Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter

[1audio]

Video can survive some extra delay in the audio. I think the THX standard is something like 15 mS although some systems are closer to 50 mS and are tolerated. Good TV's also have adjustments. Interchannel delay is never good. . .

[qusp]

but neither of those numbers is close to achievable with this design, either find a way to delay the video, or forget about using fifo in your home theatre

[iancanada]

Originally Posted by cracra

Hi Ian,

Would it be possible to minimize automatically the FIFO Latency (i.e. Minimize the number of PCM samples stored within the FIFO Memory) according to the sampling rate of the incoming Digital audio stream through I2S ?

The reason of my question is that I am using my DAC equipped with the FIFO to convert the audio stream of the movies I watch with my Blu Ray player. Currently the FIFO introduces roughly 2 seconds of latency between the image that is displayed and the associated movie dialogue or sound.

Thanks,

That's very good point. Stereophiles usually don't care much about dealy around 1 second. They are more emphasized on sound quality. But home theatre application does.

Developing a variable depth FIFO algorithm with supporting by a variable frequency low jitter clock generator such as Si570 is one of the possible solutios. But the design would become more complicated. Buffering video at same time would be easier but increasing the cost. Lips control might be alternative answer.

Ian

[iancanada]

Improvements

1. Potato GHz TTL flip-flops optimized for finial re-clock applications up to 100MHz.

2. Optional connector to introduce battery or other low noise power supply with on board LDOs bypassed.

3. Additional I2S output port with optional u.fl socket footprints at bottom side for dual mono applications.

4. Capable work with passive battery management adapter.

5. Capable interface with FIFO board via digital isolator board.

6. Providing a new serial port to communicate with potential external or third party processor sending information of MCLK frequency, *Fs and etc.


Ian

[qusp]

looking good mate!

[hochopeper]

Hi Ian,

The Si570 board progress looks very promising indeed! I have two questions though and I'm not sure if either or both have been covered previously.


1. The hirose datasheet suggests removing ground plane from beneath the signal portion of the u.fl receptacle (see attached extract from hirose datasheet), it looks like your boards have the gnd continuing through this area? Is this a concern? I am far from knowledgable on these things just asking a honest question here.

Originally Posted by iancanada

6. Providing a new serial port to communicate with potential external or third party processor sending information of MCLK frequency, *Fs and etc.

2. Is it possible with the Si570 that the output stream may select a different *MCLK based on the input Fs? I have a DAC (ES9023) that I would like to use Fs*512 for Fs<=96kHz and Fs*256 for Fs>96kHz. Is this a possible configuration option via serial port? I will have either a microcontroller or an arm based linux board in the DAC enclosure so will be able to leave either of those permanently connected to the FIFO kit for 3rd party control if necessary.

[qusp]

thats true it does prefer to have the ground eroded underneath to reduce capacitance, the ackodac boards have ground removed from this area. good catch mate, I hadnt noticed that before on the fifo boards. i'm sure its not a terminal error, particularly given the success of the boards, but definitely a detail worth taking care of

[nattonrice]

The stock landing I originally had didn't have a restrict either.
I can't remember if I noticed it when doing your board of after *meek face*.

[marce]

The signalpart of the connector goes underneaath through that areato the central pin of the actualconnector, if the solder resist is a bit flakey it can short out. I would reccomend adding a keep out area to the actual footprint if your CAD package allows it, as it is one of those things you'll forget with this device otherwise, next time you use it! I know I have done in the past

[qusp]

aha so its not about messing with the bandwidth due to capacitance, but rather a safety measure? I presumed the former and thought I remembered acko saying something of the sort

Originally Posted by nattonrice

The stock landing I originally had didn't have a restrict either.
I can't remember if I noticed it when doing your board of after *meek face*.

restrict fail =) nvm dude I didnt notice it till afterwards either, if its not about capacitance then its nothing for a short run hand soldered board, if it is, i'm sure its not such a big deal since its being used for spdif on my board rather than i2s and its all reclocked at the dac in the portable.