**5. Experimental results**

The proposed transcoder has been evaluated by using four representative QCIF sequences with different motion levels were considered. These sequences were coded at 15 fps and 30 fps using 150 frames and 300 respectively. In the DVC to H.264/AVC transcoder applied, the DVC stage was generated by the VISNET II codec using PD with BP = 3 as quantification in a trade-off between RD performance and complexity constraints but with whatever BP could be used. In addition, sequences were encoded in DVC with GOPs of length 2, 4 and 8 to evaluate different patterns. The parallel decoder was implemented by using an Intel C++ compiler (version 11.1) which combines a high-performance compiler as well as Intel Performance Libraries to provide support for creating multi-threaded applications. In addition, it provides support for OpenMP 3.0 (OpenMP, 2011). In order to test the performance of parallel decoding, it was executed over an Intel i7-940 multicore processor (Intel, 2011), although the proposal is not dependent on particular hardware. For the experiments, the parallel decoding was split into 9 parts where each core has thus a ninth part of the frame. This value is a good selection for QCIF frames (176x144), 16x16 macroblocks (this is the size of the block in the SI generation and thus a QCIF frame has 99 16x16 blocks) and 4 processors (4 cores, 8 simultaneous processes with hyper-threading).

For more complex patterns, which include mixed P and B frames (main profile), this method can be extended in a similar way with some changes. Figure 8 shows the transcoding from a DVC GOP of length 4 to a H.264 pattern IBBP. MVs are also stored by always following the same procedure. However, in this case the way to apply them in H.264/AVC changes.

For P frames, MVs are multiplied by a factor of 1.5 because MVs were calculated for a distance of 2 and P frames have their references with a distance of 3. For B frames, it depends on the position that they are allocated and it changes for backward and forward

As can be observed, this procedure can be applied to both K and WZ frames. Therefore, following this method the proposed transcoder can be used for transcoding from every DVC

The proposed transcoder has been evaluated by using four representative QCIF sequences with different motion levels were considered. These sequences were coded at 15 fps and 30 fps using 150 frames and 300 respectively. In the DVC to H.264/AVC transcoder applied, the DVC stage was generated by the VISNET II codec using PD with BP = 3 as quantification in a trade-off between RD performance and complexity constraints but with whatever BP could be used. In addition, sequences were encoded in DVC with GOPs of length 2, 4 and 8 to evaluate different patterns. The parallel decoder was implemented by using an Intel C++ compiler (version 11.1) which combines a high-performance compiler as well as Intel Performance Libraries to provide support for creating multi-threaded applications. In addition, it provides support for OpenMP 3.0 (OpenMP, 2011). In order to test the performance of parallel decoding, it was executed over an Intel i7-940 multicore processor (Intel, 2011), although the proposal is not dependent on particular hardware. For the experiments, the parallel decoding was split into 9 parts where each core has thus a ninth part of the frame. This value is a good selection for QCIF frames (176x144), 16x16 macroblocks (this is the size of the block in the SI generation and thus a QCIF frame has 99 16x16 blocks) and 4 processors (4 cores, 8 simultaneous processes with hyper-threading).

Fig. 8. Mapping from DVC GOP of length 4 to H.264 GOP IBBP.

searches.

GOP to every H.264/AVC GOP.

**5. Experimental results** 

During the decoding process, the MVs generated by the SI generation stage were sent to the H.264/AVC encoder; hence it does not involve any increase in complexity. In the second stage, the transcoder performs a mapping from every DVC GOP to every H.264/AVC GOP using QP = 28, 32, 36 and 40. In our experiments we have chosen different H.264/AVC patterns in order to analyze the behavior for the baseline profile (IPPP GOP) and the main profile (IBBP pattern). These patterns were transcoded by the reference and the proposed transcoder. The H.264/AVC reference software used in the simulations was theJM reference software (version 17.1). As mentioned in the introduction, the framework described is focused on communications between mobile devices; therefore, a low complexity configuration must be employed. For this reason, we have used the default configuration for the H.264/AVC main and baseline profile, only turning off the RD Optimization. The reference transcoder is composed of the whole DVC decoder followed by the whole H.264/AVC encoder. In order to analyze the performance of the proposed transcoder in detail we have taken into account the two halves and global results are also presented.

