Hi Nico,
I've had a good listen. First is sharper, quite sibilant, with a slightly forward instrumental backing, voice a little more recessed and mechanical. Second is smoother, little sibilance, with a forward vocal, recessed instruments, except for the castanets.
This is pretty much what I've come to expect from PP versus SE. However, the nested feedback mod I suggested to you earlier for the PP circuit will greatly enhance the imaging issues, render the vocal more prominent and add depth to the instruments.
Subjectively speaking, there is for me more coloration in the SE sound but it is musical. The coloration of the PP design is slight, but not as benign. The best way to describe the difference lies with the word 'fatigueing', an urgency of presentation in the PP circuit which militates against a relaxed, laid back listening experience.
This is all very subjective, and people's views vary, but a good portion of the listening public would prefer the SE sound.
Thank you for making a strong effort to show the differences. Of course, a very good compromise here is to use a high bias quasi complementary output stage - arguably the best of both worlds. But that needs a phase inverter, and in this single emitter follower output stage configuration, that's probably wasteful. Better to refine the fully complementary circuit further.
Cheers,
Hugh
I've had a good listen. First is sharper, quite sibilant, with a slightly forward instrumental backing, voice a little more recessed and mechanical. Second is smoother, little sibilance, with a forward vocal, recessed instruments, except for the castanets.
This is pretty much what I've come to expect from PP versus SE. However, the nested feedback mod I suggested to you earlier for the PP circuit will greatly enhance the imaging issues, render the vocal more prominent and add depth to the instruments.
Subjectively speaking, there is for me more coloration in the SE sound but it is musical. The coloration of the PP design is slight, but not as benign. The best way to describe the difference lies with the word 'fatigueing', an urgency of presentation in the PP circuit which militates against a relaxed, laid back listening experience.
This is all very subjective, and people's views vary, but a good portion of the listening public would prefer the SE sound.
Thank you for making a strong effort to show the differences. Of course, a very good compromise here is to use a high bias quasi complementary output stage - arguably the best of both worlds. But that needs a phase inverter, and in this single emitter follower output stage configuration, that's probably wasteful. Better to refine the fully complementary circuit further.
Cheers,
Hugh
Last edited:
Hugh, I don't know your opinion on this however:
I believe class AB is an interesting compromise sound-wise. During normal volume the output stage will be in class A - this will be conducive to even harmonics. But on rough parts, say explosions or dramatic moments which are generally not laid-back, the amplifier is pushed into B, where 3rd harmonics are injected and usually dominant. In this way, perhaps AB is more suited to movie sound. (I suppose the question is whether you want to "experience" the drama as much or more than you want to experience everything else)
If we go for a class A SE then it will be a different animal in this respect.
Do you have any comments on this?
- keantoken
I believe class AB is an interesting compromise sound-wise. During normal volume the output stage will be in class A - this will be conducive to even harmonics. But on rough parts, say explosions or dramatic moments which are generally not laid-back, the amplifier is pushed into B, where 3rd harmonics are injected and usually dominant. In this way, perhaps AB is more suited to movie sound. (I suppose the question is whether you want to "experience" the drama as much or more than you want to experience everything else)
If we go for a class A SE then it will be a different animal in this respect.
Do you have any comments on this?
- keantoken
Anthony,
I agree with you!
The PP HAKSA has a monotonic decreasing distortion profile at lowish volume, no more than 1Vrms. Beyond that, and coinciding with entry into Class B switching operation, the H3, H5 and H7 artefacts start to increase sharply, particularly H3 and H5. All this is visible on LTSpice, BTW. This abrupt change in the harmonic profile is not musical, and manifests as aural unease and hardness, aka 'fatigue'. This also happens with the SE amp, of course, but only at clip when half the waveform falls apart completely.
In HT applications, we are generally only registering peaks on special effects (unless it's a musical!). Consequently, the ear/mind interface, whose powers of aural discrimination are greatly reduced owing to visual stimulus, seeks out real world sounds, lacking in any musicality however, so the loss of musical pitch harmonics is no big deal. Consequently, we can choose chip amps, and run 'em close to Class B for highest efficiency, necessary for the high packaging density of multichannel amps.
Cheers,
Hugh
I agree with you!
The PP HAKSA has a monotonic decreasing distortion profile at lowish volume, no more than 1Vrms. Beyond that, and coinciding with entry into Class B switching operation, the H3, H5 and H7 artefacts start to increase sharply, particularly H3 and H5. All this is visible on LTSpice, BTW. This abrupt change in the harmonic profile is not musical, and manifests as aural unease and hardness, aka 'fatigue'. This also happens with the SE amp, of course, but only at clip when half the waveform falls apart completely.
