Superbly put together video, Jim! Having tried almost all these solutions myself, I appreciate the time that went into this. Was especially interested in Music Construction Set results since I've wanted to cover it and include proper PC speaker recordings. Also, I'd never even thought of trying in-ear mics!

@TheOldskoolPC

4 жыл бұрын

I was so disappointed by the in-ear mics -- I was hoping I'd stumbled onto something nobody else had thought of. Well, they sucked, so I guess I was the last to know :-) Thanks for the kind words!

@whisk0r

4 жыл бұрын

@@TheOldskoolPC It's interesting how the resonant frequency of your ear seems to have a filtering effect on the power supply fan noise. I am sitting in an ear different from the one you are in now.

I'd go with option "record the speaker directly". It kind of emulates the pain of not being able to hear what your PC speaker was trying to tell you.

I think the "Tap speaker leads" succeeded the best of all.

@oscwavcommentaccount

3 жыл бұрын

same

The soundcard solutions may not be the most authentic, but to my ears the sonic quality was the best in almost all examples.

Great video! I'm not in the market for recording from the PC speaker, but the overview of the various advanced methods for generating sound from that humble device is fascinating in itself.

@TheOldskoolPC

4 жыл бұрын

Agreed. I hope to make a video on Realsound someday, including how there was prior art that should have invalidated the patent.

Was really interesting to hear all of these different methods to produce sound on the PC speaker. I agree that when you mic'd the speaker with the case off it sounded the most clear and authentic. I liked the absence of fan noise. I suppose ultimately the method of recording would depend on what you're going to do with it. I could see a line recording working well if you're capturing video directly from a video card, and the mic'd speaker audio going good if you're filming the actual computer, but then again it probably wouldn't look great with the case off if you're trying to show the computer running something. Huh. Thought provoking video Jim, I'm going to be doing an Apple II video in a couple weeks or so and think I'm going to try to mic the speaker like you did here.

@TheOldskoolPC

4 жыл бұрын

I recorded a speaker for several days so you don't have to! :-)

I came to this video to show a much younger friend what he missed out on - and stayed for your engaging, warm, informative and demonstrative style. You are really good, and you deserve way more subs. Let me add one.

@TheOldskoolPC

Ай бұрын

Thanks for the kind words! Maybe your young friend might like some other videos on the channel.

When recording via the sound card with the cable, be sure to mute all of the unused inputs/sources in the sound card's mixer (line in, mic, etc.) to reduce the amount of background noise and interference from the CPU. This applies even today with modern sound cards and onboard sound chips!

@TheOldskoolPC

4 жыл бұрын

Good point! While not shown in the video, I did indeed do that, for both the SBPro I was using with a cable, and the PAS16. What I did *not* do, however, was make sure the RealSound filter bit was turned on for the PAS -- not because I forgot, but because I found it didn't make any difference (to me, with a limited sample set) so I just left it alone and didn't fiddle with it in autoexec.bat.

"Done is better than perfect" - Wise words!

@TheOldskoolPC

4 жыл бұрын

It helped me finish the video. I would have obsessed over it even longer had I not heeded that advice...

@noapparentfunction

4 жыл бұрын

i learned that phrase a long time ago and it totally transformed my projects

Recording the speaker directly is my favorite because it sounds exactly like I remember but the Sound card with without internal cable is interesting.

Back when I was my younger self playing mostly DOS games, I was a big fan of the cable to soundcard method. I actually modded a cable that ran the PC speaker through the CD line in that I used for at least one computer. A clever trick for the convenience of always using headphones. Never had any issues with the CD audio having to take the long way around, but it's not like I was doing this kind of thing in the mid 90s.

Mic with the case off sounded the best & I was glad when the score reveal came and you said the same. NONE of the other methods for capturing the BEL character sounded natural, but just micing speaker with case off sounded spot on. I didn't realize capturing how crappy the tinny speaker is would be important. However, some of the PCM tests you did sounded best captured directly with the soundcard. Shiru's music & the guy counting, there is some detail lost (authentically) when played through an actual PC speaker and not just generating sine waves. If I were an emulator developer, I would implement a tick box to apply a "tinny speaker" effect, trying to calibrate it to the BEL character, along with the rest of the samples in your video. This would keep the detail that the authors hoped and intended for you to hear, in works that push the PC speaker beyond what its tiny cone is capable of reproducing. Simultaneously, at the tick of a box, you can approximate what the noise coming out of the speaker actually sounded like. This was an important video, thank you for your work in preserving the sound of the IBM PC speaker. P.S. your old ears might not be able to hear it, but your binaural mics have a high pitched squeal that no other mic input method has. You should use a lowpass filter if you ever use those mics for anything.

@TheOldskoolPC

3 жыл бұрын

Thanks for the note about the binaural mics. It's possible they picked up the 15.7 KHz of the monitor's flyback because they were closer to it (at ear level). I'll never use them for this purpose again though. I have a feeling that putting a lav mic inside the case is probably the way I'll have to do it moving forward. It's just that SB output is so convenient...

Very interesting to see all the different techniques they used to squeeze sound out of the pc speaker.

@TheOldskoolPC

Ай бұрын

Some of the results surprised me when I was making the video.

I think the conclusion is that a combination of techniques would be the best... You can use the soundcard or tap cable to bring the signal to the outside of the case, while the case can remain closed. Then you could connect an original PC speaker to that signal outside the case, and you can record that with the mic. You could even create a sort of 'isocab', where you put the PC speaker and the mic inside some isolated box, to prevent it from picking up ambient noise, so it can even be used in somewhat noisy environments.

@Scalibq

4 жыл бұрын

Another thing is... the most noise is generated by the PSU and the HDD. HDD can be replaced by a CF-IDE solution, which is silent. And I'm pretty sure some handy electronics hobbyist could whip up a passively cooled replacement PSU built from modern components. With a CF-IDE solution, you probably won't even need the full 130W power anyway. Your average modern laptop PSU is more powerful than that, and requires no active cooling.

@TheOldskoolPC

4 жыл бұрын

I like the "isocab" solution for if I ever have a dedicated space to do captures in. Currently I need to be able to quick-change my layout due to lack of space and varying needs.

Excellent video! All the best with the recordings :)

I think your conclusion is good, in room, either directly by the isolated speaker or otherwise sounds best, it gets rid of much of the harshest bits of the sound

How incredibly specific. I love it. Must have taken days to edit.

@TheOldskoolPC

3 жыл бұрын

It did!

100 ohm variable potentiometer in series with the speaker as a volume control, turn the 100ohm pot to minimum to bring it to line level and not speaker level. fit a mono 3.5mm jack to the case to output into your wav recorder (zoom/tascam etc) leaving it at speaker level will destroy the recorders line input circuitry. and leaving it at speaker level will reveal electromagnet noise from the motherboard. in my chassis i fitted a 5watt 2" speaker with rubber surround, as many games play real music and fx through the speaker it sounds excellent, obviously as i said you need that 100ohm pot as a volume control as standard pc beepers are very high impedance, directly connecting a speaker without it will be extremely loud on a 4ohm speaker driver.

I'd be interested in seeing how an audio isolation transformer of the right impedance stacks up to these approaches- it's my usual trick for recording audio from devices that are expecting a speaker.

Thank you from Italy for this very clever and useful video!!!

Very nice video. Great level of detail. Didn't know there were so many different "types" of sound to contend with. Good explanations! Thanks! ~ CHAP

How come I never seen this video and all my favourite youtubers are commenting. Subscribed

Great video, thanks. Rambled in more detail on twitter.

wow each different characteristic resulted in varying results from each method. Depending on the ambience you want, several options are good.

Great work and great video, thanks!

Very interesting comparison, my favourite is probably direct recording from the soundcard via the cable, because of the inherently cleanest signal. The audio card with its own beep generating circuitry is a very interesting oddity though, I had no Idea that existed. PS. I love how you used BotB music as the background later on :3

Just because the way you phrased it leaves much to the imagination: "beeper designed to run at 0 or 5 volts"; nothing runs at 0 volts, I think you mean it's either off or on and then the motherboard controls the frequency for pitch. Awesome video. Thanks!

@TheOldskoolPC

2 жыл бұрын

I agree "zero volts" is bad wording! But yes, you're correct, it's up to the computer, whether through timer control or CPU control, to adjust the frequency. The fun begins when you use two timers together to pulse the speaker faster than it can physically react, to get intermediate amplitude positions :-)

Great video. I've come to the same conclusion (on newer PCs). Recording close to the PC-speaker (case off, or with the speaker outside of the case) is the best method. I have a few recordings on my channel. There's quite a bit of background noise but otherwise the recordings sound quite close the the real thing.

FYI, the IBM PS/2 model 25 (the 8086 models, at least) also has a headphone jack for the internal speaker.

@TheOldskoolPC

4 жыл бұрын

Indeed, it was disqualified since it is too fast. Also, because it isn't line-level output -- I tried to record from it a few years ago (there's a vcfed.org forums thread on it) and found that line-level input recording devices couldn't record from it for some reason. It was very odd, and I never got to the bottom of it.

@NikolayNikolov_Tralala

4 жыл бұрын

@@TheOldskoolPC Interesting. I've never tried recording from it, in fact I've never used the headphone jack on this computer. :)

The pitch drop on the tandy 1000 is interesting, but overall for me is close miced. However that odd cabinet resonance of the mic in front of the case is certainly authentic.

@TheOldskoolPC

4 жыл бұрын

Most have the same sentiments. Something I didn't try was putting a small lav mic inside the case -- might try that at some point.

Was looking exactly for something like this so I could hook up PC Speaker out directly into an amplifier setup for my ultimate retro gaming setup, I personally found the Direct Speaker Tap to produce the best sound (though for your perspective of perfect archival of what the actual Speaker sounded like that the Mic inside the case was the most authentic to my ears too) from the perspective of you could clearly hear a lot more nuances that were lacking from the other methods though it probably needs to be run through a filter or two for my purposes

Damn good video! nice to see such an in depth look into speaker recording. Now I'm interested to know more about how music construction set works, do you know any resources that I could use to find out a little bit more?

@TheOldskoolPC

4 жыл бұрын

My friend reenigne disassembled the code and commented on it, but I haven't written up our conclusions yet, sorry. I do hope to do that in a future blog article or video. The basic idea is that it's a 4-channel, 1-bit, ZX-Spectrum like pulse player. Three of the channels have 2 volume levels (pulse width 1 or 2 samples) and the fourth channel has the same plus two additional possible louder volume levels (3 and 4 samples).

I preferred the micing of the speaker directly with the case off for most instances. It sounds the best over KZread and mobile speakers which a lot will be listening to. I bet you could tap the speaker with a line cable and put a ground loop isolator between that and your recording device to eliminate the external frequencies you hear in the quiet parts. Maybe mix the two to get some of that reverberation you get from the actual speaker.

@TheOldskoolPC

4 жыл бұрын

I hadn't considered which method would sound best over headphones, I admit. Now I need to scurry off and see what the analytics look like for my channel, to see where people are coming from. One scenario I missed is to mount a mic inside the case as a more feasible permanent setup. I may try that with a cheap lav as a test.

An Shure SM57 placed in front of a Marshall Amp has a distinctive sound that cannot be duplicated by running the guitar cable directly into a mixing board. Similarly, the 2.25 inch 8Ω speaker on a PC creates it's own distinctive coloration to the audio signal. Also, the motion of a voice coil in a magnetic field cannot accurately reproduce a square wave. How the speaker deals with those square waves is part of the unique PC speaker sound. Proximity helps minimize fan noise and room acoustics. Binaural microphone recordings when played back through quality headphones may represent what a listener hears, but this does not make it the best representation of how a PC speaker sounds considering environmental noise, room acoustics, and hiss from the required gain. Similarly, the microphone placed outside the case is picking up a lot of fan noise. My favorite is the microphone placed next to the speaker.

@TheOldskoolPC

4 жыл бұрын

If I read your response correctly, I think you are supporting my conclusion that different people will have different needs based on what exactly it is they're trying to capture. I agree the mic next to the speaker sounds the most "authentic" but unfortunately I need something that is more practical for capturing many hours of footage, so I've settled on the sound blaster method, because it's the only line-output method that works with bootable programs/games (PAS needs a driver to activate it).

@ThriftyAV

4 жыл бұрын

@@TheOldskoolPC The sound blaster method is likely the most practical and most faithful to the square waves. All speakers will "round out" the edges on the square waves, even if most don't sound quite like the old 2". Your video is a very interesting experiment regardless, and I'm glad I stumbled into it.

Great, informative video! Looking forward to any future uploads. "Done is better than perfect." Very true... Read that while working on a turntable, hiding tiny chips in the black plastic using legal ink and a syringe. Oof. 😉

@TheOldskoolPC

4 жыл бұрын

Yes, sadly I worked on this video for 6 months, on and off, and on weekends. Next video will be up in October due to VCFMW planning (and recovery!) in September.

THE ultimate guide! Only thing missing is a mic drop at the end, but that mic looks expensive. :-) (What make/model is it?) Thanks for linking to that PC speaker program collection too. Before owning a sound card I used to collect those "beeper speech" programs off BBSs and the like, but the one you used in the 1-bit PCM section is new to me.

