Research

Why the first screen starts with one recording

How Ptichi's developing Home screen connects one clear recording action to guided practice, and what its weekly activity can actually tell you.

You can begin voice practice with one short message and a reason to listen to it. You do not need to decide whether your pitch, pace or confidence needs a higher score first. Ptichi's developing Home screen puts an explicit recording action inside a guided practice session, so the first take already has a next step: listen back.

The same screen shows recent activity. Those numbers answer a smaller question: how much recorded material remains in the local practice history? They cannot tell you how much you learned.

Development source review · 2 October 2026: This article describes the current desktop source and our design reasoning. We did not run the app, a native build or microphone tests for this article. There is no public desktop release or installer; the website does not record audio.

Give the first recording somewhere to go

A Record button is easy to understand. The harder part comes when the recording stops. What should you listen for? Should you make another take, open a report or choose an exercise from the beginning?

We want the first action to belong to a practice sequence rather than leave those decisions disconnected. In the current desktop source, Home uses the existing guided Practice workspace. It does not create a separate quick recorder with its own rules for what happens next.

The native starting task is Status + next step. Its listener goal is concrete: make clear what changed and what happens next. Home presents the prepared material and recording control. When the setup permits recording, the first Record action creates the guided session and starts its first take. That session carries the exercise forward into listening, one change, another take and comparison.

This is a deliberately narrow starting point. The current entry is not an automatic diagnosis of your speaking problem, and it does not select a personal exercise from an analysis of your voice. Choosing a different practice remains a separate option.

There is a useful distinction here for anyone practising with an ordinary recorder. You can decide what the message needs to achieve before you know which part of your delivery, if any, should change. “Make the next action clear” gives you something to listen for. It does not require you to predict that the answer will be slower speech or a longer pause.

The first recording gives that question an actual example. Listening may reveal a delivery issue. It may also reveal that you never stated the action, or that the message was already clear enough. Each is a more specific starting point than a general instruction to improve your voice.

A simpler entrance still needs explicit control

Reducing the number of steps must not remove the moment when a person chooses to record.

Opening Home can read microphone availability. The current availability check reads information about access and devices; it does not request permission or open an audio stream. A connected device is therefore a limited status claim. It does not mean a take has started or that the app has established recording quality.

If the recording setup is missing, Home's main action leads to microphone setup. It does not silently create a practice take while the person is sorting out access. Once the setup allows recording, Record remains an explicit action.

The source also treats a delayed start carefully. Creating a session can take time. If the person leaves Home while that operation is pending, its eventual result must not start recording on the page they have left. The implementation checks whether that entry action is still current before starting capture, and guards against duplicate activation.

These are source-level safeguards, not evidence that every real device or interaction has passed testing. Their design purpose is clear, however: one visible action should remain one understandable intention, including when the person changes their mind.

Why we reuse the practice workspace

There are several plausible ways to design this entrance. The following comparison is our current analysis, not a record of user experiments or alternatives the team previously tested.

An analysis-first screen could offer pace, pitch and other measurements immediately. That may suit someone who already has a specific measurement question. For the starting job here, it asks the person to choose what to measure before hearing the message they want to practise. It also makes the first useful action depend on understanding the measurements.

A large exercise menu could offer more choice. That is useful when the person knows their task or wants to browse. It also leaves the person responsible for selecting the route before making a first attempt. A default task can reduce that decision, but only if its purpose is visible and there is a clear way to choose something else.

A separate quick recorder could make the first screen very small. The development cost appears later: the product then has to decide how that recording joins a guided exercise, how a second attempt relates to it and where the person returns to listen. A quick recorder can be useful in its own right. It does not automatically provide that continuity.

Reusing Practice keeps Home's first take inside the existing session. The team has one practice sequence to maintain rather than a second entry that imitates only its Record button. The user-facing intention is equally modest: after starting, you should not have to assemble the rest of the rehearsal from unrelated tools.

This choice does not remove every navigation question. In the current source, Recent practice finds the session containing the newest retained take. Its “Listen back” action opens that session in Recordings without automatically playing audio. Finding the right session, choosing a take and playing it remain distinct actions. A return link helps only if it takes you somewhere you can recognise and use.

What “this week” includes

A recording-focused Home screen also needs a careful account of activity. Otherwise, a small history panel can quietly become a judgement about the person.

The current weekly summary uses the duration attached to each retained source recording. It does not substitute the duration of an analysis result. A take can therefore contribute recorded time without a successful automatic analysis.

The summary follows three rules:

  • Recorded time adds the known durations of retained takes made during the week
  • Sessions counts distinct sessions with at least one retained take in that week
  • Practice days counts days with at least one such take, using the device's local calendar

The week runs from Monday to Sunday. The date of the take determines where it belongs, rather than the date its session was first created. A new take in an older session can therefore contribute to this week's activity. Deleted takes and future-dated takes are excluded.