In HT applications, we are generally only registering peaks on special effects (unless it's a musical!). Consequently, the ear/mind interface, whose powers of aural discrimination are greatly reduced owing to visual stimulus, seeks out real world sounds, lacking in any musicality however, so the loss of musical pitch harmonics is no big deal. Consequently, we can choose chip amps, and run 'em close to Class B for highest efficiency, necessary for the high packaging density of multichannel amps.
Cheers,
Hugh
Last edited:
Hello Hugh and all the guy's
I've read all the tests and opinions for HAKSA PP and SE amps.
I just have an old Sennheiser HD-420 headphone and few cheapo ones, there is now much better ones today, so my old HD-420 will not have the best sound compared to the newest Sennheiser.
But at close listening, like with an headphone, I think a warm SE amp sound would have my favor.
So I vote for the SE.
I will wait for the final decision of all the guy's.
Bye
Gaetan
I've read all the tests and opinions for HAKSA PP and SE amps.
I just have an old Sennheiser HD-420 headphone and few cheapo ones, there is now much better ones today, so my old HD-420 will not have the best sound compared to the newest Sennheiser.
But at close listening, like with an headphone, I think a warm SE amp sound would have my favor.
So I vote for the SE.
I will wait for the final decision of all the guy's.
Bye
Gaetan
Last edited:
Okay Hugh, why not add schottkeys to the class AB's emitters? This smoothens clipping and 3rd harmonics don't rise sharply after cutoff (although they are increased in class A mode)... In fact, we could operate the OPS at as low as 6mA (I tested this on the sim with MBR745 diodes). Since prototypes have been built, why not try it?
- keantoken
- keantoken
I'll say it's subjective. Hugh you do realize that you are listening to two files, one of which is inverted. Both tracks are at different levels, not only between each other but between channels.
When I had my first listen this morning, I could hear a difference. As the day progressed and I managed to get these files to where they are almost on a par, that difference has vanished. What I am listening to now is indistinguishable and I'll say it again: these files are so exactly the same, they are almost certainly from the same amp.
Look at this, left (top) and right (bottom) from both files overlaid.
Now, an interesting thing happened while I was getting that graph together, I noticed that the left channel looked very similar to the right from the other file and vice versa. I slide the overlay up and presto! exact match. That was the final piece of the puzzle - swap channels on sample #2.
I'm not impressed.
Last edited:
Hmmm. John, you are saying, I presume, that Nico has given us the same file from the one amp, inverted it and changed the level, so as to confuse?
And then, of course, perpetrated an artful dupe on all of us?
This is kinda strange. Let's ask Nico.
Nico, is this two files, modded to look different, but from the same amp? Or are they two entirely different amps?
I find it hard to believe you would do this, but this is the net, so anything is possible, so I'm asking now.
If I heard a difference, it would almost certainly be level only, and my comments are trite, and I'm a goose. If Nico did deliver two files from two different amps, then we are all smart fellas. I guess when he answers, we will know for sure!
Nico, over to you.......
Cheers,
Hugh
And then, of course, perpetrated an artful dupe on all of us?
This is kinda strange. Let's ask Nico.
Nico, is this two files, modded to look different, but from the same amp? Or are they two entirely different amps?
I find it hard to believe you would do this, but this is the net, so anything is possible, so I'm asking now.
If I heard a difference, it would almost certainly be level only, and my comments are trite, and I'm a goose. If Nico did deliver two files from two different amps, then we are all smart fellas. I guess when he answers, we will know for sure!
Nico, over to you.......
Cheers,
Hugh
Hello guy's
I did rely on Nico post #842, his SE sound description say a lot to me.
I cannot download the Nico's test files since I'm in a low speed internet (5Kb download speed), so I can't check them myself, but I don't think Nico would have given us the same file from the one amp.
They can be very small differences into the harmonic content between those two files, that can not be easy to see on the images, but the resulting sound would be differents between the PP and SE test files.
Bye
Gaetan
I did rely on Nico post #842, his SE sound description say a lot to me.
I cannot download the Nico's test files since I'm in a low speed internet (5Kb download speed), so I can't check them myself, but I don't think Nico would have given us the same file from the one amp.
They can be very small differences into the harmonic content between those two files, that can not be easy to see on the images, but the resulting sound would be differents between the PP and SE test files.
Bye
Gaetan
Last edited:
Yes, Gaetan, I think you are right.
Nico, I am sure this is male bovine excreta. I can't see you doing anything but being 100% genuine and helpful!
