|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
74.74.162.231
Magazines like Stereophile generate single sample pulse in digital domain, and show the impulse response. The slow roll-off filter shows less pre-ringing than the fast roll-off filter.I am wondering whether this difference is realized in real world music play.
Let us imagine how music CDs are made. Either at the recording step or at the mastering step to 44.1/16 format, the signal goes through fast roll-off filter.
So something like single sample pulse used for Stereophile impulse response test does not exist in music CDs.
I do not know much about AD/DA theory, but I guess that pre-ringing is made by fast roll-off filter, and it does not significantly increase or decrease by going through AD/DA with fast roll-off filters multiple times afterwards. The pre-ringing, once produced, just remains.
P.S.: I just visited Wadia web site. None of their new products claims the usage of Digimaster slow roll-off filter.
Edits: 07/17/15 07/17/15Follow Ups:
You might find the white paper from Ayre linked below to be interesting.
I prefer the minimum phase with slower roll-off "Listen" setting vs the "Measure" setting on my Ayre QB-9 DSD DAC. It is the best sounding DAC that I have owned and I prefer it by a wide margin over the Luxman DA-06 at nearly 2x the price, and a few others I've had.
"I am wondering whether this difference is realized in real world music play."
Nobody will agree on this one.... All things equal, I prefer a "slow rolloff" filter because I think transient behavior is just as important as frequency response.... I also prefer "minimum phase" filtering, because "ringing" is more compatible with "decays" of a music transient than with "attacks"....
If I were to use a custom digital filter, I'd use one with a "Lanczos 2" characteristic for the pre-ringing, "Lanczos 4" for the post-ringing.
Hi Mr. Krieger,
i have a question about filters.
It is quite clear to me that digital filters have an impact on sound quality, and maybe often a negative impact.
I read that SACD allows for the use of analog filters.
I know that you are against upsampling ... but could it be that upsampling the cds to sacd format the benefits of using analog filters could be superior to the drawbacks of the upsampling process ?
Just thinking.
I read also that SACDs tend to sound a little flat and undynamic, but quite natural. For me this should be a priority.
I like natural e not synthesized sounds ... natural to me means real.
Many chips these days i guess upsample cds to sacd ... but i am not sure.
Thanks a lot.
Kind regards,
bg
Edits: 07/21/15
i have a question about filters.
"It is quite clear to me that digital filters have an impact on sound quality, and maybe often a negative impact."
It even depends on the recording as well.....
"I read that SACD allows for the use of analog filters."
Actually any digital medium allows for that.......
"I know that you are against upsampling ... but could it be that upsampling the cds to sacd format the benefits of using analog filters could be superior to the drawbacks of the upsampling process ?"
Upsampling implies the use of digital filters.... They're inseparable. Although analog filters can still be used as "post filters", after the upsampling conversion.
If the CD is upsampled via DSD (which I think is compatible w SACD conversion), it's synchronous.... The upsampling I don't like is "asynchronous" conversion, where the conversion is prone to jitter and noise artifacts becoming embedded into the signal. DSD does not have this problem. (Remember, your resolution for CD playback is still 16/44, no matter how it's upconverted.)
"I read also that SACDs tend to sound a little flat and undynamic, but quite natural. For me this should be a priority."
Whatever works.... I don't like SACDs because they become unbearable after 10 minutes of listening.........
"I like natural e not synthesized sounds ... natural to me means real.
Many chips these days i guess upsample cds to sacd ... but i am not sure."
Unless it's a DSD conversion, I don't think it "upsamples to SACD".... But then again, I personally wouldn't want it "upsampled to SACD".... The end result is the disadvantages of both CD and SACD (CD resolution with extra RFI), with the advantages of neither.......
Hi and thanks a lot Mr. Krieger for the very valuable advice.
I am quite ignorant.
Could it be that " synchronous DSD conversion " of CDs may have some potential for natural and good sound ?
I like sound relaxed but with detail.
I am mostly interested in CDs and maybe 16/48 files. Because i heard nice sounds also from some cd players. Today i am mainly interested in dacs because i am using more various types of streamers (i.e. Squezeebox and the like).
Thanks a lot again.
Kind regards,
bg
Best results will depend on the filtering used in the ADC or downsampling. This can be shown by theory, measurement and listening tests.
If the original recording was apodized and used a minimum phase filter, then it will be possible to use a steep linear phase filter without introducing any preringing (or adding any additional post ringing) or adding additional phase shift. If not, then other filter settings may be needed for best results. As a general rule, however, if one wants to use the same settings for all recordings then it is probably best to use a linear phase filter with slow roll-off (and slightly apodizing). This minimizes the maximum sonic damage, IMO.
I spent a lot of time reaching these conclusions. In the end, I concluded that the 44 kHz formats were inadequate and that it was probably not worth obsessing over this unfortunate fact of history. Filter choices at higher sampling rates are much less signficiant.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
I beleive Ayre's "MP" filter they use in the "listen" switch position is exactly this sort of filter arrangement on their C-5xeMP disc player.
"I spent a lot of time reaching these conclusions. In the end, I concluded that the 44 kHz formats were inadequate and that it was probably not worth obsessing over this unfortunate fact of history. Filter choices at higher sampling rates are much less signficiant."
I agree that it probably isn't worth losing sleep over, too. I own a Rega DAC (was my main DAC but now in back room) - and it allows you to choose filters at the press of a button. Fun stuff to tweak away with 48kHz and below material - and would generate some minor differences. At higher resolution, didn't seem to have very much in the way of impact at all.
The Berkeely Audio DAC-2 has a defaul setting. But it also has several other filter settings, that are really meant for stdio use - since they "simulate" crummy filters that might be in use in some playback devices. I have cycled through them a couple of times and agree with their assessment. (I think their regular DAC's are "one foot in the studio")
============================
As audiophiles, we take what's obsolete, make it beautiful, and keep it forever.
Hey! I have a blog now: http://mancave-stereo.blogspot.com or "like" us at https://www.facebook.com/mancave.stereo
My last DAC the PS Audio PWD had different filters and I thought the filters made only a small difference.
I agree!
I have some limited experience with a Heed Obelisk DA dac which has 3 filter settings.
F1 is 'linear phase soft knee'
F2 is 'linear phase brickwall' (default setting)
F3 is 'minimum phase soft knee'.
It's a very good sounding unit, used with the Heed Obelisk DT transport.
I've used it in my 2nd system (headphones only via a good amp) and in my main system.
Recently I tried changing the filter from F2 to F3. At first I found it hard to hear any difference; but on further listening, I noticed improved tonality and a more natural sound. So I have left it that way, and the Heed now remains in the main system in preference to a Resolution Opus 21, Luxman DA06, and Astin Trew At3500+. I find my Accuphase slightly better but the Heed is now closer with the F3 filter; the Accuphase stays in the headphone system. Both sound gorgeous.
My limited experience suggests different filters can make a meaningful difference. But I'm sure personal taste and system matching are major factors.
Consider the CA DACMagic+ which gives you a choice of 3 filters.
Linear: Minimum: Steep
The steep starts cutting into the audio frequencies just below 20khz but otherwise is good. I use this filter for TV sound and OTHER 32khz sample rate material.
Too much is never enough
My attitude about technology is that for anything more complicated than a basic screwdriver, it is very hard in many cases for an outsider to determine whether something sounds good "because" of a particular technical feature, or "despite" that technical feature.
A certain cable has as a selling point that it is loosely braided. I like the sound of that cable. But is what I like caused by the loose braiding, or is the loose braiding irrelevant?
In general, I prefer "Minimum Phase" digital playback filters, the selling point of which is that they do not pre-ring, at the cost of twice as much after-ringing (all things being equal). But is it the lack of pre-ringing that makes such filters sound better to me, or is there some other aspect of the implementation of that filter that causes me to like the sound?
Every February 2nd I emerge from my burrow and caution audiophiles that the deep-seated need to have a logical explanation for everything is first and formost a deep-seated and largely un-met EMOTIONAL need, which is why "Objectivists" often can be relied upon to carry their search for an answer to irrational extremes.
Not talking about you or anyone on this thread, just a general observation.
Me, my passport says "Aristotlean."
If it sounds good, it is good.
ATB,
JM
But how do you define sounds good then. What does it have to do with fidelity or isn't that relevant(It isn't for any individual but it has to be for a large population). Some 'measurements' are needed(to go along with listening too at the given state of the art) or is there a Tsar of subjective fidelity. This is a conundrum first made clear to me by Percy Wilson's Zanzibar Fallacy.
When you hear hoofbeats, don't think of zebras. Think about an actor on a Foley soundstage clattering two coconut shells on an apple crate.
— John Marks
With a computer audio system and an audio editor with resampling software you can explore these tradeoffs. You will also need a few high resolution audio files that you can convert down to CD quality in different ways. You will be able to look at plots and see ringing and, more importantly, you will be able to relate the filtering settings to what you can hear.
All the various possible filtering settings used to make and play back 44 kHz recordings are compromises of one form or other. There will either be high frequency roll off or harsh aliasing distortion if a slow roll-off filter is chosen (depending on filter offset). If a high roll-off filter is chosen these tradeoffs will be less severe, but then there will be ringing which will smear the transient response. Depending on the filter the ringing can occur before and after the transient or entirely after the transient, but the price will have to be paid one way or the other and there will be a tradeoff between tonality and imaging. The choice of best upsampling filter (as used in playback) will depend on the choice of downsampling filter (as used in recording).
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Of the many settings offered by the Sabre chip, fast/slow rolloff has the least amount of sonic impact imo.
With that said, few manufacturers actually use/or have the filters available for consumers to tweak/play with...
Dynobots Audio
Music is the Bridge between Heaven and Earth - 音楽は天国と地球のかけ橋
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: