Remote Simultaneous Interpreting Options and Standards We shape

Remote Simultaneous Interpreting: Options and Standards “We shape our tools, and then our tools shape us. ” John Culkin Cyril Flerov, AIIC, TAALS Russian Conference Interpreter ATA 2015 Miami, FL, USA

As a community, we need to have an intelligent balanced conversation about Remote Simultaneous interpretation (RSI). This presentation is generic and theoretical in nature and does not imply any flaws or advantages of any RSI platforms or developers.

UNESCO Symphonie Satellite 1976 NYC-Buenos Aires United Nations experiment 1978 European Telecommunications Standards Institute study on ISDN video telephony for conference interpreters 1993 Various tests by the European Commission 1995, 1997, 2000 United Nations tests 1999, 2001 European Parliament 2001 History

RSI – circa 1996 Extensive interpreter training Medical environment Mothers and physicians preferred RSI Interpreters preferred consec. Source: Interpreting in the 21 st Century: Challenges and opportunities (Benjamins Translation Library)Dec 20, 2002 by Giuliana Garzone and Maurizio Viezzi History

Internal cable External cable “From home” Types of Remote Interpreting

Interpreters are in adjacent room Fewest issues Support by onsite technicians Standard booths Standard interpreter console Not VOIP Your boothmate is next to you. Only difference: Quality video feed: 2 screens (presenter & Power. Point) “Internal cable”

Interpreters in next building or far away No easy access to delegates Support by onsite techs May or may not be VOIP (dedicated VOIP line) Professional booth and console Your boothmate is next to you in booth Video may or may not be available: a need for video “External cable” RI

All interpreters were located at the Main Press Center Dedicated feed (including video)

“Internal cable” – possible “External cable” – may be possible IF Professional equipment and booth Onsite tech support Dedicated ISO audio and video feeds Proper hiring practices Interpreters physically working together (same booth) Done in same Interpreter Center

RI “from home”

Individual interpreters work from home Your boothmate can be half world away Multiple platforms VOIP, interpreters provide hardware (headset, cable internet, computer) Audio only or audio and video Various business models: platform only, freelancers buy SI time, full time? No standardization Hiring standards No common approach

Early stages: you come to China

Marketing messages: how “from home” RI may be presented to clients and interpreters

Lumped together, but we cannot equate 1) “internal”, “external” and 2) “from home”

Panacea for all ills Cutting edge state of the art technology WE developed Crystal clear sound Revolution Innovation “Disruption” "universal translator“ Star Trek come true for any language or situation For clients:

You may disrupt something, but what will you put in its palce?

Luddites Technophobes Are unwilling to change: change will be there sooner or later and we must jump on the bandwagon and guide it There is INCREASED demand for it Interpreters, separate reality from marketing language

You will be jobless if you do not do it. Competition is tough so if you do not do it, someone else will. It shows you cannot adapt. Better paid by the hour than nothing at all. Technology is as good as conventional simul and you do not have to travel. Other negative or wrong messages

Does it have genuine potential? Is the technology ready NOW? Will quality be compromised compared to conventional simul? User friendliness for client and interpreters How to separate marketing language from real capabilities? What standards will be used? Who is driving the changes? What are hiring practices? Practical considerations, what can we do? What the REAL questions are:

Red Flags

Remote simultaneous translation, User friendly interface for a successful remote simultaneous translation Crystal clear sound (crystal clear is a marketing term not a technical term) Real time audio and video. (How “real time” can it be? ) Real-time video is available as an option. (So interpreters will not have video? ) Audio platform only Cell phones will be used though for short periods Red Flags

Sound Quality

Sound quality of remote SI platforms should not be evaluated by interpreters subjectively (like, do not like, OK). For that ISO standards exist.

Human voice 80 Hz to 14 k. Hz Telephone calls limit 300 Hz to 3. 4 k. Hz. ISO SI Standard: 125 – 12. 500 Hz Wideband audio 50 Hz to 7 k. Hz Wideband audio = Another Buzzword Why VOIP quality is bad?

VOIP Quality Comparison.

Why it happens?

Treble is higher frequencies Treble is important for speech intelligibility (especially in old age!) Treble boost needed 12, 500 Hz vs 7, 000 Hz: why the upper 5, 500 Hz are important?

Comparison: low pass filter at 7, 000 Hz (frequencies above 7, 000 Hz deleted): “Muddy” voice Treble may be artificially removed even more to reduce feedback

High Pass filter 300 Hz: anything lower than 300 Hz deleted (phone line) Bass Loss and Intelligibility: Bass is important for speech intelligibility

VOIP quality loss is partly because of narrow frequency reproduction Missing treble (high frequencies) make voice sound “muddy” and less intelligible “Tinny voice” is the result of cutting off frequencies less than 440 Hz and greater than 3500 Hz Metallic voice – loss of bass High and low pass filters combined

“…at least one of the audio connections is via the conventional telephone network. ” From an RSI patent Telephone input

cell Long dist HD Audio Different codecs Long Distance: G. 229 Cell: GSM-HR (“half rate”) G. 711 best legacy codec Landline and cell phone input Source: Turbo. Bridge

Human Voice Studio quality Wideband audio is no advantage if landline or cell phone input is used – still phone quality.

