Why would that be? With an attenuator at the input the preamp generated noise is always the same irrespective of the volume position. The s/n ratio is the best at full volume.
I wasn't suggesting anything, only pointing out that people look at this from different perspectives. A lot of people swear by vinyl when it's measurably worse on many fronts than hi-res digital audio. Or what about THD+N of a tube amp versus a solid state amp? I'm not one to judge.
Last edited:
I think this is more about that All pass filter corrections doesn’t match with rest of the output equipment. I have similar problems, my amplifier has thermal phase drift and after heat up high frequency phase roll-off and magic disappears. Today I changed to tubes and I was surprised how dull sound is. I have to match new filter for that change. People call that synergy but in fact it can be compensated by filters. I prepared set of 4 filters that have slight different corrections. For me 780 with F5 works with tubes but you can try to find your own matching. I should help with perceived dynamic apart volume.
I'm no sure how popular a set of filters tuned for a specific unique tube amplifier will be...
//
//
Well, not sure either...
I would love to play with a set of NOS filter but without loosing too much gain from those in 1698...in that pack I found the NOS filter to be the most clear, dynamic and lifelike...but I can see as well why some people would call the a bit rough or distorted some how.
So, If they can be a bit cleaner without looser the original value proposition about life-like...perfect...I listen from 44k to 192k...PCM
I would love to play with a set of NOS filter but without loosing too much gain from those in 1698...in that pack I found the NOS filter to be the most clear, dynamic and lifelike...but I can see as well why some people would call the a bit rough or distorted some how.
So, If they can be a bit cleaner without looser the original value proposition about life-like...perfect...I listen from 44k to 192k...PCM
So try one from #2135
F7 is improved NOS stack to 352.8/384kHz, F4, F5 are evolution of NewNOS.
Loosing gain can’t be avoided.
F7 is improved NOS stack to 352.8/384kHz, F4, F5 are evolution of NewNOS.
Loosing gain can’t be avoided.
Having listened to F7 for -what?- over a week, I'm going to evaluate F5 for a longer period of time now. gumisb are you still tweaking or are you about satisfied? 🙂
Still tweaking and discovering new things. As I mention above I have to move from solid state to tubes and all my super duper setup was ruined and I had to start over again. But for now only allpass filter is what I’m changing. All the rest is ok for now. I’m interested your comparison of both F5 vs F7 filters.
Well, because my tubes have high amplification and generate a slight hum, which is intensified by increasing the volume.
I'll pretend I haven't read it.
I don't really want to write about it in this thread 🙂
... as well as with the highest THD from the DAC. Søren published some measurements of the dac1421 on ASR at -1 dBFS and commented how this works out at lower levels with sign magnitude ladders: Review and Measurements of Soekris dac1421 Multibit DAC | Page 5 | Audio Science Review (ASR) Forum
The fact that this audio engineering thing is all about trade-offs is exactly why I love the heck out of it. Raise line level --> DNR goes up by "x" depending on the interconnection resistance, leakage to ground and EM susceptibility. At the same DAC THD goes up by "y" (not just the dam but most any DAC). I don't have the gear to measure but there's an optimum here somewhere.
Meanwhile I too run at 0 dBFS and attenuate with the preamp. I also ran with -10 dBFS for a while and less attenuation, but heard no difference.
First evening's subjective report: it seems brighter, smoother but perhaps also thinner. Which I don't understand yet as they magnitude response is the same according to your graphs.
All pass filter was always there to correct phase deformation of IIR LPF.
There is very small difference in phase between this two filters which influence on perceived sound not to much on FR mag . Second thing it’s upsampling filter not “NOS” so all coefficient play role in coloring sound. Maybe pleasant.
Attachments
gumisb, May I ask a question ?
I use indeed as well tube equipment with the Soekris Dam1941...the build in feature of this is due to the multiple usage of interstage transformers and the OPT of the power amp it self a "bandwidth"-filter: You have on the LF side the inductance of the Transformer and at the high side the capacitance...so typical frequency response of 10hz to 40khz is quiet normal in such an equipment, especially when there are 2 or 3 transformer in the chain.
So my question is: When you use this assumption designing a filter: Could the filter become even more simpler / more direct/ more gain ? If this could be the case, I would strongly wish for a tube-version of your filters...
(I am feeding directly the grid of an 801A DHT which is than playing into a Interstage Transformer, so like a Mini-Power-Tubeamp...as the DAC-Outputstage).
I use indeed as well tube equipment with the Soekris Dam1941...the build in feature of this is due to the multiple usage of interstage transformers and the OPT of the power amp it self a "bandwidth"-filter: You have on the LF side the inductance of the Transformer and at the high side the capacitance...so typical frequency response of 10hz to 40khz is quiet normal in such an equipment, especially when there are 2 or 3 transformer in the chain.
So my question is: When you use this assumption designing a filter: Could the filter become even more simpler / more direct/ more gain ? If this could be the case, I would strongly wish for a tube-version of your filters...
(I am feeding directly the grid of an 801A DHT which is than playing into a Interstage Transformer, so like a Mini-Power-Tubeamp...as the DAC-Outputstage).
My take on this is that it wont help suppressing aliasing/mirrors as a f0 at 40k is way to high - in the DA process, this has no meaning other than it will act as a HF filter for the NOS noise.
For 44,1... now for 96 it starts to help a little - but if one want to follow the sampling theorem, it is still way to little.
//
For 44,1... now for 96 it starts to help a little - but if one want to follow the sampling theorem, it is still way to little.
//
Main function of filters are anti-aliasing/anti-imaging and up sampling/reconstruction.
Up sampling filters are already maximally simplified.
Anti aliasing with IIR filter is in fact quite simple and effective but introduce phase roll-off and group delay. As you get closer to cut-off frequency, so still there in your setup.
That has to be corrected with IIR all pass filter and allow to go a lite bit further and make correction of output stage.
It’s like additional analog element in your chain that has to be matched with others.
Every element of output stage has influence on final effect. Your output stage is complex compared to main direct output so fitting something there could be challenging.
So for now there is set of 4 filters in post #2163 with 4 different allpass filters and you can try to find better fit than standard one from post#2135 offer.
I made 3 rough filters sets close to the target but with more relaxed correction. Let me know is there any improvement vs standard.