Home

Results 1 - 10 of 269,773 for larger sized. Search took 8.108 seconds.  
Sort by date/Sort by relevance
The intent is to understand the impact of larger transform sizes on the compression gain. In order to evaluate the compression gain from larger coding block sizes, we set the largest transform size to be 64x64, 32x32, 16x16 and 8x8 respectively in the configuration file of TMuC. (...) But the encoder selects the transform size based on RD optimization results. This implies that a larger transform size does not improve the overall compression performance in the prediction unit where large transform blocks are not chosen. 3. (...) Given the cost and compression gain from larger transform block sizes, we suggest limiting the maximum transform size to be 32x32 or 16x16.
Language:English
Score: 1257549.6 - https://www.itu.int/wftp3/av-a..._10_C_Guangzhou/JCTVC-C117.doc
Data Source: un
Inherited deficit- versus debt- driven tax changes Long-run growth Inherited fiscal factors Difference (1) (2) (1) - (2) Initial tax rate 21 19 2** Size of tax change 1 1 0 Debt-driven Deficit-driven Difference (1) (2) (1) - (2) Initial tax rate 19 18.5 0.5 Size of tax change 2 1 1** Long-run changes have higher initial tax rates than inherited fiscal factors Debt-driven changes have larger size of tax changes than deficit-driven changes Non-linearities can explain differential tax multipliers  Moreover…  (i) non-linear arguments + (ii) differences in initial tax rate and/or size of tax change Explain the ex-ante “puzzling” differences in tax multipliers of different types of exogenous tax changes In Summary…  The impact of tax changes on output is highly non-linear. The effect is larger (in absolute terms) (i) the higher is the initial level of tax rate and (ii) the larger is the size of the tax rate change.  Estimating tax multipliers involve much more than “academic debates” and have critical practical implications In Summary…  Our estimates for the region show a quite heterogeneous response to tax shocks: In Summary…  The impact of tax changes on output is highly non-linear. The effect is larger (in absolute terms) (i) the higher is the initial level of tax rate and (ii) the larger is the size of the tax rate change.  Estimating tax multipliers involve much more than “academic debates” and have critical practical implications Paraguay (10%) Guatemala (12%) Costa Rica (13%) Example for Costa Rica: ∆VAT = 2pp → ∆GDP ≈ 0 Ecuador (14%) Colombia (16%) Peru (18%) Example for Colombia: ∆VAT = 2pp → ∆GDP = -1.7% ∆1pp → No ∆GDP ∆2pp → No ∆GDP ∆1pp → No ∆GDP ∆2pp → ∆GDP ∆1pp → ∆GDP ∆2pp → ∆GDP 5% 13% 14% 19% 20% 27% Argentina (21%) Uruguay (22%) Greece (24%) Example for Greece: ∆VAT = 1pp (June 2016, from 23% to 24%) → ∆GDP = -2% Policy implications in action I: On the size of the government and economic development  Relationship between GDP per capita and the size of the government spending with respect to GDP: Policy implications in action I: On the size of the government and economic development  Relationship between “excess” spending and VAT rate: - Countries like Guatemala may be able to “catch up” to a more typical (i.e., larger) provision of public goods by increasing the VAT rate.
Language:English
Score: 1189531 - https://www.cepal.org/sites/de...ument/files/riera-crichton.pdf
Data Source: un
They find that larger towns tend to have larger wage and smaller employment responses than smaller towns and argue that this is driven by heterogeneity in the labour supply elasticity across different city sizes. (...) In order to match the stylized fact that more productive firms are located in larger cities Gaubert (2018) assumes there is a complementarity between raw efficiency (z) and city size (Lc) such that ex-ante more productive firms increase their productivity by more by location in a larger city. (...) Empirically, these are located in larger cities (Davis and Dingel, 2015) and in the model they would locate in larger cities if 3A more technical discussion can be found in the online apendix. 19 the relative price of skilled labour decreases with city size which is in line with empirical evidence (Bernard et al., 2008).
Language:English
Score: 1155616.8 - https://www.wto.org/english/ne...jan_bakker_essay_12sep19_e.pdf
Data Source: un
In the TMuC software[3], large transform blocks up to 64x64 size has been adopted with large intra and MC block structures. (...) Average absolute value of coefficients for 4x4 transform block These results show that the distribution function of transform coefficients differs depending on the transform block size. One thing it is obvious is that the number of continuous high-frequency ±1 coefficients (Trailing Ones) in a large block in scanning order should be larger than that of small size transform block. (...) Especially, more gain can be observed for “Nebuta festival”, which has strong noise throughout the contents resulting higher number of large transform blocks selected. Table 3 Luma BD-rate Sequence Larger than 8x8 Larger than 16x16 Traffic -0.106 -0.061 PeopleOnStreet -0.152 -0.036 Kimono -0.435 -0.592 ParkScene -0.078 -0.076 Cactus -0.292 -0.089 BasketballDrive -0.496 -0.152 BQTerrace -0.057 -0.048 Steam locomotive train -0.263 -0.287 Nebuta festival -4.039 -3.639 Avarage -0.658 -0.553 5 Conclusion This contribution reported an approach for the entropy coding of transform coefficients in large block size.
Language:English
Score: 1148577.5 - https://www.itu.int/wftp3/av-a...010_07_B_Geneva/JCTVC-B070.doc
Data Source: un
As JVT employs 6-tap FIR filter for MC, the neighboring pixels (black ones in Fig 2.1) are needed for motion compensation. If the size of the MC block size is 16x16, the overhead is: 1.72265625 16 16 ) 5 16 ( ) 5 16 ( = ´ + ´ + and if the size is 4x4, the overhead is: 5.0625 4 4 ) 5 4 ( ) 5 4 ( = ´ + ´ + That is, the smaller the MC block size is, the larger the overhead is. (...) Our simulation results show that for P-pictures, employing 6-tap filter for larger MC block size than 8x8 and 4-tap filter for smaller MC block will reduce computational complexity while keeping the image quality degradation to a minimum. For B-pictures, employing 4-tap filter for larger MC block size than 8x8 and 2-tap filter for smaller MC block will cause a loss in coding efficiency by up to 2-3%, however such loss will not greatly have a negative impact on visual quality.
Language:English
Score: 1135831.6 - https://www.itu.int/wftp3/av-a...02_05_Fairfax/JVT-C052r1-L.doc
Data Source: un
As a CU becomes larger, it wants to exploit its bigger market power by raising its CET. (...) The equilibrium CU structure thus comprises two blocs, one larger and one smaller. Members of the larger bloc have higher welfare than members of the smaller bloc. (...) On the other hand, when the smaller bloc is constrained by Article XXIV, the larger bloc would want to accept fewer members because the welfare loss that the smaller bloc can inflict on the larger bloc is smaller.
Language:English
Score: 1116463.5 - https://www.wto.org/english/re...forum_e/wtr11_18july11_d_e.htm
Data Source: un
As a CU becomes larger, it wants to exploit its bigger market power by raising its CET. (...) The equilibrium CU structure thus comprises two blocs, one larger and one smaller. Members of the larger bloc have higher welfare than members of the smaller bloc. (...) On the other hand, when the smaller bloc is constrained by Article XXIV, the larger bloc would want to accept fewer members because the welfare loss that the smaller bloc can inflict on the larger bloc is smaller.
Language:English
Score: 1116463.5 - https://www.wto.org/spanish/re...forum_s/wtr11_18july11_d_s.htm
Data Source: un
As a CU becomes larger, it wants to exploit its bigger market power by raising its CET. (...) The equilibrium CU structure thus comprises two blocs, one larger and one smaller. Members of the larger bloc have higher welfare than members of the smaller bloc. (...) On the other hand, when the smaller bloc is constrained by Article XXIV, the larger bloc would want to accept fewer members because the welfare loss that the smaller bloc can inflict on the larger bloc is smaller.
Language:English
Score: 1116463.5 - https://www.wto.org/french/res...forum_f/wtr11_18july11_d_f.htm
Data Source: un
As JVT employs 6-tap FIR filter for MC, the neighboring pixels (black ones in Fig 2.1) are needed for motion compensation. If the size of the MC block size is 16x16, the overhead is: 1.72265625 16 16 ) 5 16 ( ) 5 16 ( = ´ + ´ + and if the size is 4x4, the overhead is: 5.0625 4 4 ) 5 4 ( ) 5 4 ( = ´ + ´ + That is, the smaller the MC block size is, the larger the overhead is. (...) Solution 1: FIR1 for larger MC blocks than 8x8; FIR2 for smaller MC blocks Solution 2: P-picture …FIR1 for larger MC blocks than 8x8; FIR2 for smaller MC blocks B-picture …FIR2 for all MC blocks Solution 3: P-picture …FIR1 for larger MC blocks than 8x8; FIR2 for smaller MC blocks B-picture …FIR2 for larger MC blocks than 8x8; FIR3 for smaller MC blocks FIR1 {1, -5, 20, 20, -5, 1} // 32 FIR2 {-1, 5, 5, -1} // 8 FIR3 {1, 1} //2 Table 3.1 MC Block Size FIR1 FIR2 FIR3 16x16 1.72265625 1.41015625 1.12890625 16x8 2.1328125 1.6328125 1.1953125 8x16 2.1328125 1.6328125 1.1953125 8x8 2.640625 1.890625 1.265625 8x4 3.65625 2.40625 1.40625 4x8 3.65625 2.40625 1.40625 4x4 5.0625 3.0625 1.5625 Table 3.2 4. (...) Our simulation results show that both for P- and B- pictures, employing 6-tap filter for larger MC block size than 8x8 and 4-tap filter for smaller MC block will reduce computational complexity while keeping the image quality degradation to a minimum.
Language:English
Score: 1114167.8 - https://www.itu.int/wftp3/av-a...2002_05_Fairfax/JVT-C052-L.doc
Data Source: un
In ABT the size of the transforms is connected to the size of the blocks used for motion compensation. (...) The purpose of this experiment is to · Verify the improved coder performance of ABT in TML-4 · Investigate the performance of integer transforms of all applicable block sizes · Investigate coding methods that exploit the run lengths in the larger blocks 2 Experiment Description Since the size of the block transforms is connected to the size of the blocks used for motion compensation the motion estimation stage has to be modified as well as the transform coding and the quantization scheme. 2.1 Motion Estimation In the TML, the SATD is used as the error criterion for motion estimation. (...) The core experiment shall define and investigate transforms for application in ABT: · Integer transforms for the blocks of size 4x8, 8x4, 8x8, 8x16, 16x8, and 16x16 · Integer transforms for the blocks of size 4x8, 8x4, and 8x8 if larger transforms are not applied. 2.3 Scanning and Quantization The TML defines a quantization table for the 4x4 transform coefficients as well as two scan modes.
Language:English
Score: 1112488.1 - https://www.itu.int/wftp3/av-a...video-site/0005_Osa/q15j68.doc
Data Source: un