際際滷

際際滷Share a Scribd company logo

Letter from CEO

Daniel Ischenko
Daniel Ischenko

This document discusses including automatic speech recognition, text to speech, and voice compression technologies as standard libraries within operating platforms. It argues that including these technologies can help platforms by increasing the number and quality of applications, lowering costs for developers, and attracting more developers to support the platform long-term. The CEO of Speereo Software proposes licensing their speech recognition technologies directly to platform owners.

1 of 2
Confidential document. Partners only. Automatic Speech Recognition, Text To Speech & Voice Compression as parts of operating platforms. Technical and Marketing aspects. Konstantin Lamin CEO Speereo Software 2009 Speech recognition technology will become an inseparable part of human interaction with machines and very soon. Speech recognition will ease such interaction, make it more comfortable. That especially is true if one thinks of mobile platforms, where standard interfaces are not applicable much. It is obvious that any platforms success is based on quantity, quality, novelty and openness of software tuned for that platform. These qualities are attainable only if platform owner has a sound, transparent and reliable technology policy; if financial and marketing needs of a software developer are taken into consideration. They are many platforms available today and developers simply will not support one with uncertain future and low software distribution chances. Such situation forces platform owners to step out of traditional boundaries majority of them are hardware manufacturers and main growth takes place in software market. That is why I believe that any platform success will strongly depend on ability of its top management staff to adapt to new game rules. No new processors, hardware designs, cameras, buttons or materials can influence platform success like developer support. Therefore the time has come for platform owners to support ASR/TTS/VC developers. Lack of clearly defined strategy in this area makes distribution of these technologies difficult, thins down efforts put in development. We believe that the best way to implement speech recognition technology is to include its libraries into a platform. This will allow: To use speech engines in standard platform applications. ASR/TTS/VC libraries will become open to platform developers; To include into a good number of ASR/TTS/VC applications into software preinstalled into mobile device by default; To give access to ASR/TTS/VC to all developers that work within the platform; To free developers working with a platform from document and payment hassle tied to ASR/TTS/VC licensing; To use standard ASR/TTS/VC libraries within all applications, therefore saving RAM and ROM; To make a final move to unified paradigm of voice interfaces in miscellaneous applications. Standard libraries and SDK will give additional support to that; To concentrate all development experience gained, support efforts made and consulting materials produced for ASR/TTS/VC application development in one place instead of dispersing them through many partners; To include into platform cost ASR/TTS/VC modules and list of applications preinstalled. This will allow ASR/TTS/VC developer i.e. Speereo to cut time and costs involved in sales and licensing of ASR/TTS/VC engines to a multitude of third-party developers, i.e. to reduce licensing prices. Lowering of licensing prices will also cut down costs that carry third-party developers in production of innovating software. Therefore the platform will gain many necessary applications and by that will grow in device sales; To include ASR/TTS/VC into a platform SDK. By taking such strategic step platform owner chooses to support one technology partner. That will allow him to plan medium and long term strategy with lower risks, to invest into technology development (additional languages for speech
recognition engine for example). With that the platform owner will develop the platform market, give it a direction to take for the future. Such responsible approach will make platform grow and be attractive to third-party developers for long-term partnership. Next step of interface development will once again show those who were left behind platform owners that did not include ASR/TTS/VC into their platforms. It is not our business to tell platform owners of their strategic development. We are simply pointing out one the possibilities and making ourselves useful. Speereo offers: 0.3-0.5 EUR per license per device ASR/TTS/VC libraries with rights to use in third-party applications in given device; support and training of third-party developers by Speereo included. 0.5-1.2 EUR per package ASR/TTS/VC libraries with rights to use in third-party applications in given device; support and training of third-party developers by Speereo; plus starting application package (launcher, organizer, translator). In case of licensing contract and technological partnership we are ready to take on the responsibility for technology support and development in accordance with platform owners strategy (new devices compatibility for example). Alternative partnerships Investment Application development Sales to end-users model is not that attractive for us because of high development costs (application development in our field is more complex than average). Investment return will be possible in case of covering of several segments of retail market. It is possible, but strategically involves slower technological development and fewer resources drawn. Our main specialty is speech recognition and its use. Investment Licensing to third-party developers model also holds some minuses. With lack of knowledge of ASR/TTS/VC technologies quality third-party developers bring in additional risks. Serious knowledge of the subject requires extensive resources involvement. Third-party developers will not know if certain ASR/TTS/VC supports the platform at all. Moreover development costs will be added to ASR/TTS/VC license price and profitability of such project will become very queasy. It is worth to mention that application developers are not very skilled in ASR/TTS/VC field. Someone must set a trend, and in order to do that much work is to be done. We have already developed basic application package that will get users and developers used to speech recognition technologies. Why then go down the same road twice? It is better to use accumulated experience and move forward. Konstantin Lamin CEO