@TheOldskoolPC

4 жыл бұрын

That mic was incredibly cheap, it's a Zoom H1n, about $85. I don't recommend it because the gain knob can change even if you put it in the LOCK position while recording. I'll use a TASCAM in the future.

@x86VileR

4 жыл бұрын

Ah, gotcha. I have a Shure SM58 here (+ tube preamp and all) - it's supposedly designed for voice, but I've had good results using it with a guitar amp cabinet, so I think it might work for this purpose too. Gotta test this thing now :-D

Methodical, well-articulated, interesting. Thank you for this. I also agree with the assessment that mic-near-speaker combo sounds the closest to the original intent. Direct wired capture sound too harsh no matter the method; and distant mics (whether outside the case or binaural) pick up so much noise that it's distracting. I'm also impressed by the PAS16's rendition. Surely there's some kind of attenuation or cleaning going on because of all the direct wired captures, it sounds the smoothest and most pleasing, keeping details in the sound that get obliterated in other wired examples. I've been told a couple of times (by purists?) that the case reverberation was intended to be part of the sound - I wonder if capturing with the H1n inside a closed case, would sound that much different from the open case.

@TheOldskoolPC

4 жыл бұрын

Putting a mic inside the case was a use case I completely missed, and I think that's worth exploring for a possible follow-up video. The interesting part is how it would sound, since we normally don't listen to the speaker from INSIDE the case, but outside it. Subbed to your channel BTW (with a name like that, how could I not?) :)

@theSoundCarddatabase

4 жыл бұрын

@@TheOldskoolPC I'm very honoured! I don't put out content as often as I would like, but I try to make each video more polished than the last. Cheers!

Your test are so great!! amazing information too, the game TRISTAN PINBALL what type of method use? intro sounds very great on that era.

@TheOldskoolPC

3 жыл бұрын

What type of sound generation method did they use? They kept two or three frequency counters in memory and just counted them down at the frequencies they were tracking, and every time it hit zero, they'd toggle the speaker. Very simple method for multi-voice :-)

@nopochoclos

3 жыл бұрын

@@TheOldskoolPC Too much data!!! haha.. but the sound is great! thanks for the technical data!!! sorry for my brute english

Jim, have you considered trying the speaker lead tap method, but adding a capacitor to the equation? If you find the right cap value, I bet you could do a really good job of simulating the physical characteristics of the original PC speaker that affect how it sounds. Take away a little bit of the "punch." You could probably get something both really clean, yet still pretty true to how it really sounds sitting in front of the machine. I don't know if it'd be worth doing that when you can just use the PAS16, but it was just a thought.

@TheOldskoolPC

4 жыл бұрын

I have not yet attempted any of the more advanced circuits, from a single capacitor to some of the amplifier designs I link to in the video's description. They would definitely produce a better result than tapping the leads, but part of my setup needs to strike a balance between authenticity and convenience.

I thought the sound card with cable and tapping the speaker directly were by far the best, the sound card being just a little better than the speaker. The sound card _without_ cable sounds good but you can tell it's not actually from the speaker. P.S. You deserve way more subs!

@TheOldskoolPC

3 жыл бұрын

It seems everyone has a different favorite. I've actually gone back on my decision and am looking to mic the speaker from inside the case.

Interesting approaches; we used to wonder what the best solution is when it comes to getting a 256b intro shown on a bigscreen - using a cable either from the motherboard or through a soundcard was usually the winner, mostly because It Just Worked and you ended up with a standard relatively clean line out signal that you could then EQ with a mixer board. The micing method is obviously the most authentic, but you kinda need laboratory conditions for it :)

@TheOldskoolPC

4 жыл бұрын

Exactly, and as you can see, my basement is hardly kept in lab conditions :) So I think I'm going for the practical approach as well.

@TheOldskoolPC

4 жыл бұрын

Also, most 256bs these days are built in/for DOSBox, so that kind-of "solves" that problem. 256bs, if we're not careful, are going to morph into PICO-8 compos.

@gargaj

4 жыл бұрын

I actually think we're seeing a trend of 256bs moving away from DOSBox simply because they're starting to use an amount of processing power that's only feasible on real hardware; we have ~3Ghz CPUs now, so the bruteforce fixed-step raymarching render that was a terrible idea in '95 is perfectly plausible now :)

@TomCatAbaddon

4 жыл бұрын

And sadly DOSBox doesn't sounds good with all sound engine...

The distortions of a PC speaker are quite unique, i wonder if a a PC speaker inside a sound proof box would sound the same to a microphone. I guess the most authentic option is just building an amp and PC speaker to play the line out recorded files.

Not done with the video but already, micing the speaker (case off) is the way to go. You get the speaker's natural color in addition to the raw signal. Just like we do with guitar amplifiers. Followed closely with the sound card with internal generator which does a surprisignly good job

@TheOldskoolPC

Жыл бұрын

Yep, these are my conclusions too.

If you need to capture the authentic drone of the IBM PC's cooling fan, binaural mics FTW! Then you have the headaches of mixing soundcards and the beeper if using them in games..... (this applies to at least one PC game, Silpheed). The soundcard solutions solve this for the most part. Kinda curious how folks fare on other platforms with a beeper. The Apple II can be tricky even though multiple systems offered a real output jack (IIc and IIgs both have headphone jacks).

@TheOldskoolPC

4 жыл бұрын

That's a really good point about games that mix sound devices, Silpheed being a great example. I hadn't thought of that at all, which makes the soundcard solution even better as it saves you from having to buy a mixer.

The PAS16 more-or-less works in 8-bit machines even though the card is 16-bit. Obviously its Adlib, Game Port, MPU-401 and its "PC speaker snooping" features must work. Interestingly, the card has to be able to be listen to port writes from I/O addresses $40-6F, which are officially unavailable for expansion card usage in the IBM PC ($200-3FF allowed) and the IBM PC/XT ($100-3FF allowed). I guess you can write to expansion cards at any I/O address, but reading from them is more restricted. I believe the Tandy Sound Chip card in development can work at I/O $C0-CF in an IBM PC or XT.

Before I saw your results, I pretty much agreed with them. Tapping the speakers seems like it would be the best idea in theory, but in practice the sound just sounded too harsh most of the time. It's good to see that just a good mic and an open case gets the best results - anyone with a IBM PC or XT can do that without any exotic hardware.

@TheOldskoolPC

3 жыл бұрын

My challenge is to figure out how to do this with the case still on, as I'd like to set up a semi-permanent capture station while still having normal functionality and appearance of the system.

@ExperimentIV

2 жыл бұрын

@@TheOldskoolPC why don’t you try taking an IR of the inside of the closed pc case, then running the audio from the speaker tap through that in post?

Good video. It helped me find out more about the sound generation used in Digger, since you sort of connected it to "RealSound" for me with referencing it as pulse-modulation, so I went looking... Found this: www.gamejournal.it/the-sound-of-1-bit-technical-constraint-as-a-driver-for-musical-creativity-on-the-48k-sinclair-zx-spectrum/ I finally understand what "switching the speaker on and off faster than is physically capable" really means. The in-between states acquired via speaker cone inertia is what gives the other states beyond on or off, and therefore more diversity of sound output.

@TheOldskoolPC

3 жыл бұрын

Yep, now you've got it :-) Fairly mind-blowing to a teenage me.

Hey man, great video ! I'm into recording some pc speaker and I'm trying to buid the cable you're suggesting but wiith no success. Is there a photo of the finished circuit somewhere ? I think my mistake was that I've unplugged the speaker. I'd like to have kind of a permanent setup, maybe with Dupont connector instead of alligator clips... Many thanks for your help !

@TheOldskoolPC

7 ай бұрын

Awesome! There are some additional circuits and diagrams on the article I link to in the video's description; maybe that article can help? For my own testing with a simple alligator clip cable, sometimes I have to move around (really scrape) the clip attached to the case ground before it actually connects. Also, while it shouldn't matter, I've heard some advice that if you don't hear anything, try swapping positive for negative, or hot for ground. Let me know if you run into more trouble, happy to help with this.

@francesco_v

7 ай бұрын

@@TheOldskoolPC Hey, many thanks ! Finally I managed to get it working. It's easy as long as I dont cut the circuit from the speaker ! By the way, I've replaced the 10k resistor by a 1k one. The signal was very low on my mixing desk. With the 1k resistor it's perfect, hot enough without clipping. Thank you again for your work !

Nice. Weirdly, I liked the Tandy in several instances.

@TheOldskoolPC

4 жыл бұрын

It certainly has an extremely low noise floor. I liked it too, but the speed issue spooked me.

Also, regarding emulators... I agree 100% that the end goal should be to make someone feel like they're sitting in front of the real hardware. A commercial for the original Lexus LS400 comes to mind, v=3ao69itRG9Y. In that commercial they show all the typical car noises they've suppressed. That made me think the emulator should have check-boxes for "PC fan noise" "reverberation from PC case", etc.