Consider an authored example. On Monday you make a 40-second take in a session started last week. On Thursday you add a 50-second take to that session. You also have a Thursday take in a second session, but its duration is unavailable.

The known durations add up to 90 seconds. There are two sessions with retained takes and two practice days. The duration total is partial: the current display marks it as at least 1.5 minutes and explains that some recording durations are unavailable. It does not quietly assign zero seconds to the unknown take.

If all relevant durations are missing, the duration display uses a dash. If the history itself cannot be loaded, Home shows an error and a retry action instead of reporting an empty week. Those states need different meanings: no retained takes, an incomplete duration total and an unreadable history are different situations.

The scope also explains why a total can change after deletion. It is an account of recordings still retained in this history, not a permanent record of everything you ever practised. Reading progress stays in the Library rather than being added to an invented total of all learning activity.

Minutes can help you return; they cannot explain the result

A known recording duration is useful. It tells you something about the material available to revisit. A day with a retained take can remind you that there is a session worth reopening. Neither tells you what happened during the practice.

The same 90 seconds could come from a careful comparison, repeated attempts at an unclear sentence, or a message that already worked. Duration alone cannot distinguish them. Recorded time also leaves out time spent listening, thinking or practising without recording. A quiet day in this panel is not evidence that you did no useful work on your speech.

That is why we read these numbers as activity. To say that a change helped, we need to know what was being attempted and what became different. To say that it helped another listener, we need an observation from a listener. To say that it lasted, we need evidence beyond the current attempt. A duration total contains none of those answers.

Our product interpretation is that Home should help someone begin or return, while more detailed review keeps the relevant recording and task together. We do not need to turn every minute into progress to make the history useful.

There is also a risk in showing activity at all: a person may still read a taller bar as a better day. Clear labels cannot establish that people interpret a display as intended. If observed use showed that the panel encouraged unnecessary recording or made short, useful sessions feel inadequate, that would be a reason to change its prominence or presentation. We have not established that outcome in either direction.

Try the entrance before judging your voice

You can try the underlying idea with a recorder you already have. Use non-confidential material and keep the recording setup steady. The following is an authored rehearsal, not a tested Ptichi intervention.

Suppose a workshop has moved from Room 2 to Room 5, and the start time remains ten. Your colleague needs the new room and needs to know whether to change their arrival time.

Take A: Record how you would tell them, without writing a polished script first. Stop when the message is complete.

Listen: Can you hear which room they should use? Is it clear that the start time has not changed? Notice the words that carry those facts before judging whether you like the sound of your voice.

Change one thing: If the new room arrives after a long explanation, put the room change first. Keep the unchanged time. If both facts were already clear, choose a different short message that actually needs work rather than inventing a fault to correct.

Take B: Say the same message again with that one change. Compare the takes for how directly the facts can be heard. A shorter version that loses the time is not a useful repair.

Transfer: Put the original message away. Explain that a collection point has moved to the side entrance while the collection time stays the same. Try carrying over the decision to state the change clearly, without copying the workshop sentence.

Your own comparison is a self-listening judgement. It does not establish what a colleague would understand or prove lasting learning. Even so, it can leave you with a concrete note: “Putting the new room first made it easier for me to find on replay.” That is more informative for your next attempt than “I recorded two minutes.”

The starting point is still a hypothesis

A single guided entrance can be too narrow. Someone preparing a story, reading a poem or returning to a known exercise may need a different first action. Someone who has difficulty choosing a useful speaking task may need more orientation before Record. Making the button prominent does not answer what they should say.

The strongest unresolved question is therefore whether this Home screen helps people understand and complete a useful practice sequence. Reading the code cannot answer that. Nor can it show that the sequence offers enough value over an ordinary recorder and a short checklist.

We would reconsider the default if observed use showed that people regularly entered the wrong task, could not explain what to listen for after the first take, or had to leave the sequence to find a useful next action. A simpler screen would not compensate for those failures. We would also keep navigation success separate from evidence that speaking improved.

For a real short update, continue with the project status update guide. For a conversation coming soon, use the pre-meeting rehearsal with your existing recorder. Both give the recording a purpose without waiting for the desktop release.

That is the practical idea behind the first screen: begin with a message you can hear, make the next action clear, and let the history say only what it knows.

What you can verify on this page

  • This page includes a Ptichi-authored example built for explanation or rehearsal.

    What it does not support: It is an editorial example, not an observed-user result, experiment or proof that Ptichi improves speech.

  • This page includes a bounded listen-and-compare exercise that you can run with your own recorder.

    What it does not support: The exercise does not prove that Ptichi improves speech or that a second take will generalize to other listeners or situations.

Reviewed: