Wilbraham predates Gibbs! Who knew?
http://www.math.umn.edu/~garrett/m/mfms/notes_2013-14/04a_Wilbraham-Gibbs.pdf
Like many things in which the more famous person's exposition gets his or her name attached.
http://www.math.umn.edu/~garrett/m/mfms/notes_2013-14/04a_Wilbraham-Gibbs.pdf
Like many things in which the more famous person's exposition gets his or her name attached.
George, the OFFICIAL calender in Serbia is midnight of December 31 for the old year to be changed to new, always has been since 1918. So, rest assured on that night there will be a lot drunk and singing Serbs, I would guess like 95+% of the population. But traditions die hard, so we'll repeat it all again some two weeks later. Some true enthusisasts will have a third go in late January, when it's the Chinese new year.
Just for the sake of good international relations, you understand. Peace, brother!
Just for the sake of good international relations, you understand. Peace, brother!This deserves an answere..... A partial one, at least ..... I was only changing the rise time of the generator transient edge. There was no ringing on the generators' square wave when changing the rise time (Tr). The faster the Tr, the more "Gibbs" ringing showed on output.
Since the internal computations/math is complex, i would not venture to guess what slope is best for everyone. Best done by the mfr for their machine.
And, some/many do it but does it have low enough GD?
As far as I am concerned, it must be Bessel or similar for best GD and considering the processes own filter and what it does... maybe correction as well to get GD low enough.
I started looking into GD ever since Dean Jensen told me he designed all his transformers for flat GD response characteristics.... it was the key to its good sound over others.
I can tell that because he isnt with us any more. It was one small step for mankind. But what do the digital guys do when they filter/process? Did anyone tell them to make flat and very low GD filters? I don't think so. And, if they couldnt make a filter that was not audible, better to have little or no filtering and let go wide open?? Well, what ever one decides, you have no excuse today for not being able to make an audibly transparent LPF filter. Right?
THx-RNMarsh
Last edited:
Thank You
Yes, Shorter rise time means increasing the bandwidth (more HF content) of the signal.
The signal generator has no LP filter at it’s output, nor the input of your measuring instruments, so they don’t show any waviness.
But when that richer in HF content SQR wave signal is to be reproduced by the legitimate bandlimited CD player it will display more waviness compared to the less steep SQW wave
Well Richard
I test myself for GD issues at low frequencies for a long time now and I find I am not sensitive at all to the effect.
Recently I visited this site:
blogohl: Phase audibility
Doing the tests not only with the signal generator but processing music files (.wav) through it, I confirmed my deafness to GD effect at mid to high frequencies as well.
So, I am not the appropriate person to comment on this.
I can’t be that optimistic as you on this, even if my hearing was better than it is.
But as you know, less GD means less steep filter slope, so no free lunch.
George
I think I found and explained a good way to find the best compromise for filter cut-off with a given recorder/player.
I do not know how many people out there can measure GD or design GD correction filters. But the results i hear is promising and worthwhile.
A DIY variable group delay equalizer would be nice to see. I used such equipment with the military to condition data lines. Wish there were some consumer PRO versions today.
THx-Richard
I do not know how many people out there can measure GD or design GD correction filters. But the results i hear is promising and worthwhile.
A DIY variable group delay equalizer would be nice to see. I used such equipment with the military to condition data lines. Wish there were some consumer PRO versions today.
THx-Richard
Last edited:
This would be a good starting point for Group-Delay EQ.
IIR Filter Design Given a Prescribed Group Delay - MATLAB & Simulink Example
THx-RNMarsh
Last edited:
1968? It was a very good year! That year I independently invented the bipolar complementary differential input stage, discovered what bipolars could make 0.4nV/rt Hz when used in parallel, and even first met Mark Levinson. This was all when I was working at Ampex in the Audio Division designing circuitry for master recorders, both present and future.
Oh, to have so many more successful and interesting years! '-)
Oh, to have so many more successful and interesting years! '-)
Maybe, just maybe, you should consider being a little bit more careful about what you write. On an international forum things that can be misread will be. But I gave you a chance to demean me as well as Richard so I guess by your Modus Operandi it's a good evening.
Does-it applies to an analog signal (out of a DAC), with analog filter ?
;-)
Last edited:
They look OK to be honest how do you fake a part with such a complex functionality? Fakes would just do nothing dropped into a circuit.
Yes, it is Christmas and the beatles are now available to listen in streaming.
Remastered, I dicovered a lot of details I never heard before (HF filtered or not).
A lot of details, but, Oh, Lord, the magic is gone. Or my youthfulness with my old transistor radio (Germanium) ?
Happy Christmas to all of you, my friends...
Remastered, I dicovered a lot of details I never heard before (HF filtered or not).
A lot of details, but, Oh, Lord, the magic is gone. Or my youthfulness with my old transistor radio (Germanium) ?
Happy Christmas to all of you, my friends...
Attachments
Last edited:
When you are dealing with a situation like I had..... a less than ideal filtering situation, besides the HF removal, you can measure GD and with that info, start a correction process.
I am re-listening with the added filter... WoW!. That flat dull sound is gone and much more life-like and clear/focused. I think I am going to listen to music all day today.
I am so impressed with the result, I am not taking it out and have another ML-9600 coming to keep for R&D and more tests. I also have an OPPO 105D coming..... will check it as well.
THx-RNMarsh
Last edited:
You just over print on anything with the right number of pins. Then if the customer complains offer a full refund if they return the part. As return shipping is more than the part cost, they rarely have to make good.
So I was pleased they actually shipped a real unit. Of course they don't mention it is a pull.
Now it wouldn't qualify for any certified repair, but it is fine for a 25 year old mixer repair.
Yes, because the Gibbs Phenomena is due to the mathematics that is common to all brick wall filters, whether analog or digital.
The last CD Player or DAC that I know of that had analog brick wall filters was the CDP 101 (1983). Some editions of its service manual had schematics of the brick wall filters, which looked something like this:
Only with a lot more stages of filtering - I think about 20! Even with all that complexity and expense, its performance was relatively poor compared to the digital filters that followed it.
You can find out more information about brick wall filters here:
http://www.analog.com/library/analogDialogue/archives/43-09/EDCh%208%20filter.pdf?doc=ADA4666-2.pdf
- Status
- Not open for further replies.