THE OPAMPs WAR differences between different types of opamps

[gtudoran]

Hello guys,

I've made a little experiment that was focused on differences between different types of opamps in TS808 like overdrive pedals.
The experiment was focused on:

- Harmonic content
- THD
- Freq. response

Here you have the clip with the experiment, along with some screen shoots - i will answer any questions that you might have and also i will participate in any discussions you like on this subject.
I will not pull any conclusions my self ... but you can post yours
All the necessary info is posted in the video description

RED - 4558 / PURPLE - MC1458 / YELLOW(BLUE) - NE5532 / BLUE(YELLOW) - TL072 / YELLOW -LM358

Regards,
DeX

[GibsonGM]

Visual Analyser...PC scope?  What're you using for probes?

- Can we hear harmonic content that's 24dB down, or is the scale not set up how I think it is?   

- Not a lot of variation between them, it appears.  Any opamp should be capable of fairly accurately reproducing all of our audio content (freq. response).    All I've ever noticed at these low freq's is more or less intrinsic noise caused by the chip's construction...others swear by certain company or date of manufacture...

YMMV...most ICs are the same to me for our uses, other than noise floor...how do I make that "popcorn" icon?   

[gtudoran]

It's only the Visual Analyser, the PC Scope i find it useless. Also i'm using HOLMImpulse for frequency response trace (or you can use Audio Tester V3) - all of them had the same result in 5 to 8% deviation.
Probes are low impedance cables with jacks that were directly feeding/returning the pedal/ from the pedal (i don't think i need low impedance / matched / compensated / etc probes for such application as audio )

Now to answer your question:
- can we?! Don't know some swear that they do (but let's say that the scale is not so accurate and ... well the harmonic content have a bigger amplitude) ... again ... can someone distinguishes something like that?  ... i really don't know ... i don't for sure
- i like we are on the same page (just to be clear this experiment was not made to infirm for to confirm something, was made just to see the differences) ... as you said there are not so much differences and again... i would really like to see who can distinguishes a 0.1 difference in THD ... but again ... some say they do ... i'm not to judge.

Also regarding the freq. response ... i'm quite very happy with the accuracy that my M-Audio fast track offers me in my humble experiments and ... i'm quite confident that at least between 30Hz and 20KHz it will perform excellent (what i can't do is inject big amplitude signals without a voltage divider ... but that is not a problem).

Regards,
DeX

[GibsonGM]

Yes, it's fascinating, and a good study!   I personally can only seem to hear a difference if I listen to FET vs. non-FET opamps...the FET "seem" to be brighter, to my ear.   Many people say they do hear profound differences - the sound of MOJO!     

Perhaps they do. 

I wondered if *bad* probes could alter the THD, but probably not very much.  Let's see if others jump in! 

[gtudoran]

Well...think about it like this, the bad probes are the cables that you are using day by day, in this case the cables are low impedance (for input) the cable at the output is totally irrelevant as the TS808 is a buffered effect....so I do think that it quite irrelevant how good or bad are the probes as long as you are using guitar cable to interconnect various pedals 😁

[GibsonGM]

Yup, but I was thinking more about the distortion - if the cables could be cheap enough to be affecting the readings you were getting. But I doubt it.  You'd have to have a noise source right there and all...

Your experiment simply backs up my long-time beliefs that there is little difference! 

[karbomusic]

 

I personally can only seem to hear a difference if I listen to FET vs. non-FET opamps...the FET "seem" to be brighter, to my ear.

That is what I would expect /if/ there is no buffer in front. For example I have TS type design where the opamp I use is a FET with an input impendence high enough that I don't even need the input buffer.

[Transmogrifox]

I could never tell the difference between op amps.  There is a lot more happening that isn't being worked by a steady 1 kHz tone and a frequency sweep so I have no doubt that one could find a measurable difference if the study were to continue.  That is to say I credit those who can hear the difference, but is there really mojo in it? Well, I guess I don't need these subtle touches to feel the mojo.

If you want to make a mod that unmistakably brings on the mojo, just give the gain knob a 1/4 turn and dig in hard with your pick.  As long as there's an audio-capable op amp in there, this is likely to give you an instant hard-on if you have the right gear.  The best mods on this circuit are placed in the form of 3 knobs exposed to the user right on top of the box...oh, and that stomp switch -- that will burst your zipper.

[thermionix]

I can't tell anything from the screen shots, but I've tried a number of different opamps in my (socketed) 808 clone.  There are differences even within a certain type, though they can be VERY subtle.  Hard to describe, but RC4558P, JRC4558D, and KA4558 all sound a tad bit different from each other.  TL072 and NE5532 even more so.  1458 sounds like crap to me.  YMMV, yadda yadda yadda.

Clipping diodes make a bigger difference to me.  I have sockets for those too.  Right now I'm digging on a 1979 RC4558P and Panasonic MA150s.  It's kind of raunchy.  If I switch to the JRC it gets a little mushy.  If I swap diodes to Toshiba 1S1588s it gets a little cleaner and brighter.  That's actually the opposite of what I expected, so I know it's not just my imagination.

That said, cranked up with a band going, probably wouldn't be able to hear any difference.

