Thank you for answer!
You probably think about complex bandwidth method for ssb. 
I have already designed ssb with this method and results is much better.
I would like to get the same results with hilbert.
But I think that hilbert and delay elements used for compensating HT delay
are not so accurate. Therefore my question is there any method which enable
reduction of amplitude and phase errors of HT. 
This method with HT seems to be easier for implementing in the FPGA, what
is my goal. 
Amir     

>
>"pita" <amir_hasanovic@yahoo.com> wrote in message
>news:RsmdneIOVb8cah3bnZ2dnUVZ_rKvnZ2d@giganews.com...
>> Hi all,
>> I have designed ssb modulator with HT in Matlab Simulink.
>> The unwanted components (lower or higher) has  attenuation 35dB.
>> My question is  can I get better results with this SSB method.
>>
>> Thank you in advance,
>> Amir
>>
>>
>
>You might do better with:
>
>1. mix desired sideband to baseband (complex mix)
>2. apply baseband filter
>3. mix sideband back up and take the real part
>
>This has the disadvantage that you can't do both sidebands
simultaneously
>but it's generally easier to get the filter response you want. However,
>since the HT is just a filter too it's going to come down to how many
taps
>you need to get the best seperation between sidebands.
>
>BTW, for audio 35dB seperation probably is okay.
>
>-Clark
>
>
>

"pita" <amir_hasanovic@yahoo.com> wrote in message
news:RsmdneIOVb8cah3bnZ2dnUVZ_rKvnZ2d@giganews.com...
> Hi all,
> I have designed ssb modulator with HT in Matlab Simulink.
> The unwanted components (lower or higher) has  attenuation 35dB.
> My question is  can I get better results with this SSB method.
>
> Thank you in advance,
> Amir
>
>

You might do better with:

1. mix desired sideband to baseband (complex mix)
2. apply baseband filter
3. mix sideband back up and take the real part

This has the disadvantage that you can't do both sidebands simultaneously
but it's generally easier to get the filter response you want. However,
since the HT is just a filter too it's going to come down to how many taps
you need to get the best seperation between sidebands.

BTW, for audio 35dB seperation probably is okay.

-Clark

Hi all, 
I have designed ssb modulator with HT in Matlab Simulink. 
The unwanted components (lower or higher) has  attenuation 35dB. 
My question is  can I get better results with this SSB method.
 
Thank you in advance,
Amir