Reveal hidden emotions.

Remotely and instantanously.

Trusted by 1 328 customers worldwide

What is iCode?

It is the entry point into your customers’ subconsciousness through implicit testing.

iCode is a browser-based platform where you can design, execute and analyze implicit reaction time tests. Our algorithm decodes people’s genuine emotions, attitudes, and opinions, free of conformity and rationalization. It seamlessly detects confident and hesitant responses providingaccurate and actionable results.

iCode is not yet another implicit association test. We do not torture our participants with extensive surveys, and we don’t ask them the same questions over and over again. Instead, our proprietary and patented algorithms allow asking 30 questions in just 4 minutes, and the results are available instantaneously. No more waiting for results – with iCode, you can essentially watch a live stream of your research.

Along with access to the platform, we offer full tech support every step of the way. We also provide our clients with our scientific expertise stemming from over 20 years of experience in consumer neuroscience and human behavior.

Our scientifically validated software guarantees 95% average consistency with sample sizes of only 80 participants.

Too good to be true? Keep reading or schedule a free demo 

Made so simple...

Watch a short demonstration video showcasing the ease of study design in iCode.
There is no better way to see iCode in action than by taking a demo test.
Here you can see
a browser-based report example.

... yet so incredibly versatile

iCode’s unparallel reliability and scalability made it a highly applicable tool in market research, consumer insights and beyond. With our ever-growing client base and collaborative R&D department, we constantly find new areas for iCode to shine in. We would love to show you how iCode can help your business grow.

Brand Testing

  • test your brand image
  • compare you brand to your competitors
  • track changes in your brand perception

Product Performance

  • find out what people really think about your product – no matter what they declare
  • check if your product will succeed
  • explore customers’ needs for your product

Ad Effectiveness

  • test the strength of your communications
  • check which elements need improvements
  • compare campaigns and pick the best one

Concept Strength

  • find what your customer genuinely need
  • discover what concept best describes your products
  • create a message that appeals to consumers on the emotional level

Package Design

  • design products that people desire
  • color, shape, or opening – check which areas require lifting
  • rank your products and roll out the one with the highest potential

Sensory Research

  • test any type of fragrance or taste
  • check what wording corresponds best with your product’s sensory pallet
  • pick a perfect scent for your product

Claims Testing

  • pick the proper description of your brand or product
  • evaluate your communication
  • rank your claims and pick the most effective one

HR Research

  • define the problems the company is facing
  • determine the strongest assets of your team
  • use a battery of questions for scanning and diagnosing your company’s effectiveness

Proven by science

Merging explicit declarations with implicit response time to better predict behavior

Implicit Measures of Attitudes in Market Research

Big Data Is Decision Science: The Case of COVID-19 Vaccination

Decades of experience
in consumer neuroscience and human behavior.

Now available at your fingertips.


Years of experience in consumer neuroscience leads NEUROHM to investigate implicit response tests to gain deeper insights. The team develops iCode, an implicit reaction time tool for market research and customer insights.


iCode becomes widely present as an in-lab solution across Europe. We collect data from thousands of participants, enhance algorithms, and build our benchmarks across various sectors.


After seven years of meticulous craftsmanship, we roll out iCode as browser-based platform for experiment design, data collection, and analysis. Shortly after, iCode becomes available on all mobile devices.


As remote access attracts worldwide attention, NEUROHM partners up with academia and the private sector. iCode continues to shape the industry outlook on measuring consumer non-conscious processes. Our insights result in business and science growth globally.


The US patent arrives. It’s official – there is no other software like iCode. With rapid technological development and enormous possibilities, iCode is applied whenever people’s opinions matter, becoming a vital tool for marketers worldwide.

Don't just take our word for it...

iCode easily identifies better performing fragrance

Did you know that smart test can also empower sensory research?

Case overview

Client: A hair care company
Research method: iCode Smart test
When: 2020
Goal: choosing the best fragrance for a new hair product


  • 89% of respondents marked Fragrance A as refreshing, even though only 45% truly believed this was true
  • Fragrance B was comforting on an emotional level by 60% respondents
  • Fragrance B scored better results and the company chose it for their product

Client’s challenge

Before releasing a new product, a haircare producer sought to investigate the subconscious perception of different fragrances. It was important to the company that the chosen fragrance evoked feelings of comfort and freshness, and they wanted their product to represent these attributes. Due to the pandemic, however, it was impossible to run lab tests.


Because the labs were closed due to the pandemic, the haircare company decided to conduct an in-use survey, in which samples of hair products were shipped directly to the respondents. The responders were asked to rate the aromas with an iCode test right after washing their hair with the client’s product. This provided a rare evaluation in real-life circumstances. We selected the audience group, sent hair products to respondents, and then used iCode to measure their implicit reactions juxtaposed with declarative answers.


Fragrance A
Fragrance B
  • both scents are perceived as refreshing and comforting
  • the release of each of the fragrances will achieve the intended goals
with iCode
Fragrance A
Fragrance B
  • implicitly Fragrance B was perceived as refreshing and comforting significantly more than Fragrance A
  • the second fragrance is recommended for release, as its perception is much closer to the impression envisioned by the company
With a little help of delivery services, we managed to test fragrances during the pandemic. Luckily, iCode is available on any device at all times. We established that the majority of people claimed hair product fragrance to be refreshing and comforting, even though they didn’t really feel this way. iCode revealed the truth – less than ten half of the respondents truly enjoy Fragrance A, while Fragrance B was quite well-received. We recommended the client to use Fragrance B, and they did benefit from this choice.

How iCode saved a company from running an ineffective coffee ad

Coffee is by far the most popular and effective stimulant we use on a daily basis. But how can we make a coffee advertisement just as effective at sharing desired values and nourishing the brand image? iCode is the answer!

Case overview

Client: Segafredo
Research method: iCode Smart test
When: 2019
Goal: checking the effectiveness and value-recognition within a new campaign


  • 13% more effective communication
  • 5th position among professional Italian coffee bean distributors
  • 5% growth in coffee beans distribution
  • gaining insights into emotional reactions – vital for future campaigns

Client’s challenge

Segafredo launched a new advertising campaign for the Selezione product line in 2018. The brand opted to evaluate the campaign’s effectiveness before continuing in 2019. The primary purpose was to see if the desired values were clear and associated with the brand.


We tested Segafredo’s fundamental values and attributes to their audience using iCode smart test after showing the original verstion of their ad. We measured the respondents’ implicit reactions.
Original version of the ad


  • only 29% of consumers associated Segafredo with being a barista at their own home and with motivation to act big
  • conclusion – two main attributes were not yet well established with the brand


  • the communication of the ad should emphasize the abovementioned values
  • on top of that, based on iCode results, we advised visual changes – beauty shots improvement and stronger brand exposure at key moments to better link essential values with the Segafredo brand

Road to the goal

Segafredo team rapidly implemented our recommended changes to the ad. We ran another part of iCode tests to check whether the new ad transfers desired values, and connects them to the brand.
Improved version of the ad


  • 39% of consumers now associated Segafredo with being a barista at their own home (10% increase)
  • 45% of consumers now associated Segafredo with motivation to act big. (16% increase)
  • the changes implemented to the ad resulted in boosting the brand image



  • the ad is now ready to be launched on air!
  • the communication of the ad now has a higher potential to move consumers emotions and motivate them


Using iCode, we were able to determine if consumers truly believe in values important to our client – not just on a declarative level, but also subconsciously. We measured if the ad transfers these values and puts the right amount of emphasis on the brand image. As it turned out, the advertisement required enhancements in several areas. The upgrades were eagerly implemented, and the post-test results revealed substantial growth in the ad effectiveness. As the Segafredo representatives claim, investing in neuromarketing research was a profitable experience, and they still benefit from our insights and results.

Joanna Sobyra

Marketing Director Segafredo Zanetti Poland
„According to Nielsen, after spring commercial emission in 2019, the brand moved from the 6th to the 5th position in professional Italian coffee beans. Moreover, there was also a 5% weight distribution. Gaining knowledge about emotional reactions of respondents was the key in preparing better, more efficient commercial version. The investment in neuromarketing research was profitable.”
Are you planning on launching advertising materials soon? Try iCode to measure if your ad communicates the desired values and effectively connects them to your brand. Ask us anything, or book a free demo 

How iCode ensured that an advertisement was well-adjusted for a specific market. Or was it?

Most companies would rather know that something is not right with their marketing materials before launching a campaign, but many of them only rely on declarative surveys. As Budweiser’s case proves, declarative answers tend to miss the point. With iCode, we managed to show Budweiser that an ad they wanted to launch in Poland is really enjoyable, but… there is a big difference between enjoying an ad and actually being emotionally ready to buy a showcased product. Following up on these results, in order to sell their product, Budweiser had to adjust. Read the full story!

Case overview

Client: Budweiser
Research method: iCode Smart test
Goal: making sure that an ad designed for another country would prove effective in Poland by growing the sales potential


  • more than 80% of consumers agreed that the advertisement was moving and encouraged them to buy the beer
  • on an emotional level, over 65% of consumers were confident that the ad was moving…
  • …however, only 20% of consumers were indeed encouraged to buy beer after watching the advertisement

Client’s challenge

Budweiser decided to launch a new commercial on the Polish market, which had been initially designed for another country. The client’s goal was to find out whether Polish consumers would react positively to the advertisement and whether the ad would encourage them to buy the product.


We presented the Budweiser ad to a group of nearly 200 respondents who hadn’t seen it before. We used iCode to measure respondents’ subconscious reactions and juxtaposed them with declarative answers.
  • 82% of Poles agree that women can hold managerial positions
  • conclusion of results without iCode: in Poland, the advertisement will move consumers and positively impact the brand’s image and the sales potential of Budweiser products
  • but is that really so?
with iCode
  • over 65% of consumers were confident that the advertisement was moving
  • after watching the advertisement, only 20% of consumers were indeed encouraged to buy Budweiser beer
  • advised additional sales support communication in Poland


Without iCode, Budweiser would likely launch their advertisement in Poland and then wonder why sales didn’t skyrocket. Using non-declarative metrics allowed us to discover that the new Budweiser ad is enjoyable and will likely warm the brand’s image. However, the lack of an outlined relationship between the product and the presented story results in limited acquisition potential for the Polish consumer. It will not encourage consumers to purchase beer. If the advertisement were to be launched in Poland, additional sales support communication would be needed. And, after all, a brand’s goal is never only to be liked – everybody wants to sell, too. With this study, Budweiser proved that investing in neuromarketing can be highly profitable in the long run.
Are you planning on launching advertising materials soon? Try iCode to measure if your ad communicates the desired values and effectively connects them to your brand. Ask us anything, or book a free demo 

Unraveling the truth about gender bias about managerial positions with iCode

It would seem that the culturally rooted narrative of a male leader is long gone and that managerial competencies have no gender, right? A nationwide study, „Believe in yourself” conducted on a representative sample of Poles in collaboration with SAP allowed us to uncover a secret hidden under the cloak of declarations – until now.

Case overview

Client: SAP Poland
Research method: iCode Smart test
When: 2021
Goal: determining if there are hidden emotions and attitudes towards women holding managerial positions in companies


  • on a declarative level, 82% of Poles agree that women can hold managerial positions…
  • …but, in fact, only 1 person out of 4 believes on an emotional level that women are suitable for managerial positions

Client’s challenge

SAP Poland wanted to build an organizational culture with an emphasis on self-confidence hoping that this approach will translate into the professional development of their employees. In order to do it right, they wanted to measure the emotional certainty of various statements regarding diversity in the workplace.


The test was run on a group of over 1000 respondents. After the test, we compared averaged declarative answers (what respondents selected in the survey) with averaged emotional responses (the emotional certainty of the respondent when answering each question).
  • 82% of Poles agree that women can hold managerial positions
  • conclusion of results without iCode: the belief in managerial competencies of women does not need to be strengthened in either of the sexes. But is that really so?
with iCode
  • the measurement of emotional certainty revealed that only 27% of respondents believe that women are suitable for holding managerial positions!


Without iCode, this research would miss the point, passing over the gender bias entirely. By digging deeper, we found out that gender equity is still a song of the future when it comes to emotional certainty. Thanks to iCode, we were able to establish that the belief in the managerial competencies of women should be strengthened among both women and men. Closing the gender gap in the perception of managerial competences is a challenge that both genders will have to face.

Thomas Duschek

Managing Director, SAP Polska
„I believe that the results of the study will help us build an organizational culture in which it will be important to strengthen self-confidence and this approach will translate into the professional development of our employees. I am convinced that the power of creativity and innovation lies in diversity.”



Prices start at


If you do not feel quite ready to give iCode a go yourself, we are always happy to use your experience and expertise to help you with your research.

