I can't believe I found this channel. So much high value information. Stuff people go to college for 4 years to learn about, wrapped up in a video the length of an episode of The Office.
Thanks! Most people who study computer engineering won't learn this stuff with a 4 year degree - it's usually taught from a very abstract perspective if at all. My goal was to consolidate these topics in a more practical way, and hopefully provide less abstract insight from someone in this field.
@@RTLEngineering Why did you stop uploading? Did you create another channel? Please come back, this content is amazing and there is very little content that is this in-depth.
just log in to say thank you for such nice videos. As a chip engineer who worked in GPU industry for some years, I know making a video like this really takes time. The dilemma is that for people who know this field well, they rarely have time/passion to produce content like this. So I am quite surprised to find this channel with such well organized contents with good clarity. I’d even recommend some of our new hire to watch this as intro materials. If people are interested enough to learn more about this field, they definitely won’t complain about TTS😊Looking forward to seeing more videos from you
i greatly appreciate your videos. the TTS isn't an issue at all for me because i know exactly what would have to go into making a video with voice over and the length and depth of content you make. awesome stuff! thank you!
Thanks! I'm not sure why there has been so much negative criticism of the TTS to be honest, especially considering that there are quite a few large channels that use it exclusively and in a much less fluid manner. As you said, part of it is balancing time, where the research and compilation of information depth means less time can be spent on production. I guess, that's something one can really only appreciate after making videos.
Amazingly practical content, thank you! The note about PS1 being twice as fast at filling sprites is consistent with my experience of its catalogue vs the N64’s. Games like Megaman X4 and Castlevania: Symphony of the Night had much more detailed spritework than (the relatively few) sprite-based titles like Pokemon: Puzzle League and… Mystic Makers? Lol. But it’s nice to see an explanation for that trend, especially given that I find the graphics of the N64 otherwise superior. Nintendo was truly committing to 3D and 3D alone with this console.
The PS1 was.. wild. You can forgive the Z buffer, but no floating point? Very odd. It's amazing what people pulled off with it though, like the great spyro series.
There's no source, just simple math: pixels/clk * fclk = pixels/sec. And yes, there are a lot of errors on wikipedia, although some of the contradictory sources were linked in the references. Was there something specific that you noticed?
@@RTLEngineering I believe the mhz on the gpu is 53 mhz. "The GPU clock is about 53MHz, with the exact frequency changing between NTSC and PAL - NTSC uses 53.693(18) Mhz and PAL uses 53.203425 MHz " The 53 mhz figure seems easier to find than the 34 mhz figure. The first link I got on stating that the GPU, not the cpu, was 34 mhz was a gamefaqs discussion with no source. I also rechecked wikipedia, and tried skimming it really quick, and had to do it...more than once; as I could not find any real information on the gpu speed and fillrate at all. Note, I don't always use google, I've used bing, yandex and more obscure options like marginalia. I'm really puzzled why the gpu clock and fillrate isn't common knowledge and easily checkable and why the 34 mhz figure eluded me.
Thanks for the feedback. I'm sorry that you feel that way. Unfortunately, the voice synthesis was chosen as a method to speed up production, since the previous videos I had done with recorded audio took far too long to record and edit. Previously, a video this length would have taken me 5 entire days for the audio alone, meaning 3 weekends. That would have been in addition to research, creating the graphics, writing the script, and editing the final video. Furthermore, editing audio is a task that I don't particularly enjoy doing. The final result would likely be a video of similar quality once every 2-3 months, though more likely I would stop making them altogether like I had done before switching to this workflow. Overall, I have had positive feedback regarding the voice synthesis (you have been the only one to voice a complaint). The audio does sound more normal at 1.5x speed if that helps, or you could mute the audio and read the closed captions. And of course you're free to not watch the videos, which would be equivalent to if I stopped uploading them.
@@RTLEngineering If you ever have the time you could always release a "Remastered Series" of Videos now with real human voice !!!! LOL. I know you got much more important things to do than this so do whatever works for you.
@@AxiomofDiscord I have been keeping that option in the back of my mind, although I would probably also take that opportunity to improve the video flow and correct any mistakes. I guess it would be sort of like an iterative approach, where the form is refined over time. I still have quite a few video topics in my queue that I would like to work on before making a second pass though.
I am aware, but it has been the best possible solution for me. As for passion, that's a blanket statement when there's more nuance. The passion involved is in communicating the information presented in the videos, especially since no one else has (at least at the depth presented). But I do lack the passion for spending 2 days recording 30 minutes of voice over, and then spending another 4 days editing the same audio so that it's bearable to listen to. If other solutions were more efficient for me, then I would use them - not using TTS would require an extra 6 days of production effort, which frankly is not worth the value that it adds, especially since I have other responsibilities that must come first (e.g. a day job).
Clearly you didn't watch the video, and commented after the first 5 seconds because you were annoyed by the TTS. If you're going to base a comment on that, just don't.... None of the information the video was taken from wikipedia, and in fact, some of the information on wikipedia is incorrect. I did use a few images (not the main diagrams) from wikipedia because they are in the public domain (why spend the time recreating them). So if I didn't use wikipedia, where did the information come from? Technical documentation, development conference presentations, computer graphics textbooks, computer graphics journal articles, analyzing and reverse engineering parts of the die photos, and extrapolation from all of the above. I am not trying to be a retro TH-camr. Retro TH-camrs focus on the historical context surrounding some technology, which I tend to gloss over. Instead, I focus on the engineering behind these systems and how they worked. I could just as easily cover modern systems like the PS5, but those are both more complicated and less interesting due to their reliance on parallelism (I did a video a while back going over the NVidia Tensor Core for example). The main advantage to covering something older like the PS1/N64 is that those systems have a lot more documentation, and are significantly easier for an audience to understand. Besides, all of the principles that were used in the older systems, are present in modern day hardware as well.
I can't believe I found this channel. So much high value information. Stuff people go to college for 4 years to learn about, wrapped up in a video the length of an episode of The Office.
Thanks! Most people who study computer engineering won't learn this stuff with a 4 year degree - it's usually taught from a very abstract perspective if at all. My goal was to consolidate these topics in a more practical way, and hopefully provide less abstract insight from someone in this field.
@@RTLEngineering Why did you stop uploading? Did you create another channel? Please come back, this content is amazing and there is very little content that is this in-depth.
@RTLEngineering Why didn't you narrate this yourself. Your voice is vastly superior to some AI generated voice from text.
just log in to say thank you for such nice videos. As a chip engineer who worked in GPU industry for some years, I know making a video like this really takes time. The dilemma is that for people who know this field well, they rarely have time/passion to produce content like this. So I am quite surprised to find this channel with such well organized contents with good clarity. I’d even recommend some of our new hire to watch this as intro materials. If people are interested enough to learn more about this field, they definitely won’t complain about TTS😊Looking forward to seeing more videos from you
i greatly appreciate your videos. the TTS isn't an issue at all for me because i know exactly what would have to go into making a video with voice over and the length and depth of content you make. awesome stuff! thank you!
Thanks! I'm not sure why there has been so much negative criticism of the TTS to be honest, especially considering that there are quite a few large channels that use it exclusively and in a much less fluid manner.
As you said, part of it is balancing time, where the research and compilation of information depth means less time can be spent on production. I guess, that's something one can really only appreciate after making videos.
34:46 This is what exactly what Kaze Emmanuar did with his improved Super Mario 64 engine. It requires an Expansion Pak for this reason.
This is so useful you have no idea - and it's not just about graphics, but for high performance computing in general.
Incredibly in-depth! Thanks for the lesson! I'm sure I'll be referring back to this many times in the future.
This is definitely a hidden gem of a channel. Hope those subs keep climbing, definitely deserve it.
Amazingly practical content, thank you!
The note about PS1 being twice as fast at filling sprites is consistent with my experience of its catalogue vs the N64’s. Games like Megaman X4 and Castlevania: Symphony of the Night had much more detailed spritework than (the relatively few) sprite-based titles like Pokemon: Puzzle League and… Mystic Makers? Lol.
But it’s nice to see an explanation for that trend, especially given that I find the graphics of the N64 otherwise superior. Nintendo was truly committing to 3D and 3D alone with this console.
The PS1 was.. wild. You can forgive the Z buffer, but no floating point? Very odd. It's amazing what people pulled off with it though, like the great spyro series.
Best hardware channel in TH-cam, thanks for the info my friend!
Man, this is some grade A quality content. I love it.
Superb channel , instant sub . A truly hidden gem
This channel is awesome!
16:55 didn't the Woz design the Apple II video circuitry to do DRAM refresh while drawing the screen, leading to a very unusual framebuffer layout?
Thank you for this
"Rambus go vroom vroom"
-Kaze Emanuar
this is some amazing stuff thanks bruh
I wish I had the sources for those peak power (pixels per second) figures. They seem to be contradictory to other sources, even wikipedia.
