Preamp difference : if it's not the frequency, not the slew rate, and not the harmonics, what is it ?

[Mihobro]

Capacitors and transformers have some very interesting frequency dependent characteristics especially at the high and low extremes. Phase shifting, harmonic reactance and other distortion too. And since even a plain wire can have capacitance and reactive elements, we may not be able to measure it but it is still going to be there. We used to spend a lot of time overdriving circuits with sine waves at different frequencies, and also response to square waves to measure slewing rates. The hope is to exceed the ability of the hearer, which varies a lot also according to ear structure, and ability of the brain to process audio etc. We are fortunate to live in a time where components have improved a lot,. For example, new "music" caps are so much better than the ones from days past. I have found recapping old tuners from the 70's with modern caps, can be like night and day. Also the oversizing of caps, in signal pass circuits an octave below the originals can help to really open up the bottom end.
On another note: RF filtering requirements vary from location to location. I lived down the street from the Empire State Building's transmitters and had all kind of RF bleed into a Neve console from stock Neumann, Audio Technica and to a lesser degree AKG mics. oddly enough and counterintuitively, they responded to the removal of the small silver and ceramic caps across each of the legs of the mic outputs and the ground, what I guess is a RF shunt in theory. sometimes filters setup a better antenna. Of course, just down the block and in the studios that had chickewire faraday shields built into the walls, these mics had no problem with RF. As far as preamps are concerned, most mixers being being manufacturered today use similar two stage discrete fet design and sound better than most of the famous console mic preamps with the exception of Avalons, Jensens and a few other legends, largely because of improved components and computer aided designs. If you are not ready to spend a fortune though, the latest preamps like Mackie VLZ4 and others, may work for you. Keep on trying new stuff, We all may eventually get to recording something meaningful..eventually.

 

[Newmarket]

Capacitors and transformers have some very interesting frequency dependent characteristics especially at the high and low extremes. Phase shifting, harmonic reactance and other distortion too. And since even a plain wire can have capacitance and reactive elements, we may not be able to measure it but it is still going to be there.

If it's there it is measurable. You may not have the facility to do that yourself but it can be measured.

 

[JohnRoberts]

Back when I was designing gear I could measure deviations from ideal that I could not hear. I could pretty much measure everything that I could hear, but sometimes it involved rolling my own custom test set ups.

JR

 

[rh001]

I respectfully disagree

This is not reductionist, but about statistics - the core of all psychological research, which I worked with for a long time when I was younger

The whole point is that if there is such a thing as "better", then the minimum requirement of proof would be that an average of humans would subjectively qualify it as "better" in a double-blind test. We are not at all interested in sameness, as we are not comparing absolutes, but relative measures - and we don't care about individual preference, but the grand average. We even have the possibility to come up with an effect size, if our sample size is big enough.

So IMO, for any claim of goodsoundingness there is a relevant double-blind test to be done. Complicated and expensive, yes, but it's absolutely possible in this way to extract "hard data" from any subjective experience. If you want to. Hifi don't.

/Jakob E.

Hi!

The error in your thinking is that you assume that average is somehow not yucky. I understand that appeal to average is the correct approach for dunkin donuts, doritos and such but in so many other cases average is clearly not good, such as with BMI for instance. One who has an average BMI in USA or Mexico is fat enough to have metabolic illnesses even before age of 40 and definately after 50. Average BMI in Japan is not a health risk.

Neither Sungha Jung nor Yuzuru Hanyu have an interest in average. Sungha Jung has told the story of how when he was a little kid he initially started playing a junky old guitar his father had left over from his youth. He soon realized it was junk and asked his daddy to buy him a new guitar. His daddy bought him an "average" guitar, and it played so badly he (Sungha) cried inconsolably...and daddy ponied up for a decent instrument.

Above average Sungha Jung...62 mil views on this one:

Yuzuru Hanyu has skated his way to net worth around 30 million at age 28...more than the average competitive or ice show figure skater ever by far. Only skater with his own rink. In this video he and pianist Shinya Kiyozuka discuss arranging for figure skating ... how does a mere figure skater get a talent like Kiyozuka to give him the time of day?

Kiyozuka arranged, played and recorded about 6 pieces for Hanyu's competitions. I think every other figure skater ever only had existing pre-recorded music to choose from.


I don't think George Massenburg consulted the average in the design of his highly lauded pro audio electronics.

Ampex stuff was not designed with appeal to average in mind....built more or less to military standards.

Average is just so yucky.