Furthermore, the performance of the proposed DVC parallel decoding is shown in Tables 1 (for 15 and 30fps sequences). PSNR and bitrate (BR) display the quality and bitrate measured by the reference WZ decoding. To calculate the PSNR difference, the PSNR of each sequence was estimated before transcoding starts and after transcoding finishes. Then the PSNR of the proposed transcoding was subtracted from the reference one for each H.264/AVC RD point, as defined by Equation 3. However, Table 1 do not include results for ΔPSNR because the quality obtained by DVC parallel decoding is the same as the reference decoding, it iterates until a given threshold is reached (Brites et al., 2008).

$$
\Delta PSNR(db) = PSNR\_{reference} - PSNR\_{proposed} \tag{3}
$$

Equation 4 was applied in order to calculate the Bitrate increment (ΔBR) between reference and proposed DVC decoders as a percentage. Then a positive increment means a higher bitrate is generated by the proposed transcoder. As the results of Table 1 show, when DVC decodes smaller and less complex parts, sometimes the turbo decoder (as part of the DVC decoder) converges faster with less iterations and it implies less parity bits requested and thus a bitrate reduction. However, generally speaking the turbo codec yields a better performance for longer inputs. For this reason, the bitrate is not always positive or negative. Comparing different GOP lengths, in short GOPs most of the bitrate is generated by the K frames. When the GOP length increases, the number of K frames is reduced and then WZ frames contribute to reducing the global bitrate in low motion sequences (like Hall) or increasing it in high motion sequences (Foreman or Soccer). Generally, decoding smaller pieces of frame (in parallel) works better for high motion sequences, where the bitrate is similar or even lower in some cases.

$$
\Delta BR(\%) = 100 \times \frac{\left(BR\_{proposad} - BR\_{reference}\right)}{BR\_{reference}} \tag{4}
$$

Concerning the time reduction (TR), it was estimated as a percentage by using Equation 5. In this case, negative time reduction means decoding time saved by the proposed DVC decoding. As is shown in Table 1, DVC decoding time is reduced by up to 70% on average. TR is similar for different GOP lengths, but it works better for more complex sequences.

Mobile Video Communications Based on Fast DVC to H.264 Transcoding 27

IPPP H.264 pattern

Sequence GOP Δܴܲܵܰ(db) Δܤܴ)% (ܴܶ) % (Δܴܲܵܰ(db) Δܤܴ)% (ܴܶ) % ( 2 -0.02 0.57 -41.57 -0.01 0.31 -42.54 Foreman 4 -0.03 0.86 -44.62 0.00 0.18 -41.81 8 -0.04 1.11 -45.85 -0.01 0.36 -43.71 2 -0.01 0.41 -30.04 0.00 0.05 -30.50 Hall 4 0.00 0.12 -30.77 0.00 0.06 -29.28 8 0.00 0.17 -27.21 0.00 0.01 -27.81

Coast Guard

Coast Guard

with 15fps and 30fps sequences.

with 15fps and 30fps sequences.

15fps 30fps

2 -0.01 0.27 -47.46 -0.01 0.19 -46.39 4 -0.01 0.33 -46.15 0.00 0.08 -45.59 8 -0.01 0.20 -47.61 0.00 0.09 -45.23

 2 -0.01 0.19 -38.85 -0.01 0.15 -37.86 Soccer 4 -0.04 1.18 -43.35 -0.03 0.84 -40.61 8 -0.05 1.63 -44.98 -0.03 0.90 -42.35 mean -0.02 0.59 -40.70 -0.01 0.27 -39.47

Table 2. Performance of the proposed transcoder mapping method for IPPP H.264 pattern

IBBP H.264 pattern

Sequence GOP Δܴܲܵܰ(db) Δܤܴ)% (ܴܶ) % (Δܴܲܵܰ(db) Δܤܴ)% (ܴܶ) % ( 2 -0.06 1.51 -48.86 -0.08 2.31 -48.17 Foreman 4 -0.08 2.00 -51.11 -0.07 2.18 -49.56 8 -0.07 2.18 -52.19 0.00 0.00 -37.71 2 -0.01 0.21 -39.06 -0.01 0.24 -37.31 Hall 4 -0.01 0.55 -37.11 -0.01 0.13 -35.74 8 -0.01 0.48 -36.15 0.00 0.01 -52.41

15fps 30fps

2 -0.04 1.13 -49.44 -0.01 0.26 -51.14 4 -0.04 1.24 -49.90 -0.01 0.37 -50.62 8 -0.05 1.40 -51.07 -0.05 1.59 -45.19

 2 -0.02 0.43 -42.93 -0.08 2.63 -46.90 Soccer 4 -0.05 1.54 -46.03 -0.08 2.57 -48.52 8 -0.08 2.52 -48.71 -0.04 1.21 -45.93 mean -0.04 1.27 -46.05 -0.08 2.31 -48.17

Table 3. Performance of the proposed transcoder mapping method for IBBP H.264 pattern

(5)



Table 1. Performance of the proposed DVC parallel decoder for 15 and 30 fps sequences (first stage of the proposed transcoder).