@TheOldskoolPC

4 жыл бұрын

Maybe someday we'll get such a thing. Some Apple and Amiga emulators have options to turn on emulated disk drive seek/door noises.

@rzeka

4 жыл бұрын

The thing is, if you create an accurate-as-possible emulator and play it through a real speaker, it'll sound authentic. If you try to emulate the physical space that the speaker is in _and then_ play it through a real speaker, well now you have two levels of distortion and it's just gonna sound muffled. I like the idea of those checkboxes being there, but I would never want to use them.

Ahhh memories

thanks for great video. with combination of my childhood memory native DOS game sounds great. I would need to buy a ISA sound card with PC speaker input. so.. I think PAS16 is only sound card which can pickup PC speaker directly from ISA slot..?! I guess.

@TheOldskoolPC

Жыл бұрын

PAS16 is the only one I've personally used that can work without a cable from the speaker pins to the card, however I think I read that an Aztech card can do the same thing (but I could be wrong).

I'm also curious how micing the case would sound with a unidirectional instrument microphone, like the Shure SM-57. It should cancel more of the ambient noise from the room, but still capture the acoustic effect of the computer case, and not the just the speaker. Unfortunately, the fan noise would still be heard, but the acoustics of the room would probably matter less.

@TheOldskoolPC

4 жыл бұрын

Indeed, one opportunity I missed was trying to wire up a microphone INSIDE the case. Maybe an experiment for a future, follow-up video.

Recording from behind the speaker is the best. You hear what are you supposed to hear because sounds were designed on speaker. Also there was little noise compared to method 1 and 3. You can tell on effects test. Direct recording with wires altered how the effect sounded. Despite Tandy failing on a lot of tests the sound from it was kind of rounded i suppose they use some filters. And there were no peaks at the tone attack.

@TheOldskoolPC

Жыл бұрын

I agree; I just wish it wasn't so inconvenient to record this way. I may experiment putting a lav microphone inside the case and work the cable out.

Do you have the name of the exact song starting at around 2:46?

@TheOldskoolPC

4 жыл бұрын

"Trimline" by coda.

Would it be possible for emulator authors to enhance the Pulse Wave sounds to be less harsh? Or is it something intrinsic to the way these are produced?

@TheOldskoolPC

4 жыл бұрын

Actually, they already do: If you run these programs in DOSBox, the ones that run correctly (not all do) have no audible carrier wave (I'm guessing because the carrier wave is generated by the physical speaker itself).

Great job, very enlightening, but I was totally shocked by your evaluation in the end. Case-off method has neither richness/softness/authenticity of the outside-case method, nor the fidelity of any of the direct methods. For me, it is 50/50, half-assed compromise, neither authentic nor hi-quality, just something in the middle. If you need high fidelity, just use an external amp (maybe just a basic battery powered digital buffer) and maybe apply digital LP filter after recording at higher samplerate. This will sound the same as the later methods in the list but even cleaner and without their individual shortcomings. P.S. Square wave is not even supposed to look like a square wave even on cleanest recording, if sampled at normal audio samplerate by an audio ADC.

@TheOldskoolPC

4 жыл бұрын

Thank you for your honesty! Maybe I'm misunderstanding you, but I actually conclude in the very end of the video that I'll probably use one of the direct methods, as they satisfied my curiosity that they didn't utterly destroy the audio, while affording me a level of convenience. But if it's too harsh, I might try the outside case method again but with something hopefully easier to leave hooked up, like an attached mic with a cable running to my recorder.

12:34 What's is software/music name?

@sergiofls7623

4 жыл бұрын

Fantastic World Dizzy

@MasterNama

4 жыл бұрын

@@sergiofls7623 alright: kzread.info/dash/bejne/Z2hmldmmorW0qaQ.html

The tandy not picking up from the rca jack baffled me when i was a kid. Did anything come out of the systems internal speaker when recording? The tx has a headphone jack that should work instead of the rca audio jack.

@TheOldskoolPC

Ай бұрын

I'm not sure I follow; the RCA jack has usually worked for me across several Tandys...?

I think that the mic near the speaker with the case off lead to the most "authentic" sound. Before listening to the samples I thought that tapping the speaker leads directly would have obviously been the best solution, but it's actually far harsher to listen to. The way that the speaker adds it's own subtle touches to the signal, especially when clipping, really made most of the recordings sound "better" to me. Have you heard of taking Impulse Responses for Convolution? I'm wondering if you could take the raw line audio and convolve it with an impulse response of the speaker to get that "authentic" sound.