 

[JohnRoberts]

Yucky is not an objective technical term.

JR

 

[Gold]

This sums up the debate pretty well and shows why nothing is ever good enough for anyone no matter what their position is. First of all a properly constructed A/B/X test is far beyond the capabilities of your average self employed audio engineer. Insisting that only an A/B/X test can produce meaningful results is setting an impossibly high bar. The criteria can never be met.

Even if you manage to clear that hurdle then the above "we're not looking for average listeners" argument comes in. So even a properly constructed A/B/X test will not satisfy some.

 

[RoadrunnerOZ]

Hi!

The error in your thinking is that you assume that average is somehow not yucky. I understand that appeal to average is the correct approach for dunkin donuts, doritos and such but in so many other cases average is clearly not good, such as with BMI for instance. One who has an average BMI in USA or Mexico is fat enough to have metabolic illnesses even before age of 40 and definately after 50. Average BMI in Japan is not a health risk.

Neither Sungha Jung nor Yuzuru Hanyu have an interest in average. Sungha Jung has told the story of how when he was a little kid he initially started playing a junky old guitar his father had left over from his youth. He soon realized it was junk and asked his daddy to buy him a new guitar. His daddy bought him an "average" guitar, and it played so badly he (Sungha) cried inconsolably...and daddy ponied up for a decent instrument.

Above average Sungha Jung...62 mil views on this one:

Yuzuru Hanyu has skated his way to net worth around 30 million at age 28...more than the average competitive or ice show figure skater ever by far. Only skater with his own rink. In this video he and pianist Shinya Kiyozuka discuss arranging for figure skating ... how does a mere figure skater get a talent like Kiyozuka to give him the time of day?

Kiyozuka arranged, played and recorded about 6 pieces for Hanyu's competitions. I think every other figure skater ever only had existing pre-recorded music to choose from.


I don't think George Massenburg consulted the average in the design of his highly lauded pro audio electronics.

Ampex stuff was not designed with appeal to average in mind....built more or less to military standards.

Average is just so yucky.

I think you’re pursuing the word average in the wrong context. Testing to determine if a claimed difference exists between high end pieces of audio gear by using an audience of listeners who may be from anywhere or may be a group of professional music industry people is a way to determine that if more than a certain number of people can hear a difference then said difference can indeed be claimed to exist. This is not about the difference between an “average” guitar and a “brand” guitar, but about the difference in sound between pieces of audio gear when it may be contestable if such a difference exists - the two scenarios are not really relevant to each other. An average guitar in the hands of the right player can sound amazing - that’s about talent.

 

[Migs 31]

Wow! some serious trolling/dick measuring contest going on here today...

Concerning OP. @JohnRoberts dixit (no pun intended)
"This is a very mature topic studied by design engineers for decades, including serious psychoacoustic studies and application of known phenomenon."

Time to meditate on that...
M

 

[JohnRoberts]

Wow! some serious trolling/dick measuring contest going on here today...

Concerning OP. @JohnRoberts dixit (no pun intended)
"This is a very mature topic studied by design engineers for decades, including serious psychoacoustic studies and application of known phenomenon."

Time to meditate on that...
M

I studied two years of Latin back in high school, but that was 60 years ago so I had to look that word up.

JR

 

[jeremynothman]

The DAW’s I tested were Cubase, ProTools, Logic Pro, Digital Performer and IIRC, Ableton. There was no dither, no samplerate or bitrate conversion - just a 44.1/16bit ripped straight from a commercially produced CD into each DAW, all faders at nominal 0dB so no volume changes, then a straight export or bounce (depending on the terminology of the DAW) straight to the internal HDD.
These bounces were all imported again into the same DAW - I think ProTools but I also tried import to others as well with the same net results.
Cubase and ProTools were sample identical - with one or the other phase flipped gave total total silence, all the others varied compared to these two and also to each other - changing levels did not change the differences only made them more, so it wasn’t a DAW volume thing.
As all the DAWS’s were on the same computer and only routing to the internal HDD there was no interface reaction or contributing factors to the test. The phase nulled summed track pairs were also re-exported and the resultant differential result imported into the benchmark DAW to see if there were any artefacts. Also of course we listened to the results with the gain cranked in the original comparison setup as well as the null track re-exports.
We did similar tests with varied interfaces as well comparing their D/A and A/D.

But this has been well established as usually being a question of pan law. Given that the centre:sides ratio is different by default in different consoles and DAWs, the same material may play back differently in them.
It is now less and less common, but 15 years back, it varied drastically, and was often not changeable in DAW options.

