Hi All,
This was an interesting read. I enjoy new points of view. It was mentioned in the thread that the "square" wave for 20kHz sampled at 44.1kHz should be shown on the scope for both the NOS and OS varieties. I would like to see this posted and debated. I think the real issue is the fact that 44.1kHz doesn't give you the impulse responce that you really need to record "real" music. This is instantly obvious when comparing a DSD or DVD Audio to a CD. I've even heard some DXD recordings, very nice. If the origanal music is encoded at 44.1kHz, that means the bandwidth has already been limited on the CD, so by assuming that you want to get from one point to the other without any type of filtering, you will not get "harmonics", but you will get the images. This is worse than harmonic distortion since it "creates" tones that were never there in the origanal music piece before it was bandlimited and then recorded. At Fs/2 (22.05kHz), this give an illusion that some of the "harmonics" (1,3,5,7,9....) are truly there since the images are at integer multiples of the Fs/2, thus it looks like a square wave. But remeber this only happens for specific frequencies, not all of them. Heres an example.
22.05kHz sampled at 44.1kHz give you these tones
22.05kHz (fundamental at Fs/2) 1 fundamental
22.05kHz (image at Fs - Fs/2) 1 fundamental
66.15kHz (image at Fs + Fs/2) 3rd "harmonic"
66.15kHz (image at 2*Fs - Fs/2) 3rd "harmonic"
110.25kHz (image at 2*Fs + Fs/2) 5th "harmonic"
110.25kHz (image at 3*Fs - Fs/2) 5th "harmonic"
So as you can see, the images at this exact frequency make what appears to be a non band limited square wave from this data. Now if all we listened to was square waves at 22.05kHz, this would be fine.
BUT.....
20kHz sampled at 44.1kHz gives you these
20kHz (fundamental at 20k)
24.1kHz (image at Fs - 20k)
64.1kHz (image at Fs + 20k)
68.2kHz (image at 2*Fs - 20k)
108.2kHz (image at 2*Fs + 20k)
112.3kHz (image at 3*Fs - 20k)
as you can see, these frequencies are no longer harmonically realted. This is where the NOS methods are mathematically inaccurate. If you don't filter these non-harmonics, you will now have effectivly generated them in the D/A conversion and they would not (most likely) have been in the origanal music. (Think lossy compression
)
Of course it will sound different if you filter these or not, and like eveyone can agree, the sound to the person is whats important. If some "like" the sound of non hormonics added to the signal, thats great... for those, for others, it may not be so great. I wont go into which I prefer since its really not relevant.
Just my take.
Thanks
Dustin
This was an interesting read. I enjoy new points of view. It was mentioned in the thread that the "square" wave for 20kHz sampled at 44.1kHz should be shown on the scope for both the NOS and OS varieties. I would like to see this posted and debated. I think the real issue is the fact that 44.1kHz doesn't give you the impulse responce that you really need to record "real" music. This is instantly obvious when comparing a DSD or DVD Audio to a CD. I've even heard some DXD recordings, very nice. If the origanal music is encoded at 44.1kHz, that means the bandwidth has already been limited on the CD, so by assuming that you want to get from one point to the other without any type of filtering, you will not get "harmonics", but you will get the images. This is worse than harmonic distortion since it "creates" tones that were never there in the origanal music piece before it was bandlimited and then recorded. At Fs/2 (22.05kHz), this give an illusion that some of the "harmonics" (1,3,5,7,9....) are truly there since the images are at integer multiples of the Fs/2, thus it looks like a square wave. But remeber this only happens for specific frequencies, not all of them. Heres an example.
22.05kHz sampled at 44.1kHz give you these tones
22.05kHz (fundamental at Fs/2) 1 fundamental
22.05kHz (image at Fs - Fs/2) 1 fundamental
66.15kHz (image at Fs + Fs/2) 3rd "harmonic"
66.15kHz (image at 2*Fs - Fs/2) 3rd "harmonic"
110.25kHz (image at 2*Fs + Fs/2) 5th "harmonic"
110.25kHz (image at 3*Fs - Fs/2) 5th "harmonic"
So as you can see, the images at this exact frequency make what appears to be a non band limited square wave from this data. Now if all we listened to was square waves at 22.05kHz, this would be fine.
BUT.....
20kHz sampled at 44.1kHz gives you these
20kHz (fundamental at 20k)
24.1kHz (image at Fs - 20k)
64.1kHz (image at Fs + 20k)
68.2kHz (image at 2*Fs - 20k)
108.2kHz (image at 2*Fs + 20k)
112.3kHz (image at 3*Fs - 20k)
as you can see, these frequencies are no longer harmonically realted. This is where the NOS methods are mathematically inaccurate. If you don't filter these non-harmonics, you will now have effectivly generated them in the D/A conversion and they would not (most likely) have been in the origanal music. (Think lossy compression
) Of course it will sound different if you filter these or not, and like eveyone can agree, the sound to the person is whats important. If some "like" the sound of non hormonics added to the signal, thats great... for those, for others, it may not be so great. I wont go into which I prefer since its really not relevant.
Just my take.
Thanks
Dustin
XXHE said:
I don't think it is possible to fairly judge a type of DAC or a particular DAC on the basis of one implementation of it.
I tried a 1543 NOS DAC and even though it had some nice qualities I eventually had to say goodbye to it just too much noise & distortion.
Now I have a Buffalo and in between the two I tried a Stello D100 signature.
From my experience, if you listened to the first implementation of Buffalo this really does not demonstrate the full potential of the sabre DAC - despite the amazing figures. The sound benefits greatly from much quieter power supplies. I just finished a new discrete component op amp o/p stage with extreme quiet supply & excellent components and for me this is giving richer and more vivid tonal textures than I have heard from any DAC. ( including several $10,000 - $20,000 offerings )
I'm not saying that you will prefer what I am listening to more than your own NOS DAC and perhaps I may prefer your DAC if I heard it, who knows, but it might be a better comparison between different approaches of conversion. I think when twisted pair release their Placid Power Supplies this will make a big difference to the buffalo sound ( so long as they are used for all digital & analogue boards )
I would be interested to hear more about the phasure DAC1. Have you described that somewhere already ? Also any other NOS DAC's that you would recommend
thanks for XXHE - I love it
mike
Hi Mike, thank you.
I want to apologize to those who may feel offended my some "nothing nothing" expressions I did and which just got quoted again. Allow me to be more specific;
As I have written a while ago on phasure.com, the ESS/Buffalo is the best I ever heard and it got me right back to my youth when I always was confronted with my fathers stribing for perfect music through loudspeakers, coming mostly from vinyl at that time. I don't think I heard anything in my life more close to it from digital, and in fact I was looking for that "sound" so long. And suddenly it was there from the Buffalo.
However, and I'm just honest here, after a day or so my brain wouldn't cope anymore, and I found that this was because of not being able to recognize all the instruments playing in an orchestra. Somehow this has become a too valuable measure to being able to ignore it in whatever future. So, that annoys.
I don't think anyone without "the other" experience would ever be bothered by it, and actually I never heard of someone who did not like the Buffalo, or the EVAL board for that matter.
In the end it is just oversampling, and let it be personal to have problems with it. Nothing more to say. Not without the real justifications I doen't have either.
Well, the Phasure NOS1 was made for a special purpose (say, mine) and so far nothing more has been described about it than you will have read, Mike.
If I, currently, could recommend other NOS DACs that could do 24/192 - and in the configuration possibilities this one can do, I had bought it myself. But afaik they don't exist.
Btw, it has been quiet around it lately, because I wanted to have it even better, and it will. Give me a few more weeks.
Peter
I want to apologize to those who may feel offended my some "nothing nothing" expressions I did and which just got quoted again. Allow me to be more specific;
As I have written a while ago on phasure.com, the ESS/Buffalo is the best I ever heard and it got me right back to my youth when I always was confronted with my fathers stribing for perfect music through loudspeakers, coming mostly from vinyl at that time. I don't think I heard anything in my life more close to it from digital, and in fact I was looking for that "sound" so long. And suddenly it was there from the Buffalo.
However, and I'm just honest here, after a day or so my brain wouldn't cope anymore, and I found that this was because of not being able to recognize all the instruments playing in an orchestra. Somehow this has become a too valuable measure to being able to ignore it in whatever future. So, that annoys.
I don't think anyone without "the other" experience would ever be bothered by it, and actually I never heard of someone who did not like the Buffalo, or the EVAL board for that matter.
In the end it is just oversampling, and let it be personal to have problems with it. Nothing more to say. Not without the real justifications I doen't have either.
Well, the Phasure NOS1 was made for a special purpose (say, mine
) and so far nothing more has been described about it than you will have read, Mike.If I, currently, could recommend other NOS DACs that could do 24/192 - and in the configuration possibilities this one can do, I had bought it myself. But afaik they don't exist.
Btw, it has been quiet around it lately, because I wanted to have it even better, and it will. Give me a few more weeks.
Peter
Hi peter,
Thx for your reply - but you did not say what exactly you did listen to regarding the buffalo.
I would be really interested to hear this as it will give me a reference point to understand where you are coming from.
did you use the MUX receiver ?
what power supplies for the digital boards ?
what I to V output stage ?
what power supply for output stage ?
Unless you took special measures for all of these you did not hear the Sabre at it's best.
I agree that any old NOS DAC is not good enough - I once auditioned $25,000 audionote NOS DAC and for me it sounded very strange indeed - like all the sound stage was contained inside a large bubble. So my current conclusion is simply that nearly all DAC's don't sound good !
I try not to subscribe to religious type views e.g. valves sound good, solid state sounds bad etc . . . so for now I think I cannot "believe" that NOS is fundamentally better than OS.
I do believe that the implementation of any type of DAC is critical and from what you say I guess you do to.
I also believe that good implementation of a typical delta sigma type OS DAC may require much more care than implementation of a NOS DAC or a multibit DAC.
In other words if a DAC of any kind sounds bad there will be reasons why it sounds bad and if those reasons are addressed ( if they can be addressed ) then it will not sound bad any longer.
Are you really sure OS DACs are fatally floored or is it just that your NOS DAC is very well implemented ?
mike
Thx for your reply - but you did not say what exactly you did listen to regarding the buffalo.
I would be really interested to hear this as it will give me a reference point to understand where you are coming from.
did you use the MUX receiver ?
what power supplies for the digital boards ?
what I to V output stage ?
what power supply for output stage ?
Unless you took special measures for all of these you did not hear the Sabre at it's best.
I agree that any old NOS DAC is not good enough - I once auditioned $25,000 audionote NOS DAC and for me it sounded very strange indeed - like all the sound stage was contained inside a large bubble. So my current conclusion is simply that nearly all DAC's don't sound good !
I try not to subscribe to religious type views e.g. valves sound good, solid state sounds bad etc . . . so for now I think I cannot "believe" that NOS is fundamentally better than OS.
I do believe that the implementation of any type of DAC is critical and from what you say I guess you do to.
I also believe that good implementation of a typical delta sigma type OS DAC may require much more care than implementation of a NOS DAC or a multibit DAC.
In other words if a DAC of any kind sounds bad there will be reasons why it sounds bad and if those reasons are addressed ( if they can be addressed ) then it will not sound bad any longer.
Are you really sure OS DACs are fatally floored or is it just that your NOS DAC is very well implemented ?
mike
mikelm said:
True.
Still, I found some kind of family sound in certain DAC chips that are common to different implementations.
For instance, the Sabre (ess9008) has the greatest amount of details I have ever heard before. The Wolfson 24/96 has a vinyl-like sound. The pcm1704 has the best PRAT.
I'm sure Peter will disclose more about his NOS dac when it will be ready
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.