10. Page 36, in (3.2): need to describe the meaning of “E{}” following the equation.
The sentence preceding (3.2) was revised to introduce variable that is being measured.
11. Page 36, in (3.3) and (3/4), the subscription of “t-1” should be “t”.
The equations were revised.
12. Page 38, Section 3.3.2, “ by introducing refresh …, by forcing, …”: need to rephrase this sentence. Right now, it can be mis-interpreted as “…limited by both introducing … and forcing …”.
In page 38, the first sentence of Section 3.3.1 was revised to express that the refresh points are obtained by introducing intra-coded block. The revised sentence is as follows: “In order to reduce the effects of error propagation within a single frame, spatial and temporal dependencies can be limited by introducing extra refresh points into the bitstream, which can be accomplished by forcing intra-coded blocks”
13. Page 39, last paragraph, “the Lagrangian optimisation …”: reference is need for the Lagrangian optimisation, or a brief description is needed.
The Lagragian optimisation mentioned here is the one expressed by (3.1). Before (3.1) a reference which describes the optimisation was given. This sentence was revised to point out to the original equation.
14. Page 40, in (3.7), , what is “α”? how do you select “α”. Elaborate it.
The “a” parameter is describe in the text after the equation as the leaky factor. The rest of the paragraph is used to elaborate of the impact of the “a” in trade-off between coding efficiency and error resilience. In order to discuss how “a” can be obtained two sentences were added at the end of the paragraph: “For the determination of the α parameter, one can use network information such as packet loss probability p leading to α = (1−p). However, as can be seen in previous studies empirical approach using exhaustive can also be used to optimise the α value for each sequence [86].”
15. Page 41, in (3.8): brief description about what will happen if both ft and ft-1 are lost.
The following discussion was added to the end of the paragraph after (3.8): “Although the use of two reference frames for leaky prediction is an effective way to reduce the error propagation with a small penalty in coding efficiency, the use of interpolated frame obtained using (3.8) still leads to a severe error propagation in case of burst frame loss. In this case both frames would be affect by packet loss, which would reduce the effectiveness of this method.”
16. Page 41, line 8 from top, : symbol “n” should be “t”.
Symbols were corrected.
17. Page 43, for Figure 3.4: need to mention what the optimal GOP length is.
18. Page 44, about the “FEC”: need to mention the disadvantages of FEC very briefly.
The text was revised to include a brief mention to FEC disadvantages: “These are normally associated with Forward Error Correction (FEC) algorithms, which are typically applied at the channel level which is not aware of the video content, leading to a sub-optimal error protection [103]. Using UEP, a source-channel optimisation is applied, whereby multi-levels of FEC are used to achieve different levels of error protection, overcoming this problem.”
19. Page 45, revise the sentence “since it is allowed that the configuration to change on every frame”.
Sentence was revised to: “However, for these types the configuration is allowed to change for every frame, contrary of FMPO type 0.”
20. Page 51, in (3.10), need to explain the meaning of “n”, and what is the starting value of “i”?
The sentence preceding the equation was revised to include the definition of “n” to: “The pixel value (lp) is recovered using a weighted average of n neighbouring pixel, as follows:”. Moreover, the equation was revised to define the initial value “i”.
21. Page 56: revise the sentence “To find the more accurate MV from the MV set composed described above” so that it is complete sentence.
To clarify the underlying idea this sentence and the following ones were revised to: “To discard the wrongly extrapolated MVs and find the more accurate one from the MV set described above, further refined is applied based on the distance between each pair of MVs.
Section 3.1
P2: Add description of Section 3.2.
Section 3.2
P2: Replace “Redundant coding consists in using” to “Redundant coding uses”.
Section 3.3
P1: Replace “in” with “on”.
Section 3.3.1
P1: Introduce J variable in the text, revise the sentence preceding (3.2) and fix wrong subscript.
P2: Add missing reference to the results.
Eq. (3.6): Reduce font size to fill.
Section 3.3.2
P1: Revise the first sentence to: “In order to reduce the effects of error propagation within a single frame, spatial and temporal dependencies can be limited by introducing extra refresh points into the bitstream, which can be accomplished by forcing intra-coded blocks”.
P1: Replace “intra-coded block” with “intra-coded blocks”.
P3: Add description to Lagrangian optimisation.
Figure 3.2: Add more information to the figure’s caption.
Section 3.3.3
P1: Add discussion about leaky factor “a”.
P2: Add discussion regarding the effect of losing both frames in (3.8).
P2: Fix wrong symbols.
Section 3.3.4
P2: Remove wrong commas.
P2: Add discussion regarding the GOP length.
P4: Use “neither” and “nor” as suggested.
Section 3.4
P1: Add brief mention to FEC disadvantages.
Section 3.4.1
P2: Add missing “only” and revise wrong sentence.
P4: Remove wrong “the”.
P4: Add missing reference to the results.
Section 3.4.2
P1: Replace “level” with “levels”.
P2: Add references to previous work in the literature.
P2: Fix two typos.
P5: Revise confusing sentence to: “Also based on scalable coding, the work in [114] proposes a technique for finding the optimal FEC coding rates that leads to correctly decode the EL when errors affect the base layer.
Section 3.5.1
P1: Fix equation 3.10 definition.
Section 3.5.2
P1: Replace “in” with “of”.
P2: Replace “if” with “is”.
P3: Remove extra “with”.
P5: Replace “order” with “another”.
Section 3.6
Fix 4 typos and add use “computational complexity”.
jfmcarreira