.153471222942756 |0.441558428344727 |0.728779141125759 |0.620699318941738 |'
'|PARLAQ     |0.185057687192323 |0.445261140223642 |0.784370761057963 |0.587162756572532 |'
'|AQUA       |0.223207996944378 |0.331645933512413 |0.739286336419790 |0.547804951221731 |'
'|PEAQB      |0.225217321572038 |0.278744167467764 |0.851011116004117 |0.553935720513487 |'
'|DPAM       |0.315810440183130 |0.186717781679534 |0.690564701717118 |0.460415212267967 |'
'|WARP-Q     |0.339686211572685 |0.067600137543649 |0.777119464646524 |0.475793617709890 |'
'|GVPMOS     |0.412937133868407 |0.006851162794410 |0.783946603687895 |0.412912222208318 |'

real	3m31.783s
user	177m9.966s
sys	40m50.441s
@jyrkialakuijala
Merge pull request #127 from jyrkialakuijala/tabuli
cde4ede 
several quality improvements
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@zond @jyrkialakuijala