Have a question?

Tomasz Zając

Product Manager tomasz.zajac@neurohm.com


Frequently Asked Questions

iCode™ is an interactive, fully-automated platform which allows you to design and run online surveys enriched with Confidence Index™ – a method based on Response Time latency. Quantitative studies give one point of view, Confidence Index™ provides you with an additional dimension. iCode™ helps you to gain more insights, tapping into consumers true attitudes, while still receiving findings from traditional, declarative answers . It’s an intuitive & cost-effective solution. It’s also fast, really fast: you can create your survey in a few minutes and get the results immediately after fieldwork completion. iCode™ gives you freedom and flexibility: it can be seamlessly integrated with other methods and can be embedded as a part of bigger surveys. What’s important for researchers, the platform works 24/7 and can be run on all types of devices, from PC’s and laptops to tablets and smartphones.
What people rationally declare is important, but scientific studies have proved that we can reach more precise insights by measuring the amount of time that people need to provide an answer. The Confidence Index™ is exactly such an indirect measurement that allows us to make more accurate conclusions about the emotional certainty of opinions. Academic researchers, such as Prof. Russell Fazio from Ohio State University, have shown that attitudes expressed with a fast Response Time (high certainty) have a much stronger link with actual behaviour than not fixed attitudes expressed with slower Response Time (hesitation). In other words, the more certain people are of a given answer, the more likely they are to act upon it. Confidence Index™ helps you distinguish between what is truly important and what is just a lip-service.
Confidence Index™ provides added value for many aspects of market research:
  • actionable insights – the additional dimension enables more in-depth conclusions;
  • more complete stories – additional information for Clients;
  • a comfortable way to identify valuable insights – an additional tool to facilitate the research process;
  • the emotional side of consumers’ attitudes – an additional and simple solution to provide Clients with a real way to measure emotions.
For more information on how confidence you can boost traditional results please contact us at icode@neurohm.com with your request.
Measuring Response Time might seem simple, but in reality it is a complex process where many factors have to be taken into consideration. Our unique procedures and algorithms enable us to control the impact of many aspects of the measurement process and respondents’ idiosyncratic characteristics: individual differences (age, gender), speed of reading, familiarity with computers, the type of device (PC, laptop or mobile), fatigue, (in)voluntary carelessness and many more. To be able to do that in a correct manner, we use dedicated calibration exercises and sophisticated procedures during the test to gather high quality responses. The button placement and indeed, the whole layout, is strategically designed to be optimized for the Reaction Time measurement process. The answers for each respondent are standardized and an individual baseline has been created to develop an index that allows for comparison between responses. The stimuli and the statements are rotated to minimize the order of appearance effect.
The iCode™ uses standardized scores for every single individual, derived thanks to individual calibration and baseline. There are also other important elements and additional procedures that were incorporated into the method to increase the quality and reliability of the whole process: (attention control screen, warm-up statements, advance button). All those elements are embedded in iCode™ (made under U.S. Method Patent). Thanks to this standardization, individual differences are minimized and the scores can be presented on a universal scale – either INDEX 4.0 or our Level Scoring System (LSS). Thanks to this procedure we are able to establish what fast/average/slow means for every single individual. In the case of one respondent 800ms can be the fastest while for the other respondent the same 800ms can be at the slow end of an average Response Time. This is also another reason why we do not provide Raw millisecond scores.
The calibration phase is a crucial part of every response time survey and cannot be removed. iCode uses Smart Calibration which means it is adjusted to the actual study and only the necessary elements are used. Still, every calibration consists of two parts: CALIBRATION PART I – It consists of pressing the answer buttons without any cognitive load and – a purely motoric task that establishes the individual differences in speed of movement and familiarity with the device. Moreover, it also serves as a tutorial for respondents as it makes them familiar with using the scale. CALIBRATION PART II – consists of just a statement and one answer button (no scale) that a respondent has to press as soon as he/she has finished reading the statement. This part is designed to test how fast respondents read statements of different lengths. It minimizes the influence of the statement length on corresponding Response Times, thus statements of different lengths can be compared.
NEUROHM is in possession of standardized words used in calibration that are translated to local languages. The list is available upon request.
In general no, this is not recommended. They cannot be changed because they have been standardized.
