I did only a sine wave at 128K mp3. Result was bit perfect. It seems that somethings mp3 just doesn't change.
bit perfect is as it ought to be. But the encode/decode algor. is another issue for listening.
even the famous BBC application of the "noise power ratio"/"noise fill test" to audio distortion only gave listening result correlation for "distortion" above ~ -35 dB which is "gross distortion" by high loop gain amplifier standards
http://www.keith-snook.info/Wireles...s-World-1978/A new distortion measurement.pdf
http://www.keith-snook.info/Wireles...s-World-1978/A new distortion measurement.pdf
I do not know, this was just to show that same amplitude spectrum does not mean same waveform (phase spectrum is needed as well for the complete waveform description in frequency domain). But interestingly enough, voltage swing of both waveforms I have shown is quite different (clipping may occur with one wave and not with the second one..).
I don't assume these particular waveforms must sound different. But, as you know, ear is very sensitive to rising transient of the sound wave, especially with regards to (time) difference between 2 ears. So, I would expect that differences in phase spectra of signal transients would be audible as a different subjective spatial recognition.
Phase
Hi Pavel,
This is the exact issue I have been pondering for a long time regarding static measurements vs. timbre.
It seems obvious if the harmonic falls on the slope of a filter the phase would be changed, but usually also the amplitude. However, there are cascaded phase changes in many systems containing many poles which can show phase anomalies yet maintain amplitude levels.
This issue is common in my experience with analog recording systems, where I have seen swept responses flat +-<1dB to 30kHz, but the harmonic to fundamental relationship is altered in playback. The result is a timbre change. This says to me that some aspect of the human hearing mechanism does not merely act as a multi-bank filter setup.
Mathmatically explaining the phenomenon is a bit beyond my pay grade I am afraid, but the effect is real.
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
1st on the internet
Hi Pavel,
This is the exact issue I have been pondering for a long time regarding static measurements vs. timbre.
It seems obvious if the harmonic falls on the slope of a filter the phase would be changed, but usually also the amplitude. However, there are cascaded phase changes in many systems containing many poles which can show phase anomalies yet maintain amplitude levels.
This issue is common in my experience with analog recording systems, where I have seen swept responses flat +-<1dB to 30kHz, but the harmonic to fundamental relationship is altered in playback. The result is a timbre change. This says to me that some aspect of the human hearing mechanism does not merely act as a multi-bank filter setup.
Mathmatically explaining the phenomenon is a bit beyond my pay grade I am afraid, but the effect is real.
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
1st on the internet
amp schematics
Hi,
I have been looking through this Fourm, sounds like an interesting amp, but have not come accross a schematic of the amp.?
please direct me to the schematics if posible.
Ian
Hi,
I have been looking through this Fourm, sounds like an interesting amp, but have not come accross a schematic of the amp.?
please direct me to the schematics if posible.
Ian
Sure, with something as simple as a single sine wave, there isn't much that mp3 can do. But with more complex signals the encode/decode isn't bit perfect. That's what the lossless codecs are for.
Sorry, there is no complete schematic for this PREAMP on this tread. Only partial schematics of different functions. In any case, it is almost impossible to buy and match the parts needed to make this design successfully, today.
Possibly so- unfortunately, once you have government agencies dictating how things should be measured for the purposes of advertising, you're pretty much stuck with that. One of the few positive things I have to say about Stereophile is that they do a first-rate job of measuring, and if you ignore the purple prose from the chimps who do their reviews (and apologies in advance to the noble members of the species Pan troglodytes), the measurements they give do correlate very well with the sound, assuming we're talking about ears-only listening.
Richard, I certainly agree with you about adding 'weighting' to an FFT of the harmonics of a signal passing through an audio device. This is long discussed topic, BUT nobody has ever design this 'weighting' into an instrument, to the best of my knowledge.
"there is no complete schematic for this PREAMP on this tread"
Please put one up, you can.
Considering that most members/readers of Diyaudio are amateurs ,not professionals like you and most contributors to this thread, for us diyers understanding what you are talking about and learning from it, starts with a schematic.
Please put one up, you can.
Considering that most members/readers of Diyaudio are amateurs ,not professionals like you and most contributors to this thread, for us diyers understanding what you are talking about and learning from it, starts with a schematic.
miklos, John quite properly withholds at least some details since, among other things, he still may make these. And there are "amateurs" who just want to copy the design and make it, whether they understand it or not.
But there has been more than ample discussion of the salient features, including schematics of related products showing some of the topological approaches, and including some aspects of the power supplies. The problem is you actually have to read the rather long thread 😀
Scott ans SY
You may both have some personal experience in evaluating Geddes Metric (or/and some earlier relative metrics). I don’t. So I could not answer Scott’s question for whatever higher number of distortion limit he would have set. You understand that.
Anything that can be done to evaluate the “acoustic importance” of a given distortion type and it’s magnitude on any part of the audio chain, is to be welcomed.
Why not to try-this heard but not seen- Geddes Metric?
I’ve asked for some technical assistance:
If you think I am way off, I would like to hear (read) it. 🙂
George
- Status
- Not open for further replies.