Beyond that, I also did think that before Pro Tools went 64 bit float, Logic sounded better to my ears.
My logic there is that there was more channel headroom (logic would read as if you had a peak, but not actually peak, given the essentially-infinite channel headroom, where PT would peak). This limited actual peaking to the output buss or plugins that didn’t have as much headroom.

Anyway, back to the actual question of preamp tone

 

[RoadrunnerOZ]

Pan laws were the same in all tested - Cubase (and ProTools IIRC) had ability to alter pan law settings and all variants were tried in tests. Cubase at that stage was 32bit floating point internally as well as selectable in settings for sampling.

 

[abbey road d enfer]

But this has been well established as usually being a question of pan law. Given that the centre:sides ratio is different by default in different consoles and DAWs, the same material may play back differently in them.
It is now less and less common, but 15 years back, it varied drastically, and was often not changeable in DAW options.

Does it applies here? I believe the test described by Roadrunner in post #164 was a direct bounce of stereo material.

 

[jeremynothman]

I just wonder which record / reply medium is going to cover 20Hz-80kHz range (at a price point which is viable).

Tidal and Qobuz both do already, actually.

 

[jeremynothman]

Does it applies here? I believe the test described by Roadrunner in post #164 was a direct bounce of stereo material.

A direct bounce of stereo material. Isn’t a stereo file simply dual mono panned wide? Won’t pan law still operate?

That’s a genuine question. I’m a lowly fader-pusher, not someone with a solid engineering knowledge.

I recently had a session in which I was sending all the same levels in ProTools to both my interface output and (separately, via a buss) a track for recording a rough mix. I swear that file’s output sounded different to the mix, and I don’t think it was just the psychoacoustic affect of a few samples delay. I’ve got to do a null test to check I’m not dreaming.

 

[RoadrunnerOZ]

Pan law applies to what happens to a signal at pan centre, not at the extremes - it’s reduction by 3, 4.5 or 6dB or other variants when a signal is centred, not a gain increase at the side pan position. -3dB at centre for a mono signal means that both speakers running will give the same power output as one side speaker fully panned. -6dB pan law gives same level when stereo is summed to mono. A stereo file with no panning applied should not be affected by pan law, but in the tests the different pan law settings I tried, in case there was something that affected the results obtained. There was no change to sample rate or bit depth on import or export, no dither applied - stock standard 44.1 16bit.

 

[abbey road d enfer]

I recently had a session in which I was sending all the same levels in ProTools to both my interface output and (separately, via a buss) a track for recording a rough mix.

It's not very clear what you did. Did you create a separate mix routed to another pair of bus? Are you sure you had exactly the same levels in both mixes. It's easy to forget one or two sources, particularly FX.

I swear that file’s output sounded different to the mix,

Was it at exactly the same level? How did you A/B'd them?

 

[abbey road d enfer]

Tidal and Qobuz both do already, actually.

They are not actually media; they are providers of content.
As far as I can see, their high resolution media is FLAC.

 

[jeremynothman]

They are not actually media; they are providers of content.
As far as I can see, their high resolution media is FLAC.

And as a medium, is FLAC not affordable?

 

[RoadrunnerOZ]

I recently had a session in which I was sending all the same levels in ProTools to both my interface output and (separately, via a buss) a track for recording a rough mix. I swear that file’s output sounded different to the mix, and I don’t think it was just the psychoacoustic affect of a few samples delay. I’ve got to do a null test to check I’m not dreaming.

Did you send each individual track of the mix to the stereo bus you used for the aux track source, or did you route your whole mix to a stereo bus and use that bus as source for both interface output and your recording which would be the way to do it - if using sends from each track to a bus for a separate record out you’re creating a submix and unless all sends are post fader and at exactly the same level your mix will sound way different.

 

[jeremynothman]

It's not very clear what you did. Did you create a separate mix routed to another pair of bus? Are you sure you had exactly the same levels in both mixes. It's easy to forget one or two sources, particularly FX.

Was it at exactly the same level? How did you A/B'd them?

All tracks sent to the master as well as to a buss, pre-fader and this buss was sent to an audio track. No FX were used. All levels were sent at the same level to the buss as to the Mix.
If I remember correctly, the key stroke is Opt+Cmd+H - you can copy all level, pan, etc settings to your buss send.

This allows me to monitor with my own levels, and use PFL, but not affect the rough mix. This technique would also be used if I wanted to do headphone sends from Tools.
Again, I might be tripping. I’ll need to do a null test.