A Translation Can Be Linguistically Correct but Still Look Wrong
Translation quality is often measured by asking whether the translated sentence is correct.
But digital products introduce another problem.
A translated sentence can be linguistically accurate and still create a poor user experience.
A button may become too long.
A headline may wrap onto several lines.
A translated label may overlap an image.
A caption may appear in the wrong location.
LILT’s September 2026 product release addresses this problem with its new In-Context Preview.
Seeing Translation in Its Final Environment
The new feature allows users to preview translated content in the original layout.
This means translators and reviewers can identify problems that would not necessarily be visible when reading translated text separately.
It represents a broader change in how translation quality is evaluated.
Instead of asking only:
“Is the translation correct?”
the industry increasingly needs to ask:
“Does the translated information work correctly in context?”
Context Is Also Important for Translation Devices
The same principle applies to handheld translation devices.
A translation device does not simply output text.
It may display:
- Original language
- Translated language
- Speaker information
- Conversation history
- Bilingual subtitles
- Photo translations
- AI responses
The way these elements appear on screen influences how quickly users understand the information.
Why Bilingual Subtitles Matter
The F10 includes real-time bilingual subtitles.
This is useful because users do not always want to rely only on audio.
A user can listen to the translated voice while checking the corresponding text.
This can be particularly useful in noisy environments or when discussing important information.
The visual layer can provide an additional reference point during communication.
Photo Translation Creates Another Context Challenge
Photo translation creates an even more complicated situation.
The system needs to identify text within an image while understanding where the text is located.
For example, a product package may contain several labels.
A restaurant menu may contain multiple columns.
A machine may contain small technical markings.
A translation system that simply extracts text may lose some of the visual relationship between words and objects.
This is why contextual understanding is becoming increasingly important.
Context Databases and AI Assistants
LILT’s Context Database also highlights the growing importance of structured language resources.
Translation systems can use:
- Translation memories
- Terminology
- Style guides
- Previous translations
- Product-specific vocabulary
For an OEM translation device, the same concept can be applied to specialized vocabulary.
A device designed for manufacturing customers could potentially support terminology related to:
- Electronics
- Machinery
- Logistics
- Quality control
- Production
- Safety
This could make the device more useful in professional environments.
What OEM Manufacturers Can Learn
Translation hardware developers should consider context at several levels.
Language Context
What does the speaker mean?
Visual Context
Where does the translated text appear?
Conversation Context
What was said previously?
Industry Context
Is the speaker discussing medicine, manufacturing, logistics or another field?
User Context
Is the user traveling, negotiating or receiving customer support?
These factors can influence the usefulness of a translation device.
The Next Stage of Translation Quality
Translation technology is gradually moving away from simple sentence conversion.
The industry is beginning to treat translation as part of a larger information experience.
The F10 and F80 already combine translation with recording, photo translation, subtitles and AI assistance.
For future OEM products, contextual understanding may become just as important as language coverage.
A translation device should not only translate the words.
It should help users understand where the words belong, what they refer to and how they fit into the conversation.