The Sabre ES9038 family offers 7 filter presets.
* brick wall
* corrected minimum phase fast
* minimum phase slow
* minimum phase fast
* linear phase slow
* linear phase fast
* apodizing fast
We know from several projects over here, like the Soekris filter brewing project,
that the choice of filters shouldn't be neglected.
Archimago ran a nice analysis analysis on an Oppo device with a 9038Pro inside on his blog.
The frequency responses listed on the blog, matches what you'll find in the X-Sabre Pro DAC manuals.
Archimagos conclusion not to hear any difference between any of the filters
was somehow expected. He still IMO has done a nice job.
I do think his measurements give directions.
I havn't seen anything about the ES9038Q2M that's coming with Ian's HAT,
Allo Katana and Audiophonics iSabre 9038Q2M.
I guess that the data wouldn't differ much if any at all. (Perhaps Allo could contribute some nice AudioPrecison measurements to confirm that.
)
Since there a plenty of fellow golden ear inmates around, I think
we should have a discussion about filter preferences.
Do you actually hear any differences?
If so, mind to share your preferences?
I just started listening to these DACs. I currently have a Katana and an iSabre on my desk.
My first two filters I listened to were brickwall (default) vs. apodizing.
* brickwall sounded slightly edgy/scratchy
* apodizing a bit mellow with rounded edges.
The DACs are still in the burn-in phase though. And I'm still tweaking this or that. That 1st impression might change.
For sure I'll look at the linear filters next. These look best to me on paper.
* brick wall
* corrected minimum phase fast
* minimum phase slow
* minimum phase fast
* linear phase slow
* linear phase fast
* apodizing fast
We know from several projects over here, like the Soekris filter brewing project,
that the choice of filters shouldn't be neglected.
Archimago ran a nice analysis analysis on an Oppo device with a 9038Pro inside on his blog.
The frequency responses listed on the blog, matches what you'll find in the X-Sabre Pro DAC manuals.
Archimagos conclusion not to hear any difference between any of the filters
was somehow expected. He still IMO has done a nice job.
I do think his measurements give directions.
I havn't seen anything about the ES9038Q2M that's coming with Ian's HAT,
Allo Katana and Audiophonics iSabre 9038Q2M.
I guess that the data wouldn't differ much if any at all. (Perhaps Allo could contribute some nice AudioPrecison measurements to confirm that.
)Since there a plenty of fellow golden ear inmates around, I think
we should have a discussion about filter preferences.
Do you actually hear any differences?
If so, mind to share your preferences?
I just started listening to these DACs. I currently have a Katana and an iSabre on my desk.
My first two filters I listened to were brickwall (default) vs. apodizing.
* brickwall sounded slightly edgy/scratchy
* apodizing a bit mellow with rounded edges.
The DACs are still in the burn-in phase though. And I'm still tweaking this or that. That 1st impression might change.
For sure I'll look at the linear filters next. These look best to me on paper.
For my Katana I preferred the minimum phase slow filter, and on enquiring in the "Getting the best...." thread the two responses I got were from inmates with the same preference. I wouldn't like to try and describe the differences between the filters though. I chose the one that somehow felt most comfortable to listen to.
Yep. I 'm currently running the linear-slow too.
It looks quite nice on its impulse response. Just a little pre-/post
ringing.
Of course there's the increased aliasing and the rather early rolloff.
It IMO sounds better to me on my system currently than Brick and Apodize.
Triangles and other percussion stuff get their sparkling sounds back without being overaggressive or annoying.
Back on track.
It looks quite nice on its impulse response. Just a little pre-/post
ringing.
Of course there's the increased aliasing and the rather early rolloff.
It IMO sounds better to me on my system currently than Brick and Apodize.
Triangles and other percussion stuff get their sparkling sounds back without being overaggressive or annoying.
Back on track.
The mixup is probably because I never liked minimum phase filters when running my resampling tests.
And beside that, the Sabre minimum-phase-slow filter shows an awful rolloff performance.
I'm pretty much in line with Archimago on that one. Without listening. I'll do that later.
My resampling setup I've been using on the 51xx DACs for quite some time
are also in line with Archimago preferences.
Non of the Sabre filters seem to deliver what can be done on the PC side with e.g. sox.
To squeeze the last bit out, it perhaps requires custom filters.
I'm wondering if any of the DACs around would support that upload process!?!?
You could also feed high quality 352,8/384k from the computer and bypass the Sabre filters, in OSF Bypass mode.
But that feature would also need to be provided by the DAC firmware and driver.
Not sure if anybody offers that either.
I didn't get that quite right. There was a third response, from someone who preferred the Apodizing filter.
I've made some changes to my Katana supplies since comparing filters, so I'll have another listen.
Always good to revise earlier choices once in a while.
The frequency responses and impulse responses are a very good guideline.
These are most objective.
People need to get their systems stabilized (properly burned-in) before starting
these tests. And that can take a couple of days looping your favorite burn-in track. ( Currently I use Tara Labs CASCADE .)
If you find out that your system sounds best with a filter that shows a rolloff
starting short after 12kHz, you better question your audio chain and not the
filter that runs about linear towards 20kHz.
You'll potentially mask other issues. And you'll potentially tweak on with a
flawed filter in the chain.
I do understand very well that some people report not to hear any differences at all or do prefer different filters.
There are valid reasons. Different systems, different ears, different expectations.
If you look at the Soekris filter brewing you'll realize there'd better be "no taste" involved.
And then there's a question of test tracks.
I prefer orchestral tape-transfer recordings from the 50s. There are some pearls out there. All kind of instruments, massive room info to look for, huge dynamic range and most important all but mixed to death.
You won't find that on a Diana Krall studio CD.
Actually. Nothing new here. Just wanted to bring it up once more.
Ah. I almost forgot! You need to check once in a while that the filter of choice is still activated.
I thought I was running Apodize for 3 days, thinking how damn edgy that thing sounds. I then realized that my setup went back to default "Brick" at a certain point in time.
Now I made sure that this won't happen again. Each time I start my PI I force my own setup towards the DAC.
The frequency responses and impulse responses are a very good guideline.
These are most objective.
People need to get their systems stabilized (properly burned-in) before starting
these tests. And that can take a couple of days looping your favorite burn-in track. ( Currently I use Tara Labs CASCADE .)
If you find out that your system sounds best with a filter that shows a rolloff
starting short after 12kHz, you better question your audio chain and not the
filter that runs about linear towards 20kHz.
You'll potentially mask other issues. And you'll potentially tweak on with a
flawed filter in the chain.
I do understand very well that some people report not to hear any differences at all or do prefer different filters.
There are valid reasons. Different systems, different ears, different expectations.
If you look at the Soekris filter brewing you'll realize there'd better be "no taste" involved.
And then there's a question of test tracks.
I prefer orchestral tape-transfer recordings from the 50s. There are some pearls out there. All kind of instruments, massive room info to look for, huge dynamic range and most important all but mixed to death.
You won't find that on a Diana Krall studio CD.
Actually. Nothing new here. Just wanted to bring it up once more.
Ah. I almost forgot! You need to check once in a while that the filter of choice is still activated.
I thought I was running Apodize for 3 days, thinking how damn edgy that thing sounds. I then realized that my setup went back to default "Brick" at a certain point in time.
Now I made sure that this won't happen again. Each time I start my PI I force my own setup towards the DAC.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.