Contexto Histórico y Político de la Educación Superior en Colombia

As mentioned in the introduction, the research presented here is the first stage of a larger study on the reception of respoken subtitles included within the EU-funded project DTV4ALL. Following Jensema (1998), participants in DTV4ALL have been divided into three groups: hearing, hard-of-hearing and deaf viewers. The two experiments included in this article have been conducted with the first group of participants: hearing viewers.

The purpose of this first experiment is two-fold: to find out how much visual and verbal information is obtained through respoken subtitles in news programmes and to gather the participants’ views on the speed of these subtitles. The material used for the experiment consisted of four clips from the Six O’Clock News broadcast on 4 July 2007 by BBC1. The participants chosen for the study were 30 hearing viewers, between the ages of 20 and 45, native or near native-speakers of English and very familiar with subtitling.

Comprehension and reading patterns of respoken subtitles for the news

181

Half of them were postgraduate students currently doing an MA on Audiovisual Translation at Roehampton University and the other half was formed by lecturers and professional subtitlers. All viewers were proficient readers and habitual subtitle users.

During the experiment, participants were shown two clips with two news items and were asked to answer questions about one. Following Jensema’s (1998) methodology, the clips had no sound and so viewers had to rely on the subtitles to retrieve the information. The clips were subtitled by respeaking at two different speeds: 180 wpm, which is, as explained above, the common speed for respoken news on the BBC, and 220 wpm, which allows for verbatim subtitling of most weather reports and interview programmes. The latter speed constitutes an increase of 40 wpm over the recommended speed for hearing viewers (180 wpm), exactly the same increase deaf and hard-of-hearing viewers are currently having to face (from the 140 wpm recommended by Jensema [1998] to the current 180 wpm).

In order to carry out a quantitative analysis of the amount of information retrieved by the viewers, the two news clips were notionally divided, drawing on Chafe’s (1980) concept of idea units,5 which has already been applied to respeaking (Eugeni, 2009; Romero-Fresco, 2009), into 14 semi-units: 8 verbal units and 6 visual units. In (very few) cases in which participants retrieved a semi-unit that was not included in these 14, the new unit was also factored in the analysis. For the purpose of the analysis of the findings, a simple division was made whereby any result between 0% and 25% is regarded as zero to poor information retrieval, 25%-50% goes from poor to sufficient, 50%-75% from sufficient to good and 75%-100% from very good to perfect information retrieval.

Finally, a further problem was posed by the absence of a yardstick with which to compare the results obtained by participants watching subtitled news. Can we indeed expect viewers under normal conditions (no subtitles) to obtain 100% of the visual and acoustic information of a news clip? In order to answer this question, a preliminary test was run with 14 other students (from the above-mentioned class at Roehampton University) who watched the same clips with sound but no subtitles and were asked the same questions. The following two sections include the findings obtained in the three tests (no subtitles, 180 wpm and 220 wpm) and a discussion of these findings.

2.4.1. Findings

Graph 2 shows the overall results, whereas tables 1, 2 and 3 use the above- mentioned rating (from very good to very poor) to assess the performance of the participants in the tests with no subtitles, 220 wpm and 180 wpm

Pablo Romero-Fresco

10 30 50 70 90 No Subt.

180 wpm

220 wpm Total info_obtained

Total info conveyed

respectively. For the latter two, a comparison is drawn in these tables between the performance of the participants and their opinion on the speed of the subtitles:

Graph 2: Comparison of the overall results No subtitles Performance Very good 100% Good 0% Almost good 0% Sufficient 0% Less than sufficient 0%

Poor 0%

Very poor 0%

Table 1: Performance of the participants without subtitles Subtitles at 220 wpm Opinion Performance Slow 1% 24% Good 0% 20% OK 23% Almost good 6% Sufficient 13.3% Fast 76% 76%

Less than sufficient 20%

80%

Poor 30%

Very poor 30% Table 2: Opinion and performance of the participants

Comprehension and reading patterns of respoken subtitles for the news 183 Subtitles at 180 wpm Opinion Performance Slow 13% 66% Good 3% 49% OK 53% Almost good 6% Sufficient 40% Fast 33% 33%

Less than sufficient 20%

51%

Poor 21%

Very poor 10% Table 3: Opinion and performance of the participants

with subtitles displayed at 180 wpm

2.4.2. Discussion

In the no-subtitles condition, all participants achieved what was regarded as very good comprehension, particularly of the images (90.5%, as compared to 73.2% of the verbal information), which is normal considering that no subtitles were displayed.

As for the test with subtitles at 220 wpm, 80% of the participants did not obtain sufficient information, only 20% obtained sufficient information and none obtained good information. Besides, 60% could only remember a poor or very poor account of the news. Although not surprising, given the high subtitle speed, these results warn against the possibility of producing verbatim subtitles for certain programmes such as debates, interviews and weather reports, which are sometimes spoken at this rate. Indeed, most viewers (76%) considered these subtitles to be too fast. Many of them also added that it caused them ‘stress’ and ‘headache’ and pointed out that the images were too fast, which, although not true (they were as fast as in the other clips), goes to show how the speed of subtitles can affect the overall perception of an audiovisual programme.

The test with subtitles displayed at 180 wpm is more significant, as respoken subtitles are often displayed at this speed in some sports programmes and many news programmes, interviews and debates. In this case, most participants (66%) were happy with the speed of the subtitles and yet, more than half of them (51%) did not obtain sufficient information. This suggests that viewers may be unaware of how much information they are losing due to the speed of respoken subtitles. Thus, although most of them regarded the speed as OK or even too slow, only 3% obtained good information and 31% got poor or very poor information. More worryingly, 1 out of 3 participants acquired incorrect information, believing, for example, to have seen the President of Nicaragua or Tony Blair, neither of whom appeared on the news.

Pablo Romero-Fresco