[gtudoran]

@thermionix you can save the files (they are actual screen shoots that were taken from my display that was set in HD) and you can iimport the in Inkscape for example on different layers and then modify the transparency of one layer or the other and see what are the differences that is why i've posted them.
Regarding the clipping diodes - 100% on the same page

@Transmogrifox you can tell me what kind of test you think about and i will try to have it done ... but as long as all the audio equipment is measured using this type of tools and tests ... i highly doubt that i will find that astonishing difference (not to mention the mojo)

Anyway i do hope it was useful to make a little bit of light in some dark corners.

Regards,
DeX

I can't tell anything from the screen shots, but I've tried a number of different opamps in my (socketed) 808 clone.  There are differences even within a certain type, though they can be VERY subtle.  Hard to describe, but RC4558P, JRC4558D, and KA4558 all sound a tad bit different from each other.  TL072 and NE5532 even more so.  1458 sounds like crap to me.  YMMV, yadda yadda yadda.

Clipping diodes make a bigger difference to me.  I have sockets for those too.  Right now I'm digging on a 1979 RC4558P and Panasonic MA150s.  It's kind of raunchy.  If I switch to the JRC it gets a little mushy.  If I swap diodes to Toshiba 1S1588s it gets a little cleaner and brighter.  That's actually the opposite of what I expected, so I know it's not just my imagination.

That said, cranked up with a band going, probably wouldn't be able to hear any difference.

[Transmogrifox]

@Transmogrifox you can tell me what kind of test you think about and i will try to have it done ... but as long as all the audio equipment is measured using this type of tools and tests ... i highly doubt that i will find that astonishing difference (not to mention the mojo)

It gets complicated because I'm thinking of large-signal stuff and transient responses.  Like feed it an envelope-shaped signal that is like a guitar signal envelope.  Maybe even record a guitar clip and play it through and analyze it.

Things you might be looking for might show up in a spectrogram or possibly by comparing waveforms in the time domain.

For example if you play a guitar sample track out at exactly the same level and have all the knobs set under plexiglass like you do in the first experiment then the first thing is a listening test.  Just post mp3's and see if any of us can actually hear a difference. 

If the general consensus is that most of us can hear a difference then out comes the signal analysis stuff to find out what is different between the output with one op amp vs the other.  Might even try something like time-aligning the recorded clips and subtracting then look at the spectrum of the difference between the two.  Also could try dividing A/B, and again look at the spectrum of the ratio between the two.

Looking at the difference or the ratio in the time domain might reveal a difference at higher signal peaks or something on a pick attack.  Might be best to bypass the tonestack for this kind of test so the nonlinear and narrow pulse responses don't get smeared in time.

Assuming all playback and recording equipment are the same you would expect the difference between them to show up in a difference (subtract) or ratio (divide).

Even subtracting the spectrums of the samples you have already collected with the sine generator might reveal more than trying to visually interpret it.  It's hard to tell the difference at that scale and it might help if they were overlaid, but even a 1 dB difference would be hard to pick out on overlaid plots.

[gtudoran]

Well it will always be something but yes i can do that - record a small clean guitar clip and then put it through the pedal with different opamps inside.

But i'm sure that then it will be the fact that the pedal was not attacked directly by the pickup ... because of the reluctance of the coil... the impedance was not the same ... and so on

But i will do that test in couple of days (i do have some projects on the table that needs to be finished in a timely manner )  and i will post the wav files here (exported to the highest quality).

Regards,
DeX

@Transmogrifox you can tell me what kind of test you think about and i will try to have it done ... but as long as all the audio equipment is measured using this type of tools and tests ... i highly doubt that i will find that astonishing difference (not to mention the mojo)

It gets complicated because I'm thinking of large-signal stuff and transient responses.  Like feed it an envelope-shaped signal that is like a guitar signal envelope.  Maybe even record a guitar clip and play it through and analyze it.

Things you might be looking for might show up in a spectrogram or possibly by comparing waveforms in the time domain.

For example if you play a guitar sample track out at exactly the same level and have all the knobs set under plexiglass like you do in the first experiment then the first thing is a listening test.  Just post mp3's and see if any of us can actually hear a difference. 

If the general consensus is that most of us can hear a difference then out comes the signal analysis stuff to find out what is different between the output with one op amp vs the other.  Might even try something like time-aligning the recorded clips and subtracting then look at the spectrum of the difference between the two.  Also could try dividing A/B, and again look at the spectrum of the ratio between the two.

Looking at the difference or the ratio in the time domain might reveal a difference at higher signal peaks or something on a pick attack.  Might be best to bypass the tonestack for this kind of test so the nonlinear and narrow pulse responses don't get smeared in time.

Assuming all playback and recording equipment are the same you would expect the difference between them to show up in a difference (subtract) or ratio (divide).

Even subtracting the spectrums of the samples you have already collected with the sine generator might reveal more than trying to visually interpret it.  It's hard to tell the difference at that scale and it might help if they were overlaid, but even a 1 dB difference would be hard to pick out on overlaid plots.

[Transmogrifox]

Yeah, the idea with playing back guitar samples wasn't to try to replicate the whole system, but to introduce a complex signal with a representative envelope and frequency content.  I couldn't think of a better test signal than the real thing.  It may "turn on" some behavior that will show us what the ears seem to tell us.

Well, not my ears , but I don't think all the people who hear a difference are just imagining things.  I am convinced the small difference between the way different IC's behave in this circuit is audible to certain people.

[gtudoran]

Transmogrifox sure thing, will do, if not this week then for sure in weekend.