More Related Content

Letter from CEO

  • 1. Confidential document. Partners only. Automatic Speech Recognition, Text To Speech & Voice Compression as parts of operating platforms. Technical and Marketing aspects. Konstantin Lamin CEO Speereo Software 2009 Speech recognition technology will become an inseparable part of human interaction with machines and very soon. Speech recognition will ease such interaction, make it more comfortable. That especially is true if one thinks of mobile platforms, where standard interfaces are not applicable much. It is obvious that any platforms success is based on quantity, quality, novelty and openness of software tuned for that platform. These qualities are attainable only if platform owner has a sound, transparent and reliable technology policy; if financial and marketing needs of a software developer are taken into consideration. They are many platforms available today and developers simply will not support one with uncertain future and low software distribution chances. Such situation forces platform owners to step out of traditional boundaries majority of them are hardware manufacturers and main growth takes place in software market. That is why I believe that any platform success will strongly depend on ability of its top management staff to adapt to new game rules. No new processors, hardware designs, cameras, buttons or materials can influence platform success like developer support. Therefore the time has come for platform owners to support ASR/TTS/VC developers. Lack of clearly defined strategy in this area makes distribution of these technologies difficult, thins down efforts put in development. We believe that the best way to implement speech recognition technology is to include its libraries into a platform. This will allow: To use speech engines in standard platform applications. ASR/TTS/VC libraries will become open to platform developers; To include into a good number of ASR/TTS/VC applications into software preinstalled into mobile device by default; To give access to ASR/TTS/VC to all developers that work within the platform; To free developers working with a platform from document and payment hassle tied to ASR/TTS/VC licensing; To use standard ASR/TTS/VC libraries within all applications, therefore saving RAM and ROM; To make a final move to unified paradigm of voice interfaces in miscellaneous applications. Standard libraries and SDK will give additional support to that; To concentrate all development experience gained, support efforts made and consulting materials produced for ASR/TTS/VC application development in one place instead of dispersing them through many partners; To include into platform cost ASR/TTS/VC modules and list of applications preinstalled. This will allow ASR/TTS/VC developer i.e. Speereo to cut time and costs involved in sales and licensing of ASR/TTS/VC engines to a multitude of third-party developers, i.e. to reduce licensing prices. Lowering of licensing prices will also cut down costs that carry third-party developers in production of innovating software. Therefore the platform will gain many necessary applications and by that will grow in device sales; To include ASR/TTS/VC into a platform SDK. By taking such strategic step platform owner chooses to support one technology partner. That will allow him to plan medium and long term strategy with lower risks, to invest into technology development (additional languages for speech
  • 2. recognition engine for example). With that the platform owner will develop the platform market, give it a direction to take for the future. Such responsible approach will make platform grow and be attractive to third-party developers for long-term partnership. Next step of interface development will once again show those who were left behind platform owners that did not include ASR/TTS/VC into their platforms. It is not our business to tell platform owners of their strategic development. We are simply pointing out one the possibilities and making ourselves useful. Speereo offers: 0.3-0.5 EUR per license per device ASR/TTS/VC libraries with rights to use in third-party applications in given device; support and training of third-party developers by Speereo included. 0.5-1.2 EUR per package ASR/TTS/VC libraries with rights to use in third-party applications in given device; support and training of third-party developers by Speereo; plus starting application package (launcher, organizer, translator). In case of licensing contract and technological partnership we are ready to take on the responsibility for technology support and development in accordance with platform owners strategy (new devices compatibility for example). Alternative partnerships Investment Application development Sales to end-users model is not that attractive for us because of high development costs (application development in our field is more complex than average). Investment return will be possible in case of covering of several segments of retail market. It is possible, but strategically involves slower technological development and fewer resources drawn. Our main specialty is speech recognition and its use. Investment Licensing to third-party developers model also holds some minuses. With lack of knowledge of ASR/TTS/VC technologies quality third-party developers bring in additional risks. Serious knowledge of the subject requires extensive resources involvement. Third-party developers will not know if certain ASR/TTS/VC supports the platform at all. Moreover development costs will be added to ASR/TTS/VC license price and profitability of such project will become very queasy. It is worth to mention that application developers are not very skilled in ASR/TTS/VC field. Someone must set a trend, and in order to do that much work is to be done. We have already developed basic application package that will get users and developers used to speech recognition technologies. Why then go down the same road twice? It is better to use accumulated experience and move forward. Konstantin Lamin CEO