What this does show clearly is that even relatively fine measurements and graphing fail to pick distinctions which, with care, the human ear can easily detect. This says quite a lot about measurement technology, and it isn't flattering.
Cheers,
Hugh
Nico, I am sure this is male bovine excreta. I can't see you doing anything but being 100% genuine and helpful!
What this does show clearly is that even relatively fine measurements and graphing fail to pick distinctions which, with care, the human ear can easily detect. This says quite a lot about measurement technology, and it isn't flattering.
Cheers,
Hugh
Hi Guys,
interesting you should say that. I did not give the same files as you can see the absolute difference in my samples. Furthermore, is it not a extremely stupid thing to analise the spectrum, this is completely irrelevant anyway, shows how much you know of what to measure. You are confirming is that the same song might be playing - rather ignorant I would say, you see the FFT of the music, nothing of the amplifier.
What did you expect to see in an FFT, Tracey Chapman's voice print, what guitar is playing over what amp, how much fuzz or overdrive is introduced?
I did take a lot of trouble doing this, messed up my week-end for what? To show that there is actual differences in the time domain, (oscilloscope plot) this is the only place that you will really see what you hear.
That nobody realised that the SE is inverting also shows the complete lack of engineering skill.
I am rather disgruntled by these accusations - I have been an engineer for longer than most of you have been alive, I do not cook the books I am not an accountant, nor do I hide mistakes I am not a DIY - nobody is paying me for contributing to this thread or to make up your mind into what you should or should not build. I could not care less, I am not even considering either of these amplifiers anyway, they do not measure even nearly to my standards.
I don't subscribe to hocus pocus, engineering is an exact science not a gut feel, everything is measurable and can be explained.
That is me for this thread.
interesting you should say that. I did not give the same files as you can see the absolute difference in my samples. Furthermore, is it not a extremely stupid thing to analise the spectrum, this is completely irrelevant anyway, shows how much you know of what to measure. You are confirming is that the same song might be playing - rather ignorant I would say, you see the FFT of the music, nothing of the amplifier.
What did you expect to see in an FFT, Tracey Chapman's voice print, what guitar is playing over what amp, how much fuzz or overdrive is introduced?
I did take a lot of trouble doing this, messed up my week-end for what? To show that there is actual differences in the time domain, (oscilloscope plot) this is the only place that you will really see what you hear.
That nobody realised that the SE is inverting also shows the complete lack of engineering skill.
I am rather disgruntled by these accusations - I have been an engineer for longer than most of you have been alive, I do not cook the books I am not an accountant, nor do I hide mistakes I am not a DIY - nobody is paying me for contributing to this thread or to make up your mind into what you should or should not build. I could not care less, I am not even considering either of these amplifiers anyway, they do not measure even nearly to my standards.
I don't subscribe to hocus pocus, engineering is an exact science not a gut feel, everything is measurable and can be explained.
That is me for this thread.
I didn't measure anything, I just worked with the data Nico presented and the conclusions are obvious for anyone who has their eyes open to see.
I have done this before, with my own gear and there would be differences that were plain to see. Changing an opamp in the same phono preamp netted visible differences in the waveform at this level.
Shown in post 906 is the 2 channel output from 2 completely different amps - different devices, different topology, different power supply, different class of operation, too many differences to be this exactly the same.
If they are not the same amplifier, then the only logical conclusion is that all amplifiers sound the same, because these two files do sound the same after they have been "straightened out".
The amp that Hugh presented in post #821 does not invert. Why would it invert?
Regarding the files, why, when I do all that I did to put them on an equal level, do I wind up with identical waveforms? If these were identical from the start, it would not be questionable but that's not the case - they looked different and they even sounded different. To recap what I did for those of you who obviously didn't bother to read my earlier posts:
First, I made the files start at exactly the same time and discarded the lead in time to this point. Next I made them the same length - exactly 1 minute. I then inverted the waveforms on sample #2 (the one you said was the SE and "naturally" inverted). I normalized the volume on all 4 channels to the same level. Finally, I switched left for right in sample #2.
Viola! The same! How?
I'm sorry if I have upset you Nico but I call it as I see it (you yourself said this about me earlier). Please, show me where I got it wrong.
I asked for a 1khz tone from each amp but you didn't provide it, why not? You mention FFT above, but I wasn't looking at that at all, just the waveforms, as you did earlier. Perhaps this would shed new light on the situation.
Comparison.
I can see the differences, unless my eyes are also going. These two signals are normalised to 0dB and in time only microseconds out of sync.
But we are talking PC sound cards which are at best suspect, a decent storage scope should show the differences well.