@TheOldskoolPC

3 жыл бұрын

I think you're right, and if you check the comments, others agree. My only issue is that it's a PITA to set up and keep set-up. I might try running a small lav mic in an open slot to see what that sounds like with the case on. I have not yet considered taking impulse responses -- I don't even know how to go about that, but it sounds promising!

@dbozan99

3 жыл бұрын

@@TheOldskoolPC Most people don't use convolution for modeling a speaker, they use it to model reverb; so don't mind that when looking for more info about. Creating an impulse response (DEconvolution) is a fairly straightforward process: 1. Play a frequency sweep (dry) on the speaker. 2. Record the sweep with a mic to get the "wet" version. 3. Convolve the inverted dry sweep with the wet sweep to get the "long" IR. 4. In a WAV editor such as Audacity, trim the "long" IR of excess silence before and after the impulse to get the IR. Once you have the IR, you should be able to use it with any convolution application/VST plugin/etc that accepts IRs in WAV format, which is most of them. Convolve a direct-line recording with the IR and ideally it should sound very similar to the recording. GratisVolver is free (you don't need the Pro version) and should do what you need it to, it even comes with the sweeps you need (dry and inverted, the 3 sec. will likely be fine). The hardest part I think will be actually playing the dry sound on the IBM's speaker; You'll likely need to disconnect the speaker from the IBM and use another device to drive the speaker instead. You will also likely need to isolate the speaker to ensure that it's in a very quiet area since you don't want the impulse response of the room, you just want the speaker's IR.

To my ear, the mic by the speaker with the case off made the best overall sound too. My NCR PC4 has a much better speaker than the IBM PC. I wonder if using a different speaker with the IBM could make the best solution? It's kind of like a Hammond Organ with a Leslie speaker. You need the physical speaker moving to get the best sound, taking a direct line feed doesn't sound right. Perhaps you could build a speaker into a small dampened box, and hook that up to the motherobard for something that's great for recording, and not as subject to room noise.

@TheOldskoolPC

4 жыл бұрын

That's not a bad idea. Taking it from line is the most convenient, but it sounds too "fat". I wish there was an emulator that could model various speaker types and housings.

Wich is the difference between the 2 schemes on the link? One has a capacitor, the other one has a 8ohm to 1kohm. Wich one should I use to connect the pc speaker out put to a stereo amplifier? Thaks.

@TheOldskoolPC

3 жыл бұрын

The one with a capacitor is meant to DC-offset and smooth the signal but leaves it at 5v which is a bit hot for some equipment. The one with a resistor is meant to reduce the +5v to something closer to line-in expected voltage. (At least, this is my basic understanding of both; if I'm wrong, I hope an engineer will correct me.) Before I knew line-in was 1V, I was feeding it 5V all the time and never had any issues, so my single experience is "it's safe". So either should work.

@MrEduedu123

3 жыл бұрын

@@TheOldskoolPC I will go for the one with a resistor. Thanks!

RDTSC is another way to get timing, this instruction is available since Pentium processors, could you please talk about it and also about the infamous Appendix H? I would like to hear the emotions and what the programmers felt back then since this was a secret from Intel. Thank you Jim, greetings from Colombia.

@TheOldskoolPC

2 жыл бұрын

I'm not sure I'll cover RDTSC because there are other videos that do so, and also because using RDTSC can be problematic in multi-core scenarios. Thanks for the comment!

For demonstration purposes, I prefer micing the speaker (case off). For archival purposes, I prefer the sound card with no cable

@TheOldskoolPC

Жыл бұрын

I agree 100% with this.

The sound card with internal cable seems like it's clipping, really not good in my opinion. I personally highly prefer the direct speaker tap. Feels authentic but with full fidelity. A close second would be the sound card no-cable. For the full experience I'd say recording with the case off is good, as that captures the tonality of the speaker but doesn't get affected as much by the room or muffled by the case. This really makes me want to hear System Beeps recorded with a PC speaker tap though! The extra bass is very welcome.

@TheOldskoolPC

Жыл бұрын

I agree the speaker tap is my favorite sonically, but unfortunately it picks up way too much system noise. Sound card no cable is best for eliminating noise.