There's no source, just simple math: pixels/clk * fclk = pixels/sec. And yes, there are a lot of errors on wikipedia, although some of the contradictory sources were linked in the references.
Was there something specific that you noticed?
@@RTLEngineering I believe the mhz on the gpu is 53 mhz.
"The GPU clock is about 53MHz, with the exact frequency changing between NTSC and PAL - NTSC uses 53.693(18) Mhz and PAL uses 53.203425 MHz "
The 53 mhz figure seems easier to find than the 34 mhz figure. The first link I got on stating that the GPU, not the cpu, was 34 mhz was a gamefaqs discussion with no source.
I also rechecked wikipedia, and tried skimming it really quick, and had to do it...more than once; as I could not find any real information on the gpu speed and fillrate at all.
Note, I don't always use google, I've used bing, yandex and more obscure options like marginalia. I'm really puzzled why the gpu clock and fillrate isn't common knowledge and easily checkable and why the 34 mhz figure eluded me.
I wonder if you know enough to make one of these about the Dreamcast.
Already done - that video will be posted tonight.
Seriously, you need to stop it with the voice synthesis, it's incredibly obnoxious.
Thanks for the feedback. I'm sorry that you feel that way.
Unfortunately, the voice synthesis was chosen as a method to speed up production, since the previous videos I had done with recorded audio took far too long to record and edit. Previously, a video this length would have taken me 5 entire days for the audio alone, meaning 3 weekends. That would have been in addition to research, creating the graphics, writing the script, and editing the final video. Furthermore, editing audio is a task that I don't particularly enjoy doing. The final result would likely be a video of similar quality once every 2-3 months, though more likely I would stop making them altogether like I had done before switching to this workflow.
Overall, I have had positive feedback regarding the voice synthesis (you have been the only one to voice a complaint). The audio does sound more normal at 1.5x speed if that helps, or you could mute the audio and read the closed captions. And of course you're free to not watch the videos, which would be equivalent to if I stopped uploading them.
@@RTLEngineeringThe text synthesis is fine.
They literally made a whole video explaining why they switched to it. If it means they can keep producing these quality long form videos, so be it.
@@RTLEngineering If you ever have the time you could always release a "Remastered Series" of Videos now with real human voice !!!! LOL. I know you got much more important things to do than this so do whatever works for you.
@@AxiomofDiscord I have been keeping that option in the back of my mind, although I would probably also take that opportunity to improve the video flow and correct any mistakes. I guess it would be sort of like an iterative approach, where the form is refined over time. I still have quite a few video topics in my queue that I would like to work on before making a second pass though.
TTS sucks, my guy.
If you have zero passion, just don't.
I am aware, but it has been the best possible solution for me.
As for passion, that's a blanket statement when there's more nuance. The passion involved is in communicating the information presented in the videos, especially since no one else has (at least at the depth presented).
But I do lack the passion for spending 2 days recording 30 minutes of voice over, and then spending another 4 days editing the same audio so that it's bearable to listen to. If other solutions were more efficient for me, then I would use them - not using TTS would require an extra 6 days of production effort, which frankly is not worth the value that it adds, especially since I have other responsibilities that must come first (e.g. a day job).
@@RTLEngineering Copypasting some Wikipedia articles into TTS is not passion.
@@RTLEngineering Wannabe retro TH-camrs...
Clearly you didn't watch the video, and commented after the first 5 seconds because you were annoyed by the TTS. If you're going to base a comment on that, just don't....
None of the information the video was taken from wikipedia, and in fact, some of the information on wikipedia is incorrect. I did use a few images (not the main diagrams) from wikipedia because they are in the public domain (why spend the time recreating them). So if I didn't use wikipedia, where did the information come from? Technical documentation, development conference presentations, computer graphics textbooks, computer graphics journal articles, analyzing and reverse engineering parts of the die photos, and extrapolation from all of the above.
I am not trying to be a retro TH-camr. Retro TH-camrs focus on the historical context surrounding some technology, which I tend to gloss over. Instead, I focus on the engineering behind these systems and how they worked. I could just as easily cover modern systems like the PS5, but those are both more complicated and less interesting due to their reliance on parallelism (I did a video a while back going over the NVidia Tensor Core for example). The main advantage to covering something older like the PS1/N64 is that those systems have a lot more documentation, and are significantly easier for an audience to understand. Besides, all of the principles that were used in the older systems, are present in modern day hardware as well.