I am not here to prove the one over the other, just to show that there is a marked difference in the graphically representation, there may be less than a difference in what one hears or hope to hear.
I can see the differences, unless my eyes are also going. These two signals are normalised to 0dB and in time only microseconds out of sync.
But we are talking PC sound cards which are at best suspect, a decent storage scope should show the differences well.
I am not here to prove the one over the other, just to show that there is a marked difference in the graphically representation, there may be less than a difference in what one hears or hope to hear.
Attachments
Sorry John about the one KHz signal you said it should be perfect, a PC cannot produce a perfect sine wave. You only have 44.1 samples per cycle, hardly anything to judge an amp by.
Besides I said I will do in using a signal generator which has at least a decent signal. This would probably take me another few hours at work just to be criticised.
But I guess that you expect me to pass every moment for your pleasure. You can build this circuit and generate your own signals and make your own choices, I have nothing to prove, to anyone.
I provided information as a service to this community, you are acting like the beggar who gets something for nothing and then bitch about it. Use it or lose it.
Hugh I am sorry but I guess you are on your own from now on. I have far better things to do than trying to protect my integrity with fruitless and stupid arguments, I have not pushed anyone into accepting anything.
Cheers
Nico
Besides I said I will do in using a signal generator which has at least a decent signal. This would probably take me another few hours at work just to be criticised.
But I guess that you expect me to pass every moment for your pleasure. You can build this circuit and generate your own signals and make your own choices, I have nothing to prove, to anyone.
I provided information as a service to this community, you are acting like the beggar who gets something for nothing and then bitch about it. Use it or lose it.
Hugh I am sorry but I guess you are on your own from now on. I have far better things to do than trying to protect my integrity with fruitless and stupid arguments, I have not pushed anyone into accepting anything.
Cheers
Nico
That's too bad Nico, I thought a man of science such as yourself could withstand a little criticism and that your measurements would bear up under any scrutiny. Nothing personal here, no argument, just my observations.
I have saved each cleaned up file to constant bit rate 256k Mp3 for anyone who wants to listen to them. I can email - they are less than 2Mb each.
I have saved each cleaned up file to constant bit rate 256k Mp3 for anyone who wants to listen to them. I can email - they are less than 2Mb each.
Nico, thanks for taking the time to prototype the amp and post the wav files. I'm an experimental scientist and so I know what a faff it can be to take supposedly 'simple' measurements like these in reality. After a quick listen they both sound very similar, I will have to do an extended comparison later. (I think I will have to try comparing shorter segments, as, by the time I get to the end of one sample I have forgotten exactly what it sounded like at the beginning when I start the second track)
I have looked at the two files in Audacity and there are subtle differences between the two. Your comment about the harmonics was interesting, I had not though of that before, a close inspection reveals the differences you talked about. You're right about sound cards being a bit suspect, if I align the tracks at to a given sample at one point they are very slightly out of sync some time after.
The differences are small but I would expect them to be. If the differences were large one of the amps would be rubbish (assuming the other was good) as its output would not be accurate copy of the input. After all the THD is a measurement of the mathematical (in)accuracy of the output compared to the input and it is pretty small for both amps.
Phil
I have looked at the two files in Audacity and there are subtle differences between the two. Your comment about the harmonics was interesting, I had not though of that before, a close inspection reveals the differences you talked about. You're right about sound cards being a bit suspect, if I align the tracks at to a given sample at one point they are very slightly out of sync some time after.
The differences are small but I would expect them to be. If the differences were large one of the amps would be rubbish (assuming the other was good) as its output would not be accurate copy of the input. After all the THD is a measurement of the mathematical (in)accuracy of the output compared to the input and it is pretty small for both amps.
Phil
distortions of a few tenths of a percent can he heard when comparing clean to clean+distorted.
Graphical comparison is incapable of showing the difference between these two signals.
I would expect all competent amplifiers to give a graphical representation that is visually identical.
Gross distortions around 10% can be seen. But I don't believe that is what these two amplifiers are producing.
Graphical comparison is incapable of showing the difference between these two signals.
I would expect all competent amplifiers to give a graphical representation that is visually identical.
Gross distortions around 10% can be seen. But I don't believe that is what these two amplifiers are producing.
Last edited:
That would be your expectation, why not give it a try? Different recording sessions using the same setup and amplifier will net visibly different results, let alone a completely different amp.
We are comparing a low distortion, multi-gain stage, push-pull class AB to a high distortion single ended amp. Frequency response would be VERY different, not this:
Both files, cleaned up and overlaid. The differences are consistent with two recording session of the same amp.