No, I think it might be fair play for some of us to post links to the vast literature on signal processing, information theory, etc. Would anyone read it? The analogy to the reading of all the material on psycho-acoustics and perceptual testing seems apt.
Can you summarise these many papers as Jakob & I do? - in most cases we just provide the linked references for those who want to verify or read further
Probably not. If people wanted to study voluminous literature, they probably would have done it already. People come to forums to argue for other reasons. For some people it appears to be a game in which they enjoy fighting as dirty as they can get away with without being banned.
Just about the only test which comes closest to what we hear is a muilti-tone test. The level of masking and other sound characteristics heard can often be seen in a multi-tone test.
Even though the THD is known to be a poorest indicator of sound quality.... reducing it, often reduce the other more audible distortions and their effect. So, in a way, it indirectly is affecting the audible distortion levels and masking.
Most people can do IMD easily and maybe should instead of THD. But multi-tone shows more correlation to audible levels than all others. If the muti-tone was below -100dBv... we would all be happier than if only THD was below -100.
Here is an illustration... your results will vary.....
THx-RNMarsh
Even though the THD is known to be a poorest indicator of sound quality.... reducing it, often reduce the other more audible distortions and their effect. So, in a way, it indirectly is affecting the audible distortion levels and masking.
Most people can do IMD easily and maybe should instead of THD. But multi-tone shows more correlation to audible levels than all others. If the muti-tone was below -100dBv... we would all be happier than if only THD was below -100.
Here is an illustration... your results will vary.....
THx-RNMarsh
Last edited:
What causes those distortions to spike so much between 2k and 5khz?
That is supposedly right in the most sensitive area for human hearing......would there be any benefit in working on decreasing that 2-5k spike?
That is supposedly right in the most sensitive area for human hearing......would there be any benefit in working on decreasing that 2-5k spike?
I think this is a great idea. Back in the neolithic era when we were perfecting mastering for magnetic media, we noticed the sum of energies had a strong effect on the actual tape biasing (of course, a fact which led to Dubly HX) as well as the associated circuitry clipping due to pre and de-emphasis. In order to analyze behavior with a more music-like signal we came up with a multi-tone generator outputting 9 tones. We set them up as a semi-harmonic series so the H2 and H3 of each would not fall on each other or the fundamentals, allowing for individual harmonic analysis. I think the frequencies started at 30 Hz and then increased as (F*2)*1.1.
The other reason to do multi-tone is crest factor which music has but static single tones do not. For this reason we constructed 9 individual oscillators in a Tek TM500 plugin. They were not synchronized and all were set in level +/- 0.1 dB. We recorded the resulting signal, originally on DAT, but then on CDs and used them extensively in equipment testing as well as recorder setting...I wonder if I still have one around....
Another blast from the past...
Howie
Except you bought up the 1970s (when Bruno was in short trousers) which does not appear to have anything to do with what is happening in 2019 in Denmark.
If you read the prelim specs here https://6moons.com/wp-content/uploads/audioreviews/purifi/1.pdf you will see that, although he leads with THD as that is what some people still want to see there is also a number of other more useful tests. No 32 tone test sadly, but no one else publishes that.
There are 3 commercial amplifiers I am aware of that push the envelope this far and two of them are Bruno designs. The 3rd is the AHB2.
And luckily for us there are affordable amps that do this. Just wish I had time to put mine in a box...
Not to disappoint anyone but IME modern well executed designs don't show well in conventional THD and clearly fail multi-tone tests. In fact that $9 DAC dongle had a perfectly clean 32 tone spectra.
Yep, all fit into USB/Lightning and 3.5mm plug body. Repost from ASR, measurement by SY.
This is the standard AP test I assume. The tones have to all be integral multiples of the sampling clock so they lie in exact FFT bins. The crest factor can be adjusted over a wide range but I don't think anyone does this. I did extensive work with multi-tone testing of ADSL drivers with synthesized QUAM signals, there is a lot of room for new work here.
BTW you can even make your asymmetrical waveforms, out of sine waves even who would have thought?
BTW you can even make your asymmetrical waveforms, out of sine waves even who would have thought?
Last edited:
Yes, we could use better ways to measure some of the stuff some skilled listeners may notice.
Went over to Jam's today and nitpicked the sound of his system. He agreed though. Then we reviewed our nitpicking of what we have been doing over at my place with ES9028PRO and AK4499. Hopefully, something good will come out of it eventually for other people who may enjoy some de-nitted sound. Would be great if we had measurements that correlated better with what at least of few of us hear. Probably save some time and effort.
Still have a good memory of Richard's M2 speakers.
Regarding the little dac you mentioned, performance is not so good with 32-ohm headphones attached IIRC.
So the output amp had limited drive, it would probably really suck driving the M2's.
It is not good enough for any of the planar type headphones which are low Z.
Those are the very popular HiFiMan, OPPO, Audeze etal.
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.