HDV@Work: Much More about 720p24 HDV, Part 1
Mar 13, 2006 1:54 PM
By Steve Mullen
Though we all know there are two forms of HDV, this fact is far too often ignored when folks talk about “HDV production.” Thus, when one reads that a post house has spent hours painting out “HDV motion blocking artifacts,” the comment is of little value to the production community unless the type of HDV is specified. That is because there are significant differences between the HDV types. 1080i HDV is a long GOP system while 720p HDV is a short GOP system.
This difference is not like the difference between DV and DVCAM. Rather, it is much closer to the difference between DVCPRO 25 and DVCPRO 50. The short six-frame GOP used by 720p prevents significant motion blocking artifacts.
Nevertheless, when working with 720p24 HDV that will be transferred to film, you need the assurance that no artifacts will appear on the big screen. Can you have the confidence you need in light of claims that HDV will exhibit artifacts because it is “more compressed” than DVCPRO HD? As we will learn in this article, you can indeed have a great deal of confidence.
In order to maximize the amount of HD video that can be stored on a P2 card, these camcorders can be run in “native” mode when shooting 24fps. (Only the 24 frames for each second are transferred to storage.) Each “720PN” frame is compressed by the same amount it would be compressed were the camera shooting 60 frames per second. (Remember that all DVC intraframe codecs employs a fixed compression ratio.) Therefore, 720p24 is recorded at 40Mbps — which includes 3Mbps for four 48kHz PCM audio channels. Thus, the DVCPRO HD 720p24 effective video data rate is only 36.9Mbps.
The Tables below show the Total Bits-per-Second (bits per frame multiplied by 24 frames per second) for both HDV 720p24 and DVCPRO HD 720p24. Also shown in each Table are the two video data rates.
|
HDV 720p 24fps |
||||
|
4:2:0 |
H |
V |
Bits |
Bits-per-Frame |
|
Y |
1280 |
720 |
8 |
7,372,800 |
|
R - Y |
640 |
360 |
8 |
1,843,200 |
|
B - Y |
640 |
360 |
8 |
1,843,200 |
|
|
11,059,200 |
|||
|
Total Bits/Second |
265,420,800 |
|||
|
Video Data Rate |
18.7Mbps |
|||
|
Bits/Rate Ratio |
14:1 |
|||
The calculated Bits/Rate ratio is 14:1.
|
DVCPRO HD 720p 24fps |
||||
|
4:2:2 |
H |
V |
Bits |
Bits-per-Frame |
|
Y |
960 |
720 |
8 |
5,529,600 |
|
R - Y |
480 |
720 |
8 |
2,764,800 |
|
B - Y |
480 |
720 |
8 |
2,764,800 |
|
|
11,059,200 |
|||
|
Total Bits/Second |
265,420,800 |
|||
|
Video Data Rate |
36.9Mbps |
|||
|
Bits/Rate Ratio |
7:1 |
|||
The calculated Bits/Rate ratio is 7:1. The twice-as-large 14:1 ratio does seem to indicate that HDV is twice as compressed as DVCPRO HD. However, these ratios do not take into account the role of compression efficiency.
MPEG-2 is considered 3 to 5 times as efficient as DVC-based compression (DV, DVCAM, HDCAM, DVCPRO, DVCPRO 50, and DVCPRO HD). We see the practical benefits of this efficiency when we burn 25Mbps DV/DVCAM/DVCPRO25 to a DVD at 5Mbps to 8Mbps.
The Table below displays the Bits/Rate ratio of DVCPRO HD and HDV. These ratios are divided by three Compression Efficiency values to yield three Effective Ratios.
|
Bits/Rate Ratio |
Compressor Efficiency |
Effective Ratio |
Data Rate Multipliers |
Equivalent Data Rates |
|
|
DVCPRO HD |
7:1 |
|
7:1 |
|
36.9Mbps |
|
HDV |
14:1 |
3X |
4.7:1 |
1.5X |
28.1Mbps |
|
HDV |
14:1 |
4X |
3.5:1 |
2.0X |
37.4Mbps |
|
HDV |
14:1 |
5X |
2.8:1 |
2.5X |
46.8Mbps |
The DVCPRO HD 7:1 Bits/Rate ratio is then divided by the three HDV Effective Ratios, yielding three Data Rate Multipliers. The product of 18.7 and the three multipliers generates three Equivalent Data Rates.
Worst case, HDV has a 25 percent lower average data rate — and thus a higher compression ratio than DVCPRO HD. Best case, 720p24 HDV has a 25 percent greater average data rate that allows a lower compression than DVCPRO HD. Make note of the 2.5X Data Rate Multiplier.
How do intraframe (DVCPRO HD) and interframe (HDV) motion-handling capabilities affect these calculations? There are two answers to this question, one general and one specific. First, the natural variability of interframe compression is why I have provided a range of Compression Efficiency values. Second, in the specific case of 720p24 HDV, there is a very good reason why the best-case situation is representative of typical performance.
When the same level of compression is applied to 720p30 as is applied to ATSC 720p60, the result is a data rate of only 7.8Mbps. HDV 720p camcorders, however, use a data rate of 18.7Mbps. The 2X greater data rate explains why motion blocking artifacts are not a significant issue with 720p HDV. Therefore, in our comparative analysis, there is no real need to consider motion degradation.
In the next newsletter, we will confirm the best-case calculation and examine the real difference between 4:2:0 color-sampled HDV and 4:2:2 sampled DVCPRO HD.
Continue the discussion on “Crosstalk” the Millimeter Forum.


Multimedia
Blogs
Forum
Affordable HD
Whitepapers
Advertisers
Blogcast
Millimeter