Fuzzy Voice - too much gain on the signal Synthetic Voice – packet loss or delay Choppy Voice - consecutive packets that are lost or excessively delayed Clipped Voice - words are cut off. Listener and talker echo: caused by transmission issues or speaker’s acoustic environment Noise from source mic Other issues in VOIP not related to frequency range:

An office or a room at home are just as good as the booth. No need for booth. Wrong. Booths exist for a reason. Extraneous noise. Sound insulation issues. Flooring, echo, ventilation, household noises, sudden outdoor noises “Reverberation time (see ISO 3382) inside the booth shall be between 0, 3 s and 0, 5 s measured in the octave bands from 125 Hz to 4000 Hz (booth unoccupied). ” Hard to achieve at home – determined by acoustic properties of by surrounding surfaces.

Room 10 x 5 x 3 meters Floor: linoleum on concrete Wall: plastered Ceiling: Gypsum Door: heavy Window: 3 mm Reverberation Calculator

5. 1 – 10. 9 times worse than in a booth!

Compare to 0. 3 – 0. 5 ISO standard for booths. Even a conference room is not ISO compliant Bathrooms: 0. 5 - 2 seconds Bedrooms: 0. 5 seconds Source: buero-forum

Ceilings: acoustic ceilings, baffle systems, edge absorbers Walls: textiles, wall absorbers Floors: textile floor coverings, carpets Sound absorbers in furniture The shorter reverberation time - the better speech intelligibility. Acoustic absorbers

Systematic Use of Relay?

From an RSI patent

“In this example, words of an English-speaking delegate … conversing with a Spanishspeaking delegate would be translated four times before reaching the other party. ” From an RSI patent Use of Relay

Codecs Noise Relay Acoustic Shock

If sound quality is bad for the interpreter, it is bad for participants too. Everyone is on equal ground. Wrong. Very different cognitive task. Delegate may afford to skip a word or a sentence. . Interpreter needs better sound quality than a delegate

ISO 2603 --Fixed booths for simultaneous interpretation ISO 4043 --Mobile booths for simultaneous interpretation “The overall system (comprising microphone input at the speaker’s position, amplifier stages, level controls, output terminals and interpreters’ control panel for headphones), shall correctly reproduce audio-frequencies between 125 Hz and 12500 Hz. A gradual rolloff at the lower end of the frequency response is recommended in order to improve speech intelligibility. ” ISO Standards are there for a reason:

Other messages and considerations:

Equipment will be in different locations but it does not matter. Yes, it does. Higher probability of hardware failure. Unreliable home broadband, no dedicated line. Using chat and interpret at the same time? Handover? Failure scenarios? Communication? Redundancy? Need for common procedure, no tech support

All you need is a PC, headset and broadband connection. No standard requirements exist for headset quality, noise cancellation capabilities, computer operating systems, processing power, broadband line quality, UPS for PC and cable modem, backup means to contact technicians and participants in case of system failure etc. Interpreter selected hardware – another uncertainty factor

Shock from: the line or own headset IP network shrieks, mic feedback fax machines, dropped phone, malicious noise, power supply failures 85 d. B safe limit “…with Vo. IP connections of a high quality the sound intensity easily reaches a higher level” Bad connection = headset volume Correct headset type and adjustment Health and Safety: Acoustic Shock

Headaches Balance disturbances Nausea Pain (ear, head and neck) Hearing loss Tinnitus (ringing in the ears) Hypersensitivity to high level sounds Stress Acoustically Safe Headsets ? ? ? Possible Consequences:

Professional interpreter headsets are NOT designed to be USB e. g. Beyerdynamic DT 394 SIS (see my review here https: //app. box. com/s/s 09 t 3 lrxmn 9 i 5 tsusvl 7) Requires mic amp with phantom power and ideally a headset amp to mimic professional interpreter console. Any headset = Kitchen knife for surgery USB headset is convenient for the computer and not for the interpreter. Conversion happens in headset not in PC sound card so you depend on the quality of the headset. USB or no USB?

USB Headsets bypass computer sound card Importance of bass and treble Bass and Treble control in interpreter interface? Will USB headset support it? Bass and Treble Control

Bass and Treble Adjustment Experiment with controls, start in the middle Advanced Simultaneous Interpretation Technique 56

We work hard to make things better. Making things better does not yet mean making them ISO compliant or commercially usable and interpreter friendly. It is just an interesting experiment not fully commercially usable.

Balance reality with opportunity It may be an opportunity for platform designers or businesspeople. For professional interpreters working in substandard conditions is not an opportunity but exploitation.

RI is convenient for delegates Not at the expense of interpreters and their health RI a great opportunity for interpreters See above. RI is best for regular meetings happening periodically. We know that most assignments on the market are one time. So will it be in RI too.

who is liable in case of equipment failure how to determine in real time whose interface failed: on the interpreter, client or platform side? will the interpreter still be paid in case of equipment failure? recording interpretation? cancelation policy Legal Issues:

Not to replace face to face meetings However, longer meetings are already being supported with it and there is no guarantee that clients will not want it in lieu of standard SI, even though the RI systems are said not to be designed for it. Who will make that decision? Need for standards Then it is disruption of what?

We will use cell phones only for short periods of time. What is short? If you use it once you use it always. Who decides what “short” is? Participants are forgiving. We are judged by our output and forgiving is not a scientific term. It does not mean we can sacrifice quality. Health concerns: stress, acoustic shock, feedback, static – length of shift?

Secure connection You depend on browser, computer, provider security settings and flaws

WIFI streaming to smartphones Installed app Issues: ◦ What is added value? ◦ No universal platform support ◦ Unreliable end user hardware (battery life) ◦ Need for a backup option ◦ If streamed outside of conference room: data throttling and poor internet connection ◦ Streaming protocols add to lag, mix of smartphones and receivers: different pauses? ◦ Mixed results?

What can interpreters do?

“How do you tell a communist? Well, it is someone who reads Marx and Lenin. And how do you tell an anti Communist? It’s someone who understands Marx and Lenin.

Separate reality from marketing hype. Understand a market need for RI Appreciate commercial entities’ desire to use the opportunity Resist lowering standards of the profession Refuse to accept substandard assignments or platforms Educate clients about real state of RI and existing quality issues We need to:

- interpreters are not accountable for the quality of their work (best effort to deliver) - a need for disclaimers in contracts. - inform clients about possible difficulties when using RI (use at your own risk) - RI only for short meetings, not to replace regular SI - decision by interpreters. - participants should be informed that RI is being used - no use of cell phones, Polycoms - properly trained technicians are needed.

Discuss working conditions in advance Refuse to interpret if sound quality or conditions are substandard (“Inaudible”) Your employer and clients must know beforehand that you shall not interpret if sound quality is substandard and you shall not make an attempt to “guess” what the speaker is saying. It will be a good indication of the “quality” of the SI platform.

Source: AIIC Cd. P 10 Survey

Remote Simul will only really be ready if it provides the same working environment and quality (including ISO compliant sound quality) as conventional simul. As of today (November 2015), no such system or platform exists.

Remote Interpretation is already a reality…. . Is it?

“Until sound and video quality over the Internet are improved, I do not think it should be used for interpretation. Remote interpretation with a dedicated cable connection is feasible as long as the sound is of high quality. ” “Remote technology is not there yet as far as quality of sound and interpreter comfort as well as overall system reliability. “ AIIC USA Survey Responses

One time large scale “external cable” RSI projects with large budgets Government RFI and RFPs, trial/test projects More internal cable and external cable “From home” RI is still not ready for commercial application (ISO compliance, sound quality, cumbersome interface, using poor quality inputs e. g. cell phones, availability of video, hiring standards, onsite technician) Efforts by interpreters, Danger of dehumanization and deprofessionalization Future:

http: //aiic. net/page/1125/remote-interpretingassessment-of-human-factors-and-performanceparameters/lang/1 AIIC: Remote interpreting: Assessment of human factors and performance parameters http: //aiic. net/page/143/guidelines-for-remoteconferencing/lang/1 AIIC: Guidelines for Remote Interpreting Useful Links

Are critics of RI Luddites? You cannot stop progress.

“Yet it is to be observed, that the work thus executed was inferior in quality, not marketable at home, and merely hurried over with a view to exportation. It was called, in the cant of the trade, by the name of Spider-work. ” Lord Byron, Speech in Defense of Luddites, 1812

“Above all, the destruction of the Luddites by the State established the principle that industrialists have the right to continually impose new technology, without any process of negotiation, either with the people who have to operate it or with society at large. ”

“[Luddites] just wanted machines that made highquality goods, and they wanted these machines to be run by workers who had gone through an apprenticeship and got paid decent wages. Those were their only concerns. ” Historian Kevin Binfield

Join my group on Linkedin

Email for more info if interested: c. flerov@aiic. net

Practical recommendations to students and interpreters Exercises Use of Computer Software Voice and Equipment Skills “Improving the Interpreter Voice” by Cyril Flerov and Michael Jacobs The book will be published later this year.

Thank you! Cyril Flerov, AIIC, TAALS Russian Conference interpreter San Francisco, CA, USA c. flerov@aiic. net