How to improve Deep Vocal

I think that if you take these options into account you can improve deepvocal to a point where it is better than utau and maybe it competes with synth V and / or vocaloid:
1.-Ignore if the consonant is voiced or not, this would help to create individual consonants and for the voicebanks to include more vowel samples, as an example would be to add a growl or a raspy voice in the low tones.
2.-Treat the coda consonants as vowels, this to improve the CC connections (mostly used in English) and in turn would benefit the ability of DV to reproduce CVC syllables.
3.-Add the conjugation capacity, it is more than anything for allophones, for example in Spanish before a word with k or g, the n is turned into N.
4.-Let the user decide the order of the tones.
5.-Allow the individual use of phonemes, DV saves the position, duration and range of consonants and vowels, therefore if it is edited so that it is cut in a certain way, or failing that, the register is made by vowels and consonants separate would allow DV to handle individual phonemes as vocaloid plus it would give way to import VCV and VCCV libraries.
6._Keep the vsqx data, This is more useful for dictionaries because if the phoneme symbols are kept when importing a VSQX, it will allow standardizing the symbols of the languages ​​and therefore it would be easier to use them for new people.