NEUROHM procedures take into consideration several variables that influence the results. The length of the attribute is one of them and this variable is added to the model during the cleaning process. One of the key parts of every survey is calibration phase – thanks to this procedure we can apply our patented filters that take into account factors such as length of statements and allow us to clean the data effectively.
iCode™ utilizes RORT™ – the fourth generation of latency measures, technologies and algorithms developed by NEUROHM to assure the highest level of reliability. RORT™ technology includes: dedicated software, individual calibration to create personal baselines, warm-up procedures, as well as unique cleaning algorithms to remove artifacts and minimize the influence of various factors e.g.: individual differences, familiarity with computers, age impact, fatigue, length of statements, speed and stability of Internet connections and (in)voluntary carelessness of respondents.
Yes, you can add up to 20 additional questions before and after the implicit survey with no additional cost. iCode™ controls the data collection and presents charts for gender and age only, all other answers will be available in the respondent level data in an excel file.
Currently iCode™ does not provide any fieldwork solutions. Setting up fieldwork is the part of the research process that has to be handled by the User. Of course, iCode™ will provide you with all necessary details to set up the survey on your own or via a panel provider (for more info see: “How to use redirect links” in the HELP section in your account).
iCode™ bridges traditional declarative answers with emotional certainty of opinions, so you always get two measures in one survey. You can ask your traditional questions via iCode and on top of that you will also see the response time reactions of your respondents. The other way to integrate traditional research with the iCode module is to incorporate it in your protocol via redirection links. In the same way a Panel provider can easily run or incorporate an iCode link into any other survey.
iCode™ questionaries can be applied to all surveys where emotional aspects are important. On the platform you will find ready to use designs that will help you to quickly set up a test for brand image, product design, videos & commercial, individuals’ image (people and brand ambassadors) and concept testing. Nevertheless all instructions can be modified and the survey creator is flexible enough to test much more such as: brand assets, customers’ or employees’ needs, photo appeal, DVD or magazine covers, Point of Sales materials and many more.
The pre & post-test is recommended in ad testing. However it all depends on your research needs. If you want to test whether the ad has an impact on your brand, the pre & post-test scheme is exactly what you need so you can compare how your brand heritage (pre test) changed after the ad (post test). Thanks to this scheme, you can clearly measure the impact of the commercial. On the other hand, if you just want to see how effective the ad is or you want to check the comparative effect of a few ads, then just a post-test will be enough and will provide you with insights you need.
The short answer is NO. Isn’t it great? Unlike other providers iCode™ does not charge on a respondent basis – you can test 100, 1 000 or 10 000 respondents and the price will remain the same. The data processing fee depends solely on the number of tested items. However, Please bear in mind that the platform does not provide respondents for your study. This cost is dependent on your approach to data collection – if you use a professional panel provider, you need to be aware of their cost. However, if you run a survey on an internal client database or social media, then the fieldwork is likely to be free of charge, or at least significantly reduced.
An ‘Item’ is a single research screen of iCode survey where Response Time is measured – it consists of an attribute (a statement) and usually a stimulus (e.g. a brand, product, ad) but it might consist of the statement alone. To calculate the number of items in the survey, simply multiply the number of tested elements (e.g. brands, packages, ads) by the number of statements/attributes that describe the tested element. If there is no stimulus tested then the number of items is equal to the number of statements. For example 1 element (e.g. brand, package) tested on 1 statement (e.g. high quality, modern) is 1 item, 1 element x 10 statements = 10 items, 2 elements x 5 statements = 10 items, 3 elements x 10 statements = 30 items. You can also see the number of items, the price and the estimated duration of the survey at any stage of the study.
We recommend testing at least 160 respondents per survey for statistical reasons, however the minimum number required to complete the test is just 10 respondents. The number of respondents depends on various factors such as your target group, the total population of your target group and the incidence rate. We also advise oversampling by at least 10%. This is due to the fact that some respondents will be excluded in the cleaning process (e.g. due to too fast or too slow responses, read more in point 7).
You can conduct studies in one of the predefined languages available on the platform. The list is presented in the START step during the survey creation process. We update the list frequently and more languages will be added in the future. Haven’t found the language you need? Please contact us and we will help you to create a survey in the language of your choice. The list will be updated and more languages will be added soon. However please note that English is the only language for operating iCode™.
