We had long discussions about this sort of thing in earlier threads on this forum, I just can't remember which ones. The conversion to bit stream has to involve a substantial sample rate increase, otherwise you hear more noise than music, but the designers may have made a quasi-non-oversampling mode where they try to mimic the signal transfer of a non-oversampling DAC.
That would mean just repeating the same sample instead of interpolating with a normal oversampling filter, and using a noise shaping loop designed for unity signal transfer. In fact ESS has a patent on an asynchronous sample rate converter that keeps on repeating the input sample, except close to input data transitions, US7330138B2, Asynchronous sample rate correction by time domain interpolation.
That would mean just repeating the same sample instead of interpolating with a normal oversampling filter, and using a noise shaping loop designed for unity signal transfer. In fact ESS has a patent on an asynchronous sample rate converter that keeps on repeating the input sample, except close to input data transitions, US7330138B2, Asynchronous sample rate correction by time domain interpolation.
Was that an answer to my question? 
What you mean is called OSF Bypass = OverSampling Filter Bypass. A function that is offered by the Sabre chip.
Neither Ians nor Audiophonics nor Allo nor any other ES9038Q2M DAC I'm aware offers access to this feature
via user interface/controller/driver.
And yes, at 8*FS the internal oversampling """filters""" get bypassed with OSF-B.
BUT. You still need to feed 8*FS. That means you need to upsample your data externally.
And these data still run into an IIR filter after that btw.
Therefore all this has nothing to do with good ole Non-OverSampling (NOS).
OK. Let's say you can live with that.
Then there'll be other side effects that come with OSF-B:
* You can't use software volume control anymore.
* And e.g. Mute wouldn't work anymore, which might cause nasty noises here and then.
Bottom Line.
If
* you don't look for NOS,
* you can also live with these side effects,
* you have a real good resampler and
* you're able to activate OSF-B
give it a try and let us know how it works.
There is a slight chance that externally resampled data @ highest quality
do have a slight advantage over the internal function/filter of the DAC.
Good luck.
What you mean is called OSF Bypass = OverSampling Filter Bypass. A function that is offered by the Sabre chip.
Neither Ians nor Audiophonics nor Allo nor any other ES9038Q2M DAC I'm aware offers access to this feature
via user interface/controller/driver.
And yes, at 8*FS the internal oversampling """filters""" get bypassed with OSF-B.
BUT. You still need to feed 8*FS. That means you need to upsample your data externally.
And these data still run into an IIR filter after that btw.
Therefore all this has nothing to do with good ole Non-OverSampling (NOS).
OK. Let's say you can live with that.
Then there'll be other side effects that come with OSF-B:
* You can't use software volume control anymore.
* And e.g. Mute wouldn't work anymore, which might cause nasty noises here and then.
Bottom Line.
If
* you don't look for NOS,
* you can also live with these side effects,
* you have a real good resampler and
* you're able to activate OSF-B
give it a try and let us know how it works.
There is a slight chance that externally resampled data @ highest quality
do have a slight advantage over the internal function/filter of the DAC.
Good luck.
Last edited:
Was that an answer to my question?
What you mean is called OSF Bypass = OverSampling Filter Bypass. A function that is offered by the Sabre chip.
Neither Ians nor Audiophonics nor Allo nor any other ES9038Q2M DAC I'm aware offers access to this feature
via user interface/controller/driver.
And yes, at 8*FS the internal oversampling """filters""" get bypassed with OSF-B.
BUT. You still need to feed 8*FS. That means you need to upsample your data externally.
And these data still run into an IIR filter after that btw.
Therefore all this has nothing to do with good ole Non-OverSampling (NOS).
OK. Let's say you can live with that.
Then there'll be other side effects that come with OSF-B:
* You can't use software volume control anymore.
* And e.g. Mute wouldn't work anymore, which might cause nasty noises here and then.
Bottom Line.
If
* you don't look for NOS,
* you can also live with these side effects,
* you have a real good resampler and
* you're able to activate OSF-B
give it a try and let us know how it works.
There is a slight chance that externally resampled data @ highest quality
do have a slight advantage over the internal function/filter of the DAC.
Good luck.
For the OP but since you've deigned to pontificate to me, lets go. The OP referred to the ES9038 and not any particular implementation of it and as such your lack of awareness is irrelevant. Bypassing the 8x OS filter is a chip level feature that has been available since the days of the AD1955, CS4397, PCM1738 and the WM8740 etc. The only thing that gets bypassed is the 8x OS filter. All the other features are still available.
As it happens I am aware of a number of commercial designs that bypass the internal OS filter in favour of a propriety filter and at least one diy design that does away with the filter altogether. As to the size of the advantage, I'll leave that to those who buy or build them.
Look.
* The OP referred to NOS and ES9038. NOS on Sabre simply doesn't and I'd guess won't ever exist.
* Name me a single ES9038 device relevant to us DIY geeks that offers OSF Bypass and that would accessible through the driver or controller - because otherwise it'd be useless.
* Bringing up whatever DACs is nonsene - OFF Topic. And even if these OFF Topic DAC chips offer that feature, it doesn't mean that the device manufacturer made it accessible.
* I've been using PCM51xx DACs on my DAC HATS which simply bypass the filters at 352k8/384. So what!?? That wisdom won't help the OP and the subject.
And finally:
Just some excerpts from the ES9038Q2M DS:
You're simply wrong with:
Instead of wining about being "pontificated", which is just ridiculous, you should read what I wrote. IMO a clear and structured response related to the OPs subject.
Enjoy.
* The OP referred to NOS and ES9038. NOS on Sabre simply doesn't and I'd guess won't ever exist.
* Name me a single ES9038 device relevant to us DIY geeks that offers OSF Bypass and that would accessible through the driver or controller - because otherwise it'd be useless.
* Bringing up whatever DACs is nonsene - OFF Topic. And even if these OFF Topic DAC chips offer that feature, it doesn't mean that the device manufacturer made it accessible.
* I've been using PCM51xx DACs on my DAC HATS which simply bypass the filters at 352k8/384. So what!?? That wisdom won't help the OP and the subject.
And finally:
Just some excerpts from the ES9038Q2M DS:
You're simply wrong with:
Instead of wining about being "pontificated", which is just ridiculous, you should read what I wrote. IMO a clear and structured response related to the OPs subject.
Enjoy.
I think we all agree that it is only quasi-NOS: the noise shaper needs to be oversampled, otherwise it can't possibly work, but the interpolation filter can be replaced with something that simply repeats the same sample a couple of times, thereby sort of mimicking the waveforms coming out of a NOS DAC - except for all the ultrasonic quantization noise of the noise shaper. rfbrw has explained how to do this.
If you are a DIY geek, why don't you design and build your own ES9038 application? Or at least modify one until it does whatever you want it to do?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.