Results for the second stage of the transcoder are shown in Tables 2 and 3. In this case, both H.264/AVC encoders (reference and proposed) start from the same DVC output sequence (as DVC parallel decoding obtains the same quality as the reference DVC decoding), which is quantified with four QP values. For these four QP values, ΔPSNR and ΔBRare calculated as specified in Bjøntegaard and Sullivan's common test rule (Sullivan et al., 2001). TR is given by Equation 5. In Table 2, DVC decoded sequences are mapped to an IPPP pattern. In this case RD loss is negligible and TR is around 40%. For 30 fps sequences, the accuracy of the proposed method works better and RD loss is even lower. In addition, Figure 9 displays each plot for each of the four 4 QP values simulated. As can be observed, all RD points are much closer. For the IBBP pattern (Table 3), the conclusions are similar. Comparing both patterns, the IBBP pattern generates a slightly higher RD loss, but H.264/AVC encoding is performed faster (up to 48%). This is because B frames have two reference frames, but dynamic ME search area reduction is carried out in both of them. Figure 10 displays plots for each of the four QP points when an IBBP pattern is performed. As can be observed, the RD drop penalty is negligible.

ܴܶሺΨሻ ൌ ͳͲͲ כ ൫்ೞି்൯

Proposed DVC parallel decoder

ܴܤΔ (%)

Reference DVC decoder

> BR (kbps)

(dB)

(first stage of the proposed transcoder).

RD drop penalty is negligible.

Sequence GOP PSNR

Coast Guard ்

Reference DVC decoder

> BR (kbps)

PSNR (dB)

15 fps sequences 30 fps sequences

ܴܶ (%)

2 30.25 295.58 0.66 -75.99 32.41 504.93 -2.37 -74.29

 8 28.96 571.31 -1.04 -76.98 30.87 804.73 -0.88 -73.23 2 32.81 222.24 -2.52 -75.14 36.4 439.74 1.33 -70.89 Hall 4 33.1 224.09 7.99 -70.47 36.34 412.8 1.27 -70.83 8 33.06 224.13 9.05 -69.34 36.03 384.31 6.49 -68.96

 2 29.56 377.15 -2.67 -74.17 30.52 532.94 1.2 -74.86 Soccer 4 29.05 593.66 -2.82 -75.81 30.21 855.23 -0.58 -75.41 8 28.34 735.48 -3.2 -73.94 29.53 1069.21 -1.19 -74.65 mean 0.82 -73.73 1.55 -72.76

Table 1. Performance of the proposed DVC parallel decoder for 15 and 30 fps sequences

Results for the second stage of the transcoder are shown in Tables 2 and 3. In this case, both H.264/AVC encoders (reference and proposed) start from the same DVC output sequence (as DVC parallel decoding obtains the same quality as the reference DVC decoding), which is quantified with four QP values. For these four QP values, ΔPSNR and ΔBRare calculated as specified in Bjøntegaard and Sullivan's common test rule (Sullivan et al., 2001). TR is given by Equation 5. In Table 2, DVC decoded sequences are mapped to an IPPP pattern. In this case RD loss is negligible and TR is around 40%. For 30 fps sequences, the accuracy of the proposed method works better and RD loss is even lower. In addition, Figure 9 displays each plot for each of the four 4 QP values simulated. As can be observed, all RD points are much closer. For the IBBP pattern (Table 3), the conclusions are similar. Comparing both patterns, the IBBP pattern generates a slightly higher RD loss, but H.264/AVC encoding is performed faster (up to 48%). This is because B frames have two reference frames, but dynamic ME search area reduction is carried out in both of them. Figure 10 displays plots for each of the four QP points when an IBBP pattern is performed. As can be observed, the

2 30.14 289.84 1.11 -72.8 33.84 592.64 4.69 -71.88 4 30.13 371.62 1.38 -74.46 33.32 608.2 5.88 -72.51 8 29.65 437.85 1.91 -74.73 32.24 661.11 4.72 -70.92

Foreman 4 29.73 450.59 -0.05 -70.88 31.95 648.69 -1.91 -74.7

(5)

ܴܤΔ (%)

Proposed DVC parallel decoder

> ܴܶ (%)


Table 2. Performance of the proposed transcoder mapping method for IPPP H.264 pattern with 15fps and 30fps sequences.


Table 3. Performance of the proposed transcoder mapping method for IBBP H.264 pattern with 15fps and 30fps sequences.

Mobile Video Communications Based on Fast DVC to H.264 Transcoding 29

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00 41.00

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00

PSNR

PSNR

0 5 10 15 20 25 30

0 5 10 15 20 25 30

IBBP pattern. Sequences QCIF (176x144) 30fps GOP = 8

CoastGuard

Hall

0 5 10 15 20 25 30

Bit rate [kbit/s]

Bit rate [kbit/s]

Reference Proposed

Reference Proposed

Reference Proposed

Foreman

Soccer

CoastGuard

Foreman

CoastGuard Foreman Soccer

Soccer

IBBP pattern. Sequences QCIF (176x144) 30fps GOP = 2

Hall

Bit rate [kbit/s]

IBBP pattern. Sequences QCIF (176x144) 30fps GOP = 4

Hall

(a) (b)

Reference Proposed

Reference Proposed

Reference Proposed

Foreman Soccer

CoastGuard

CoastGuard

Foreman Soccer

CoastGuard Foreman Soccer

0 2 4 6 8 10 12 14 16 18

IBBP pattern. Sequences QCIF (176x144) 15 fps GOP = 2

Hall

Bit rate [kbit/s]

IBBP pattern. Sequences QCIF (176x144) 15 fps GOP = 4

Hall

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00 41.00

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00

PSNR

PSNR

PSNR

(c) (d)

0 2 4 6 8 10 12 14 16 18

IBBP pattern. Sequences QCIF (176x144) 15 fps GOP = 8

Hall

Bit rate [kbit/s]

(f) (g)

0 2 4 6 8 10 12 14 16 18 20

Bit rate [kbit/s]

Fig. 10. PSNR/bitrate results transcoding to H.264 IBBP GOP from DVC GOP = 2, 4, and 8 in sequences with 15 and 30 fps. Reference symbols: ■Foreman ♦Hall ▲CoastGuard ●Soccer

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00 41.00

PSNR

Fig. 9. PSNR/bitrate results transcoding to H.264 IPPP GOP from DVC GOP = 2, 4, and 8 in sequences with 15 and 30 fps. Reference symbols: ■Foreman ♦Hall ▲CoastGuard ●Soccer

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00 41.00

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00

PSNR

PSNR

PSNR

0 510 15 20 25 30

0 510 15 20 25 30

0 510 15 20 25 30

Bit rate [kbit/s]

Bit rate [kbit/s]

IPPP pattern. Sequences QCIF (176x144) 30fpsGOP = 8

Hall

Reference Proposed

Reference Proposed

Reference Proposed

CoastGuard Foreman

CoastGuard Foreman

Soccer

Soccer

CoastGuard

Foreman Soccer

IPPP pattern. Sequences QCIF (176x144) 30fps GOP = 2

Hall

Bit rate [kbit/s]

IPPP pattern. Sequences QCIF (176x144) 30fpsGOP = 4

Hall

(a) (b)

Reference Proposed

Reference Proposed

Reference Proposed

CoastGuard

Soccer

Foreman

CoastGuard Foreman Soccer

CoastGuard Foreman Soccer

0 2 4 6 8 10 12 14 16 18

IPPP pattern. Sequences QCIF (176x144) 15 fps GOP = 2

Hall

Bit rate [kbit/s]

IPPP pattern. Sequences QCIF (176x144) 15 fps GOP = 4

Hall

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00 41.00

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00

25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00

PSNR

PSNR

PSNR

(c) (d)

0 2 4 6 8 10 12 14 16 18

IPPP pattern. Sequences QCIF (176x144) 15 fps GOP = 8

Hall

Bit rate [kbit/s]

(f) (g)

0 2 4 6 8 10 12 14 16 18 20

Bit rate [kbit/s]

Fig. 9. PSNR/bitrate results transcoding to H.264 IPPP GOP from DVC GOP = 2, 4, and 8 in sequences with 15 and 30 fps. Reference symbols: ■Foreman ♦Hall ▲CoastGuard ●Soccer

Fig. 10. PSNR/bitrate results transcoding to H.264 IBBP GOP from DVC GOP = 2, 4, and 8 in sequences with 15 and 30 fps. Reference symbols: ■Foreman ♦Hall ▲CoastGuard ●Soccer

Mobile Video Communications Based on Fast DVC to H.264 Transcoding 31

15 fps for IBBP H.264 pattern

 2 -0.03 0.67 -75.12 Foreman 4 -0.03 -0.03 -70.51 8 -0.03 -1.01 -76.63 2 0 -2.47 -73.99 Hall 4 0 7.86 -69.66 8 0 8.92 -68.63

 2 -0.04 -2.61 -73.45 Soccer 4 -0.04 -2.77 -75.39 8 -0.03 -3.15 -73.64 mean -0.02 0.81 -73.10

Table 6. Performance of the proposed transcoder for 15fps sequences and IBBP pattern.

30 fps for IBBP H.264 pattern

 2 -0.04 -2.28 -73.25 Foreman 4 -0.03 -1.86 -74.06 8 -0.02 -0.85 -72.80 2 0 1.31 -69.40 Hall 4 0 1.25 -69.82 8 0 6.40 -68.13

 2 -0.04 1.20 -74.01 Soccer 4 -0.05 -0.55 -74.92 8 -0.04 -1.16 -74.29 mean -0.02 1.54 -72.00

Table 7. Performance of the proposed transcoder for 30fps sequences and IBBP pattern.

ܴܲܵܰ തതതതതതത (dB) Δത

2 0 4.60 -70.94 4 -0.01 5.78 -71.88 8 0 4.66 -70.46

ܴܲܵܰ തതതതതതത (dB) Δത

2 -0.01 1.10 -72.00 4 -0.01 1.37 -73.92 8 -0.02 1.89 -74.31

ܤܴതതത (%) ܴܶതതതത (%)

ܤܴതതത (%) ܴܶതതതത (%)

Sequence GOP Δത

Sequence GOP Δത

Coast Guard

Coast Guard


Table 4. Performance of the proposed transcoder for 15fps sequences and IPPP pattern.


Table 5. Performance of the proposed transcoder for 30fps sequences and IPPP pattern.

15 fps for IPPP H.264 pattern

 2 -0.02 0.64 -75.50 Foreman 4 -0.02 -0.05 -70.66 8 -0.01 -1.02 -76.78 2 0 -2.47 -74.48 Hall 4 0 7.86 -70.01 8 0 8.91 -68.92

 2 -0.03 -2.62 -73.82 Soccer 4 -0.03 -2.78 -75.61 8 -0.03 -3.15 -73.79 mean -0.01 0.80 -73.39

Table 4. Performance of the proposed transcoder for 15fps sequences and IPPP pattern.

30 fps for IPPP H.264 pattern

 2 -0.01 -2.32 -73.72 Foreman 4 0 -1.88 -74.34 8 0 -0.87 -72.99 2 0 1.31 -70.03 Hall 4 0 1.25 -70.24 8 0 6.40 -68.48

 2 -0.02 1.16 -74.39 Soccer 4 -0.02 -0.57 -75.15 8 -0.02 -1.17 -74.45 mean -0.01 1.53 -72.33

Table 5. Performance of the proposed transcoder for 30fps sequences and IPPP pattern.

ܴܲܵܰ തതതതതതത (dB) Δത

2 0 4.59 -71.31 4 -0.01 5.78 -72.15 8 -0.01 4.65 -70.65

ܴܲܵܰ തതതതതതത (dB) Δത

2 0 1.07 -72.41 4 0 1.35 -74.18 8 0 1.88 -74.52

ܤܴതതത (%) ܴܶതതതത (%)

ܤܴതതത (%) ܴܶതതതത (%)

Sequence GOP Δത

Sequence GOP Δത

Coast Guard

Coast Guard


Table 6. Performance of the proposed transcoder for 15fps sequences and IBBP pattern.


Table 7. Performance of the proposed transcoder for 30fps sequences and IBBP pattern.

Mobile Video Communications Based on Fast DVC to H.264 Transcoding 33

Artigas, X., Ascenso, J., Dalai, M., Klomp, S., Kubasov, D. & Ouaret, M. (2007). The

Ascenso, J., Brites, C. & Pereira, F. (2005). Improving frame interpolation with spatial motion

Ascenso, J., Brites, C., Dufaux, F., Fernando, A., Ebrahimi, T., Pereira, F. & Tubaro, S. (2010).

