Re: NewTimePitch AudioUnit increases volume?
Re: NewTimePitch AudioUnit increases volume?
- Subject: Re: NewTimePitch AudioUnit increases volume?
- From: Bruno Di Giorgi <email@hidden>
- Date: Thu, 06 Sep 2012 15:50:30 +0200
Thanks for your advice.
I did it. But I found DiracLE version is not 64bit compatible, and I'm building for 64-bit.
So I think I cannot use it.
But I will definitely check it better.
Il giorno 06/set/2012, alle ore 15:34, tahome izwah ha scritto:
> Check out Dirac: http://dirac.dspdimension.com. It's faster, offers
> better quality and is a proven solution that doesn't have any quirks.
> We use it in all our products.
>
> --th
>
>
> 2012/9/6 Bruno Di Giorgi <email@hidden>:
>> Thank you Brian for your right answer.
>>
>> I have the classic graph AudioFilePlayer, NewTimePitch and DefaultOutput.
>> Inputs and outputs of these Units have identical streamFormat (and then number of channels).
>>
>> So, it's not related to stream format but as I said you're right! It is definetly related to stereo image.
>> I made two tests:
>> 1. I made mid/side codify in between NewTimePitch and DefaultOutput to hear the difference
>> 2. I compared NewTimePitch and TimePitch
>>
>> The main result is that TimePitch has a lot better quality, and it is mainly due to phase coherency in both channels.
>> In fact: when I switch from mid to side with NewTimePitch, I can't hear any difference (when the unit is working, i.e. not at (rate==1,pitch==0)).
>> When I switch from mid to side with old TimePitch phase is consistent so I hear at (rate!=1,pitch!=0) exactly what I hear at (rate==1,pitch==0).
>>
>> Not hearing difference from mid to side with NewTimePitch means that when listening in stereo channels are summing and I hear that volume increase.
>> This lead me to think that the NewTimePitch computing algorithm is messing with phase somehow.
>>
>>
>> Il giorno 06/set/2012, alle ore 04:38, Brian Willoughby ha scritto:
>>
>>>
>>> On Sep 5, 2012, at 10:04, Bruno Di Giorgi wrote:
>>>> I'm experimenting with NewTimePitch Unit, and having success in doing time and pitch changing in real time.
>>>> However I have a little issue:
>>>>
>>>> When time (kTimePitchParam_Rate) changes from 1 or pitch (kNewTimePitchParam_Pitch) changes from 0 Volume rises of a factor of about 1.3.
>>>> Then comes back to normal when values are back to defaults, as if unit with default values is bypassed then actives (rising volume) when changing values.
>>>> Now I'm compensating in OutputUnit, lowering volume to about 0.7.
>>>> Is this the normal behavior of this unit?
>>>
>>> No. I don't think so, but I've only used the older TimePitch units.
>>>
>>> I guess you have a mono/stereo issue in your graph, such that left+right are being combined in the TimePitch unit. That would explain the 3 dB change in volume. As to how you fix this, I don't know, but you should check your graph to make sure all inputs and outputs match in channel count.
>>>
>>> Brian Willoughby
>>> Sound Consulting
>>>
>>
>>
>> _______________________________________________
>> Do not post admin requests to the list. They will be ignored.
>> Coreaudio-api mailing list (email@hidden)
>> Help/Unsubscribe/Update your Subscription:
>>
>> This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden