I don't know how this "noise" looks... But I don't think it is easy to hear the noise directly. I have been just speculating, that when there is less noise, then low level details such as room reverberation can be clearly heard. I mean, when I hear that the vocal intelligibility suffers, I speculate that it is because of noise...
In the first A-F test, I was using my low distortion speaker, and I could hear how some opamps have very clear vocal and how the TL072 suffers. I also heard "hard" bass with TL072. I didn't use laptop as the source, but copied the files using memory card to a Samsung handphone.
I didn't use laptop because the laptop is very problematic (due to audio driver issue). The noise is extremely high. But when I did ABX, of course I used laptop and headphone. And I think because of the laptop's high noise, I couldn't hear noise issue with the opamps. So I'm not sure that it is the noise that makes audible difference. But I'm sure there's contribution from the noise...
Played each of the files once through my mono laptop speaker, my first impression is that LM3900 is more musical and that the vocal sounds more like woman's voice.
Plugged in my headphone and indeed, LM3900 has stronger transient attack that gives it a more "foot-tapping" aspect. Just picked a certain musical note to ABX to show you that the difference in transient attack is obvious.
They are obviously different but I couldn't say which one is better at the moment. I need to hear/ABX the vocal (unfortunately I'm not familiar with Tracy's voice), then to listen through my speaker before I can say which one is better.
Add: I have a strange little feeling that LM3900 is not accurate tho. See later...
Attachments
Last edited:
Ooops... BTW, is this LM3900 versus other opamp, or LM3900 versus DIRECT???
If LM3900 versus DIRECT, unless there is difference in sound level, it is strange how LM3900 has more transient attack. But if it is so, then I have learnt something new...
And it explains why I have a feeling that LM3900 is not accurate. This feeling is based on 2 things: (1) I could feel the instrument with the other file but not with LM3900 (2) The Tracy's vocal with the other file is the one I'm familiar with, which is like a male voice.
If LM3900 versus DIRECT, unless there is difference in sound level, it is strange how LM3900 has more transient attack. But if it is so, then I have learnt something new...
And it explains why I have a feeling that LM3900 is not accurate. This feeling is based on 2 things: (1) I could feel the instrument with the other file but not with LM3900 (2) The Tracy's vocal with the other file is the one I'm familiar with, which is like a male voice.
It is LM3900 vs a direct rip.
The LM3900 was configured for a gain of approx 4.5 and then followed by an attenuation network and then another stage of 4.5 gain. So all four opamps in the package are in use.
The differences here are very obvious and its easy to get 100% correct in identifying them. The performance of the LM3900 by todays standards is grim in the extreme (for audio) but it kind of puts in perspective how small the improvements in opamps have become. It is an interesting device to revisit though. In the early 70's it was quite something.
The LM3900 was configured for a gain of approx 4.5 and then followed by an attenuation network and then another stage of 4.5 gain. So all four opamps in the package are in use.
The differences here are very obvious and its easy to get 100% correct in identifying them. The performance of the LM3900 by todays standards is grim in the extreme (for audio) but it kind of puts in perspective how small the improvements in opamps have become. It is an interesting device to revisit though. In the early 70's it was quite something.
Oh my! I had assumed I was comparing to TL072. Simple circuits will always have something to deliver. May be it will be interesting to design a discrete version!?
It is strange how a circuit can add such an effect as more transientt attack and musicality. This reminds me of the Harsch crossover in the other thread. The transient attack is superior but the tonality seems wrong to me.
Add: But I wish somebody would check the files for differences, because the idea that the original has less transient attack is strange to me. May be there is extra gain, even if small (Usually it is strange how small measured difference is so audible).
It is strange how a circuit can add such an effect as more transientt attack and musicality. This reminds me of the Harsch crossover in the other thread. The transient attack is superior but the tonality seems wrong to me.
Add: But I wish somebody would check the files for differences, because the idea that the original has less transient attack is strange to me. May be there is extra gain, even if small (Usually it is strange how small measured difference is so audible).
Last edited:
From the quick listen I had, the LM3900 seems 'closed in', the direct rip opens the soundstage up dramatically. If you saw the squarewave response... well its laughable. Hugely asymmetric, more like a triangular wave on the up slope (its very much slew limited) but not so bad on the falling edge. 10kHz square is a distorted mess.
But it doesn't sound that bad in reality. If I was forced to choose between having to use 741's or LM3900 I would go for the 741.
But it doesn't sound that bad in reality. If I was forced to choose between having to use 741's or LM3900 I would go for the 741.
Is it THAT bad?
Yes, for some inaccurate sounds, it takes us some listening time to realize that the sound is wrong. There are a lot of examples for that.
Sounds like you have been very busy
Hi Mooly,
I took a quick scan through all postings, but I did not succeed in finding an overview about the exact outcome.
So would it be possible for you to make a resume of:
1) how many participants joined the test
2) was there any result that can be regarded as conclusive, and how did this look like.
3) how did the best sounding OPA compare to the original sound.
I would very much appreciate your reaction.
regards,
Hans
I'll add that anyone interested in training their ears to hear spectral balance, harmonic distortion, and phasing issues, ought to invest in trainyourears. Their most recent edition includes many premade training sessions. You can also create your own training sessions and use a program like cooleditpro to manipulate your content.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.