Brites, C., Ascenso, J., Quintas Pedro, J. & Pereira, F. (2008). Evaluating a feedback channel

De Cock, J., Notebaert, S., Vermeirsch, K., Lambert, P. & Van de Walle, R. (2010). Dyadic

Fernández-Escribano, G., Kalva, H., Cuenca, P. & Orozco-Barbosa L. (2007). A first approach

Fernández-Escribano, G., Cuenca, P., Orozco-Barbosa, L., Garrido del Solo, A. J. & Kalva, H.

Girod, B., Aaron, A.M., Rane, S. & Rebollo-Monedero, D. (2005) Distributed Video Coding.

ISO/IEC International Standard 14496-10 (2003). Information Technology – Coding of

Martínez, J. L., Kalva, H., Fernández-Escribano, G., Fernando, W.A.C. & Cuenca, P. (2009).

Momcilovic, S., Wang, Y., Rane, S. & Vetro, A. (2010). Toward Realtime Side Information

Oh, R., Park, J. & Jeon, B. (2010). Fast implementation of Wyner-Ziv Video codec using

OpenMP. (2011) The OpenMP API specification for parallel programming. URL:

Peixoto E., Queiroz R. L. & Mukherjee, D. (2010). A Wyner-Ziv Video Transcoder, Circuits

Sheng, T., Zhu, X., Hua, G., Guo, H., Zhou, J. & Chen, C. (2010). Feedback-free rate-

machine learning. Multimedia Tools Appl. 35(2): 225-240.

Audio – Visual Objects – Part 10: Advanced Video Coding.

transcoders. Multimedia Tools Appl. 38(1): 1-25.

Intel Processor Core family. (2011). URL: http://www.intel.com/.

Processing (ICIP), Cairo, Egypt. pp. 2941-2944.

Signal Processing (MMSP), Saint-Malo, France.

Broadcasting (BMSB), Shangai, China.

http://openmp.org

Systems, 16(2): 127-137.

Proceedings of the IEEE 93, 71-83.

Symposium (PCS), pp. 1-4. Citeseer.

Signal Processing Conference (EUSIPCO).

Smolenice, Slovak Republic.

Communication 23, 269-297.

16(2): 139-149.

DISCOVER codec: architecture, techniques and evaluation. In: Picture Coding

smoothing for pixel domain distributed video coding. 5th EURASIP Conference on Speech and Image Processing, Multimedia Communications and Services,

The VISNET II DVC Codec: Architecture, Tools and Performance. In: European

based transform domain Wyner-Ziv video codec. Signal Processing: Image

spatial resolution reduction transcoding for H.264/AVC - Multimedia Systems,

to speeding-up the inter mode selection in MPEG-2/H.264 transcoders using

(2008). Simple intra prediction algorithms for heterogeneous MPEG-2/H.264 video

Wyner-Ziv to H.264 video transcoder, 16th IEEE International Conference on Image

Decoding on Multi-core Processors, IEEE International Workshop on Multimedia

GPGPU, IEEE International Symposium on Broadband Multimedia Systems and

and Systems for Video Technology, IEEE Transactions on , vol.20, no.2, pp.189-200.

allocation scheme for transform domain Wyner–Ziv video coding - Multimedia

Finally, to analyze the global transcoding improvement, Tables 4, 5, 6 and 7 summarize global transcoding performance. In this case, Bjøntegaard and Sullivan´s common test rule (Sullivan et al., 2001) was not used because it is a recommendation only for H.264/AVC. Then, to estimate the PSNR obtained by the transcoder, the original sequences were compared with the output sequences after transcoding. For each four QP points, the PSNR measured is displayed as an average (Δܴܲܵܰ തതതതതതതത)). To estimate the BR generated by the reference and the proposed transcoder, the BR generated by both stages (DVC decoding and H.264/AVC encoding) was added. Then equation (1) was applied and it was averaged for each four H.264/AVC QPs (Δܤܴതതതത). As the DVC decoding contributes with most of the bitrate, results are very similar to those in Tables 1. In order to evaluate the TR, total transcoding time was measured for the reference and proposed transcoder. Then Equation 5 was applied and a mean was calculated for each of the four H.264/AVC QPs (ܴܶതതതത). As DVC decoding takes up most of the transcoding time, improvements in this stage have a bigger influence on the overall transcoding time, and so the TR obtained is similar to that in Table 1, reducing the complexity of the transcoding process by up to 73% (on average).