Actually, IBM PS/1 computers have a sound card (a rather fringe one with not a lot of support, but for example ibmulator (extremely accurate IBM PS/1 emulator, for different models with different CPU's, etc. - mine was model 2011 I think, a 286 with 1MB RAM, 30 or 30MB HDD, CPU clockspeed of ~12Mhz (no lousy 4+Mhz 286's) that also used the sound card for PC speaker sounds and had headphone jacks. The output is cleaner than PC Speaker, and I would like to hear whether the hacks used to play digitized samples (PCM) through PS speaker, as the trick relied very much on how the PC speaker worked, but PS/1 sound card worked differently as it sent analog sound signal to regular speakers inside the PC, so I wonder if it would work fine, acceptably, poorly or if it works at all. Anyway, as this shows, there's other systems than ones mentioned in video, that can output audio meant for PC Speaker from headpdone jack as well - the IBM PS/1 computers.

@TheOldskoolPC

2 жыл бұрын

Appreciate the clarification! The PS/2 Model 25 also has a 1/4" headphone jack for speaker output, but it isn't line level and I also wanted to limit the scope for this particular video to 4.77 MHz systems.

Informative and a little surprising. But not the binaural mic's result, that was obvious throughout.

@TheOldskoolPC

4 жыл бұрын

I was hoping it would provide some sort of magical "you are there" result. It did... in the worst way.

@Scalibq

4 жыл бұрын

I think the binaural mic fails because it lacks a brain. As in, it probably captures exactly what your ears do, but your brain is good at focusing on a specific sound, and drowning out background noise (which is why you can hold a conversation in busy environments like a bar). People with hearing damage have problems with that. And your brain probably can't do it ont he binaural recording because some ambient 'cues' are missing.

Congrats, it’s a very nice work! I've repeated some of your tests with a different method. I used the motherboard's HDA mixer on my recent PC and notebook under native DOS. Here are the results: abaddon.hu/misc/PCS-HDA.zip

@TheOldskoolPC

4 жыл бұрын

For anyone who wants to capture the speaker off of a modern machine, this is a really great trick, and I encourage people to check it out!

why is there not an emulated pc (dosbox) in your list?

@TheOldskoolPC

3 жыл бұрын

Good question! It's because traditionally, emulation of the PC speaker has not always been accurate, especially in DOSBox where it is a kludge to save time and complexity. Some DOSBox forks have more accurate PC speaker emulation, such as dosbox-x which I recommend, but those more accurate emulations weren't finished when I was making the video. Also, I try to use real hardware whenever I can.

Ah, the "one bit honker"

Sound recordings are nyquist limited due to the sampling frequency of the recording device. Trying to judge the ”purest” square wave is fraught with difficulties. I.e. If you produce a perfect square wave and remove all frequencies above 22kHz (can’t hear anyways) and record it at 44 kHz sample rate it will not look like a square wave. It will wobble significantly whenever it hits a step change. If it looks perfectly square it has not been bandwidth limited and will alias badly. What happens is that frequencies above 44 kHz fold back into audible frequencies that are not harmonically related to the square wave and it sounds just terrible.

@TheOldskoolPC

2 жыл бұрын

Given that, what's the best procedure to sample the speaker? I don't think I mentioned it when I made the video, but all my recordings were at 96 KHz.

Hi Jim, I know this is an older video and this is probably largely settled but I wanted to pose the possibility of relocating the speaker outside of the case and recording it that way. Specifically, it would allow you to get the speaker away from the noisy system but also would allow pointing the speaker's cone directly at the mic which is how a recording engineer would (check out Steve Albini's excellent but admittedly overkill explainer video where he covers this far better than I can in a dumb KZread comment: kzread.info/dash/bejne/kaGmxsZuXcjOkps.html); currently you are capturing the sound waves irregular rebounding off the chassis and expansion cards. I wonder how much of a difference this could make. There's also the question of positioning the microphone various distances from the speaker and comparing the frequency expansion/collapse to determine which distance gives the most accurate and/or revealing sample. I wish this comment were more timely.

@TheOldskoolPC

Жыл бұрын

I agree outside of the case is another valid option that I wish I'd explored. The question, then, becomes what you are optimizing for. Meaning, since most people hear the speaker inside the case and bouncing off everything inside the case, would that not be the most appropriate for presentation? (I don't know, just posing the question.) Thanks for your thoughts!

_Translator:_ Is there any way to adjust the volume of this speaker in MS-DOS?

@TheOldskoolPC

Жыл бұрын

Usually no, although some clone systems did have a volume control, like the Tandy 1000 series.

I have a IBM PC from 1989 and it still has the PC Speaker in it

and 0.2W is still 0.2W too much

@TheOldskoolPC

11 ай бұрын

LOL

@chaoticsystem2211

11 ай бұрын

@@TheOldskoolPC i remember trying the soundcard approach once. i ripped the cable out after a few seconds :D