It all depends on the test’s type and scheme. The maximum capacity of the survey is 60 items. For example: 1 element (e.g. brand) tested with 1 statement (e.g. question like: is it trustworthy) – equals 1 item. The survey consisting of 60 items will not last longer than 8 minutes and this is the optimal time we recommend to use for online studies that utilize Response Time. This is important due to the attention span of respondents. It is possible to increase the number of items per survey in certain situations, please contact us at icode@neurohm.com with your request.
In iCode™ we have prepared a list of statements recommended for each type of survey. You can choose a few of these for your survey or treat them as inspiration and create your own statements best matched to your research questions. It’s also good practice to organize a workshop with the client (if you are a research agency or consultancy) or customers (e.g. during the focus group) to gather creative ideas corresponding with the research objectives. Another effective way to create relevant attributes is to review core brand values to check which of them are the most convincing. Remember, all statements must refer to emotions and attitudes not to facts and knowledge. It’s important to create a list of statements that:
    1. are connected with the stimuli (e.g. brand),
    2. are representatives of the most important values and features (e.g. of the product image),
    3. are a reflection of the way in which the stimulus is communicated (e.g. via TV commercials),
    4. are ready to be used as an inspiration to produce actionable results – e.g. you can build your communication on these statements,
    5. should be simple in meaning, avoid complex or ambiguous statements,
    6. should be frequently used by the tested target audience,
    7. avoid combined meaning in one go (e.g. “smart and intuitive”, in iCode you should split this statement into 2 separate ones),
    8. avoid negation, as on the implicit level the brain takes additional cognitive effort to process it.
Please be sure that your statements are tailored to Reaction Time testing. Your statements should be short, intuitive, easy to understand for respondents and have only one clear meaning. The statements should be no longer than 80 characters.If you follow these rules you can be certain that you’ll have relevant attributes for your reaction time test. Do not hesitate to contact our team if you would like to discuss your ideas in greater depth.
iCode ™ uses the NEUROHM internal Implicit Benchmark Database ™ gathered since 2004, which contains of over 16 200 000 registered responses from over 50 countries and six continents. Benchmarks are created as a general reference scale created from data collected.
Yes, we do. The results aggregated for the whole group are available on iCode in both forms: as an interactive web report and xls or csv format. The respondent level data is available only in the downloadable xls or csv format, which can be used for further computation or data processing.
We do not provide raw millisecond data because it will not provide you with the data you need or with the quality you deserve. Raw data requires a different paradigm and approach than that applied in iCode. Thus, in such a setting they are meaningless and can be misleading. As an outcome, we provide you with a set of cleaned and standardized data processed by our proprietary algorithms which take into consideration many factors such as: length of words, order of presentation, individual differences, motoric skills etc. All of that is done to deliver the highest quality outcome. The raw milliseconds approach does not account for all those necessary factors, and even if treated according to state of the art academic rules, will still provide weak scores, far from the quality we want to deliver. Moreover, one of the core principles of our company policy is to protect IP rights – yours and ours. If we would provide you with both raw milliseconds data and cleaned data a reverse-engineering of our algorithms would be possible, thus the raw data is not accessible through iCode platform for any of the projects.
It is possible, though we recommend making sure that these words are opposites, for instance: „I agree/I disagree”. However, you must remember that the change in the scale will result in changing the scale within the whole study, including the calibration phase. Regarding longer words, you need to be careful, even if the button description will be facilitated thanks to applied procedures, it still needs to match the size of the buttons. Additional checking in such cases is required as what is well visible on a computer screen might be unclear on a smartphone display. Thus, our recommendation is to keep it short and simple.
The warm-up statements are part of the main study and their hidden function is to familiarize respondents with the content and teach them how to provide answers, thus, such short training is necessary to obtain reliable results. It is recommended to use 6 such statements and they always appear as the first ones. What is more, you can select statements of your own choice. They should refer to the topic of your main study and be similar to the actual tested values in terms of structure. After this short exercise, the respondent is well prepared and knows what to expect as warm-ups teach respondents how to press buttons properly, so important data is not lost later on. Having this goal in mind you can select any warm-up you want.
We need 6 statements that refer to the topic of your study, they need to have the same structure as the tested statements.
No, these warm-up statements are not tested and no data is recorded for them. They are warm-up words and they are just for training purposes. The data from these warm-ups is likely to be distorted.